CN118010060B - Image processing-based single wheel lane route searching and planning method and system - Google Patents
Image processing-based single wheel lane route searching and planning method and system Download PDFInfo
- Publication number
- CN118010060B CN118010060B CN202410424071.8A CN202410424071A CN118010060B CN 118010060 B CN118010060 B CN 118010060B CN 202410424071 A CN202410424071 A CN 202410424071A CN 118010060 B CN118010060 B CN 118010060B
- Authority
- CN
- China
- Prior art keywords
- data
- track
- boundary
- generate
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000012545 processing Methods 0.000 title claims abstract description 176
- 238000000034 method Methods 0.000 title claims abstract description 61
- 238000004458 analytical method Methods 0.000 claims abstract description 53
- 238000005457 optimization Methods 0.000 claims abstract description 53
- 238000007405 data analysis Methods 0.000 claims abstract description 25
- 238000012544 monitoring process Methods 0.000 claims abstract description 20
- 238000001228 spectrum Methods 0.000 claims description 70
- 230000003595 spectral effect Effects 0.000 claims description 63
- 230000001502 supplementing effect Effects 0.000 claims description 35
- 238000004422 calculation algorithm Methods 0.000 claims description 30
- 238000004364 calculation method Methods 0.000 claims description 24
- 238000010191 image analysis Methods 0.000 claims description 23
- 238000012937 correction Methods 0.000 claims description 19
- 230000008569 process Effects 0.000 claims description 19
- 238000003708 edge detection Methods 0.000 claims description 11
- 238000009499 grossing Methods 0.000 claims description 10
- 230000002787 reinforcement Effects 0.000 claims description 10
- 238000001514 detection method Methods 0.000 claims description 9
- 230000007613 environmental effect Effects 0.000 claims description 9
- 238000012360 testing method Methods 0.000 claims description 9
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 238000012549 training Methods 0.000 claims description 8
- 230000004888 barrier function Effects 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 4
- 238000013075 data extraction Methods 0.000 claims description 3
- 238000003702 image correction Methods 0.000 claims description 3
- 238000005429 filling process Methods 0.000 claims 1
- 230000008859 change Effects 0.000 description 16
- 238000005516 engineering process Methods 0.000 description 12
- 238000005286 illumination Methods 0.000 description 12
- 238000010586 diagram Methods 0.000 description 6
- 238000013461 design Methods 0.000 description 5
- 230000007246 mechanism Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 238000002310 reflectometry Methods 0.000 description 4
- 230000004044 response Effects 0.000 description 4
- 230000009466 transformation Effects 0.000 description 4
- 230000001133 acceleration Effects 0.000 description 3
- 230000009471 action Effects 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000033001 locomotion Effects 0.000 description 3
- 230000007704 transition Effects 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 230000001154 acute effect Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 230000001105 regulatory effect Effects 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 description 1
- 238000012300 Sequence Analysis Methods 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000005452 bending Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 229940088597 hormone Drugs 0.000 description 1
- 239000005556 hormone Substances 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 210000001503 joint Anatomy 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000010183 spectrum analysis Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Landscapes
- Image Processing (AREA)
Abstract
The invention relates to the technical field of image data processing, in particular to a method and a system for searching and planning a single-wheel lane route based on image processing. The method comprises the following steps: the method comprises the steps of collecting track image data by using monitoring equipment arranged in a wheelbarrow, and generating track image data; performing dynamic binarization threshold analysis processing on the racetrack image data to generate dynamic binarization threshold data; performing track center line data analysis processing on the track image data according to the dynamic binarization threshold data to generate track center line data; carrying out single wheel lane route planning processing according to the track central line data to generate planned single wheel lane route data; designing wheelbarrow control parameters according to the planned wheelbarrow road data to obtain wheelbarrow control parameters; and executing the wheelbarrow driving optimization operation based on the wheelbarrow control parameters. The invention realizes a more accurate and efficient single wheel lane route searching and planning method.
Description
Technical Field
The invention relates to the technical field of image data processing, in particular to a method and a system for searching and planning a single-wheel lane route based on image processing.
Background
Along with the continuous development of the intelligent technical field, an intelligent wheelbarrow is developed according to the intelligent technology, and various games are extended, wherein the collected image data and the control balance of the wheelbarrow are automatically adjusted after computer hormones are utilized, so that the intelligent wheelbarrow can automatically complete the games. However, the conventional wheelbarrow road searching and planning method cannot accurately identify and process the driving paths of the wheelbarrow in the wheelbarrow track open-circuit area, the obstacle area and the like, so that the wheelbarrow road searching and planning efficiency is poor, the driving path deviation of the wheelbarrow cannot be effectively corrected, larger deviation occurs between the driving path and the planning path of the wheelbarrow, and the follow-up wheelbarrow road searching and planning accuracy is low.
Disclosure of Invention
Based on the above, the present invention provides a method and a system for searching and planning a road of a single wheel based on image processing, so as to solve at least one of the above technical problems.
In order to achieve the above purpose, a method for searching and planning a road of a single wheel based on image processing comprises the following steps:
S1, performing track image data acquisition by using monitoring equipment built in a wheelbarrow to generate track image data;
S2, performing dynamic binarization threshold analysis processing on the racetrack image data to generate dynamic binarization threshold data; performing track boundary data analysis processing on the track image data according to the dynamic binarization threshold data to generate track boundary data;
step S3, including:
S31, when track boundary data is detected, track boundary central line data analysis is carried out according to the track boundary data, and track boundary central line data are generated;
s32, when the track boundary data is not detected, performing track boundary line supplementing processing according to the track image data to generate track boundary line supplementing data; analyzing the track boundary line-supplementing central line data according to the track boundary line-supplementing data to generate track boundary line-supplementing central line data;
s4, performing track center line fitting optimization processing according to track boundary center line data and track patch boundary center line data to generate track center line data; carrying out single wheel lane route planning processing according to the track central line data to generate planned single wheel lane route data;
S5, carrying out path offset distance analysis according to the track image data and the planned single wheel lane road data to generate path offset distance data; designing wheelbarrow control parameters according to the path offset distance data and the planned wheelbarrow road data to obtain wheelbarrow control parameters; and executing the wheelbarrow driving optimization operation based on the wheelbarrow control parameters.
The invention uses the built-in monitoring equipment of the monocycle to collect the track image data, so that the monocycle can capture the detailed information of the surrounding environment in real time, and the instant environment perception capability provides accurate track information for the monocycle, thereby being the basis for realizing autonomous navigation and intelligent decision-making. The dynamic binarization threshold analysis processing is carried out on the track image data, and the monocycle can adaptively adjust the binarization threshold under the image condition, so that the track boundary is effectively identified, the accuracy of track boundary identification is greatly improved by the adaptive adjustment mechanism, and the influence of environmental change on the identification accuracy is reduced. The dynamic binarization threshold value data is used for analyzing and processing the track image data, so that the wheelbarrow can adapt to various road conditions, including different track materials, road marking definition, external illumination change and the like, track boundary data can be identified, keys for stable running can be kept in complex and variable environments, the complexity of the image data is reduced through the dynamic binarization processing, the image is converted into a format which is easier to analyze and process, and therefore the data processing efficiency of subsequent track boundary identification and road planning is optimized. When the boundary of the track is clear and identifiable, the central line of the track can be accurately calculated by the monocycle, which is important for the monocycle to safely and stably run along the track, and the optimal navigation performance of the monocycle under ideal conditions is ensured. Under the condition that the track boundary data is incomplete or unrecognizable, the missing track boundary information is supplemented through image data analysis, so that the adaptability of the wheelbarrow to complex or changing environments is improved, the track boundary information of a broken area, an intersection area and the like is identified, the wheelbarrow can find the most reasonable running path, the accurate calculation and the line supplementing processing of the track center line can be carried out, the wheelbarrow can be more accurately kept to run on the track, the deviation risk is reduced, and the running safety is improved. Through the track center line fitting optimization processing, the generated track center line data can reflect the center of an actual track to the greatest extent, so that a more accurate navigation path is provided, the optimized track center line can provide a smooth and direct driving path for the monocycle, unnecessary steering and adjustment are reduced, driving efficiency is improved, manual intervention is reduced in automatic road planning, autonomy of the monocycle is improved, and the monocycle can adapt to more complex and changeable environments. The design of the control parameters of the wheelbarrow is carried out according to the path offset distance data and the planned wheelbarrow lane data, the wheelbarrow is allowed to be dynamically adjusted according to the actual running condition, the capability of adapting to different track conditions is improved, the wheelbarrow can run along a preset path more accurately by analyzing the path offset distance and optimizing the control parameters according to the path offset distance, and the deviation and the error are reduced. The driving optimization operation of the monocycle is carried out, wherein the driving optimization operation of the monocycle comprises the steps of continuously learning feedback data of driving in control parameters by using a learning model, and iteratively updating the control parameters according to the driving feedback data.
The present disclosure provides an image processing-based single-wheel lane road searching and planning system for executing the image processing-based single-wheel lane road searching and planning method, which includes:
the track image data acquisition module is used for acquiring track image data by using monitoring equipment arranged in the wheelbarrow to generate track image data;
The track boundary analysis module is used for carrying out dynamic binarization threshold analysis processing on the track image data to generate dynamic binarization threshold data; performing track boundary data analysis processing on the track image data according to the dynamic binarization threshold data to generate track boundary data;
the track boundary data acquisition module is used for analyzing the track boundary central line data according to the track boundary data when the track boundary data are detected, so as to generate track boundary central line data; when the track boundary data is not detected, carrying out track boundary line supplementing processing according to the track image data to generate track boundary line supplementing data; analyzing the track boundary line-supplementing central line data according to the track boundary line-supplementing data to generate track boundary line-supplementing central line data;
The single wheel lane road planning module is used for performing track center line fitting optimization processing according to track boundary center line data and track line supplementing boundary center line data to generate track center line data; carrying out single wheel lane route planning processing according to the track central line data to generate planned single wheel lane route data;
The wheelbarrow driving module is used for carrying out path offset distance analysis according to the racetrack image data and the planned wheelbarrow road data to generate path offset distance data; designing wheelbarrow control parameters according to the path offset distance data and the planned wheelbarrow road data to obtain wheelbarrow control parameters; and executing the wheelbarrow driving optimization operation based on the wheelbarrow control parameters.
The method has the advantages that the method can accurately identify and process the driving paths of the wheelbarrow in the wheelbarrow track open-circuit area, the obstacle area and the like, and plan the whole driving paths of the wheelbarrow by carrying out line supplementing processing on the unidentified road boundaries, so that the efficiency of the wheelbarrow road searching and planning is high, the driving path deviation of the wheelbarrow can be effectively corrected, the wheelbarrow control parameters are continuously optimized and adjusted through the reinforcement learning model, the driving path of the wheelbarrow accords with the planning path, and the accuracy of the subsequent wheelbarrow road searching and planning is higher.
Drawings
FIG. 1 is a schematic flow chart of a method for searching and planning a road of a wheel road based on image processing;
FIG. 2 is a detailed flowchart illustrating the implementation of step S1 in FIG. 1;
FIG. 3 is a detailed flowchart illustrating the implementation of step S2 in FIG. 1;
FIG. 4 is a flowchart illustrating the detailed implementation of step S4 in FIG. 1;
FIG. 5 is a flowchart illustrating the detailed implementation of step S5 in FIG. 1;
the achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
The following is a clear and complete description of the technical method of the present invention, taken in conjunction with the accompanying drawings, and it is evident that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, are intended to fall within the scope of the present invention.
Furthermore, the drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus a repetitive description thereof will be omitted. Some of the block diagrams shown in the figures are functional entities and do not necessarily correspond to physically or logically separate entities. The functional entities may be implemented in software or in one or more hardware modules or integrated circuits or in different networks and/or processor methods and/or microcontroller methods.
It will be understood that, although the terms "first," "second," etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another element. For example, a first element could be termed a second element, and, similarly, a second element could be termed a first element, without departing from the scope of example embodiments. The term "and/or" as used herein includes any and all combinations of one or more of the associated listed items.
In order to achieve the above objective, referring to fig. 1 to 5, the present invention provides a method for searching and planning a road of a wheel road based on image processing, comprising the following steps:
S1, performing track image data acquisition by using monitoring equipment built in a wheelbarrow to generate track image data;
S2, performing dynamic binarization threshold analysis processing on the racetrack image data to generate dynamic binarization threshold data; performing track boundary data analysis processing on the track image data according to the dynamic binarization threshold data to generate track boundary data;
step S3, including:
S31, when track boundary data is detected, track boundary central line data analysis is carried out according to the track boundary data, and track boundary central line data are generated;
s32, when the track boundary data is not detected, performing track boundary line supplementing processing according to the track image data to generate track boundary line supplementing data; analyzing the track boundary line-supplementing central line data according to the track boundary line-supplementing data to generate track boundary line-supplementing central line data;
s4, performing track center line fitting optimization processing according to track boundary center line data and track patch boundary center line data to generate track center line data; carrying out single wheel lane route planning processing according to the track central line data to generate planned single wheel lane route data;
S5, carrying out path offset distance analysis according to the track image data and the planned single wheel lane road data to generate path offset distance data; designing wheelbarrow control parameters according to the path offset distance data and the planned wheelbarrow road data to obtain wheelbarrow control parameters; and executing the wheelbarrow driving optimization operation based on the wheelbarrow control parameters.
The invention uses the built-in monitoring equipment of the monocycle to collect the track image data, so that the monocycle can capture the detailed information of the surrounding environment in real time, and the instant environment perception capability provides accurate track information for the monocycle, thereby being the basis for realizing autonomous navigation and intelligent decision-making. The dynamic binarization threshold analysis processing is carried out on the track image data, and the monocycle can adaptively adjust the binarization threshold under the image condition, so that the track boundary is effectively identified, the accuracy of track boundary identification is greatly improved by the adaptive adjustment mechanism, and the influence of environmental change on the identification accuracy is reduced. The dynamic binarization threshold value data is used for analyzing and processing the track image data, so that the wheelbarrow can adapt to various road conditions, including different track materials, road marking definition, external illumination change and the like, track boundary data can be identified, keys for stable running can be kept in complex and variable environments, the complexity of the image data is reduced through the dynamic binarization processing, the image is converted into a format which is easier to analyze and process, and therefore the data processing efficiency of subsequent track boundary identification and road planning is optimized. When the boundary of the track is clear and identifiable, the central line of the track can be accurately calculated by the monocycle, which is important for the monocycle to safely and stably run along the track, and the optimal navigation performance of the monocycle under ideal conditions is ensured. Under the condition that the track boundary data is incomplete or unrecognizable, the missing track boundary information is supplemented through image data analysis, so that the adaptability of the wheelbarrow to complex or changing environments is improved, the track boundary information of a broken area, an intersection area and the like is identified, the wheelbarrow can find the most reasonable running path, the accurate calculation and the line supplementing processing of the track center line can be carried out, the wheelbarrow can be more accurately kept to run on the track, the deviation risk is reduced, and the running safety is improved. Through the track center line fitting optimization processing, the generated track center line data can reflect the center of an actual track to the greatest extent, so that a more accurate navigation path is provided, the optimized track center line can provide a smooth and direct driving path for the monocycle, unnecessary steering and adjustment are reduced, driving efficiency is improved, manual intervention is reduced in automatic road planning, autonomy of the monocycle is improved, and the monocycle can adapt to more complex and changeable environments. The design of the control parameters of the wheelbarrow is carried out according to the path offset distance data and the planned wheelbarrow lane data, the wheelbarrow is allowed to be dynamically adjusted according to the actual running condition, the capability of adapting to different track conditions is improved, the wheelbarrow can run along a preset path more accurately by analyzing the path offset distance and optimizing the control parameters according to the path offset distance, and the deviation and the error are reduced. The driving optimization operation of the monocycle is carried out, wherein the driving optimization operation of the monocycle comprises the steps of continuously learning feedback data of driving in control parameters by using a learning model, and iteratively updating the control parameters according to the driving feedback data.
In the embodiment of the present invention, as described with reference to fig. 1, a step flow diagram of a method for searching and planning a road of a monowheel based on image processing according to the present invention is provided, and in the embodiment, the method for controlling a flood diversion thyristor based on information fusion includes the following steps:
S1, performing track image data acquisition by using monitoring equipment built in a wheelbarrow to generate track image data;
According to the embodiment of the invention, the built-in monitoring equipment of the monocycle, such as a multispectral high-definition camera, is ensured to be in a good working state, the focal length and the angle of the camera are calibrated, and the track image is captured at the optimal visual angle. Parameters of image acquisition, such as resolution, frame rate and the like, are set so as to adapt to different illumination conditions and track environments, and corresponding exposure settings are adjusted in strong light or low light environments to ensure image quality. Starting a continuous acquisition mode of the monitoring equipment, capturing an image sequence of the track in real time, ensuring to acquire dynamic change track information, acquiring image data under different spectrums, carrying out image reconstruction based on the track images under different spectrums to obtain track images with more color discrimination, and preprocessing the reconstructed track images, such as denoising, adjusting brightness and contrast, to generate track image data.
S2, performing dynamic binarization threshold analysis processing on the racetrack image data to generate dynamic binarization threshold data; performing track boundary data analysis processing on the track image data according to the dynamic binarization threshold data to generate track boundary data;
In the embodiment of the invention, according to the illumination intensity and the color distribution characteristic of the racetrack image data, a binarization threshold value is dynamically calculated by adopting an adaptive algorithm, and the method can comprise the steps of analyzing the histogram distribution of the image and determining an optimal threshold value to divide racetrack and non-racetrack areas. And (3) performing binarization processing on the racetrack image data by using the calculated dynamic threshold value, converting the image into a form which only contains two colors of black and white, and simplifying the identification and analysis of racetrack boundaries. And processing the binarized image by adopting an edge detection algorithm, such as Canny edge detection, and identifying the boundary line of the race track. And extracting boundary line data of the track based on the edge detection result, and performing preliminary shape analysis, such as straight line and curve detection, to determine the rough layout of the track and generate the track boundary data.
Step S3, including:
S31, when track boundary data is detected, track boundary central line data analysis is carried out according to the track boundary data, and track boundary central line data are generated;
s32, when the track boundary data is not detected, performing track boundary line supplementing processing according to the track image data to generate track boundary line supplementing data; analyzing the track boundary line-supplementing central line data according to the track boundary line-supplementing data to generate track boundary line-supplementing central line data;
In the embodiment of the present invention, step S3 includes a case where track boundary data is detected and a case where track boundary data is not detected. When the track boundary data is detected, the accuracy of the track boundary data is confirmed, for example, by image processing software or an algorithm, the track boundary data is ensured to be correctly identified, the center line distance of the boundary line is extracted based on the left boundary line and the right boundary line of the track, the center line distance is marked as track boundary center line data, and the track boundary center line data can be subjected to post-processing such as outlier removal and smoothing processing. Or when track boundary data is detected, for example, track image data is photographed right at an intersection area and a break area of a track, a track boundary line is not recognized at a middle portion of the intersection area, a boundary line evaluation process is performed on the unidentified track boundary line, track boundary information in front and rear frame images is used, or an estimation of the track boundary is performed according to known characteristics of the track, and a line compensation process is applied to the unidentified track boundary by selecting an appropriate line compensation algorithm, such as a method based on linear interpolation or curve fitting. Extracting the track centerline from the completed track boundary data requires the application of specific image processing algorithms to ensure the consistency of the centerline of the patch portion with the original track centerline. And (3) optimizing the extracted line-supplementing central line to ensure that the central line after line supplementing is smooth and truly reflects the running path of the track, wherein the line-supplementing central line can be subjected to smoothing treatment and abnormal point correction.
S4, performing track center line fitting optimization processing according to track boundary center line data and track patch boundary center line data to generate track center line data; carrying out single wheel lane route planning processing according to the track central line data to generate planned single wheel lane route data;
In the embodiment of the invention, firstly, curve fitting is carried out on track boundary central line data and track line supplementing boundary central line data by using polynomial regression analysis, smoothness of the curve is ensured by adopting a cubic spline curve, the curve is obtained by taking the cubic spline curve as a fitting parameter through least square method calculation, best fitting of the central line to an actual track is ensured, smoothing treatment is carried out on the central line obtained by fitting, high-frequency noise is reduced by using a Gaussian filter, and smoothness of the central line and running safety are ensured. Based on The algorithm performs path planning, takes track center line data as a basis of a navigation path, simultaneously takes dynamic constraint of the wheelbarrow into consideration to perform path optimization, sets heuristic functions as Euclidean distance of the path, calculates an optimal path from a starting point to an end point according to the algorithm, takes obstacle avoidance and path smoothness into consideration to plan a wheelbarrow road, and generates planning wheelbarrow road data.
S5, carrying out path offset distance analysis according to the track image data and the planned single wheel lane road data to generate path offset distance data; designing wheelbarrow control parameters according to the path offset distance data and the planned wheelbarrow road data to obtain wheelbarrow control parameters; and executing the wheelbarrow driving optimization operation based on the wheelbarrow control parameters.
In the embodiment of the invention, the path offset distance analysis is performed based on the central node position of the track image data and the planned wheelbarrow road data, and the offset distance of the real running path and the expected running path is calculated, wherein the offset value under the time window is considered to obtain the path offset distance data, and the control parameters of the wheelbarrow are designed according to the path offset distance data, including the speed, the steering angle and the like, so that the wheelbarrow is adjusted to return to the planned path, the wheelbarrow can be ensured to quickly and effectively correct the path deviation, and the running on the optimal path is kept. Based on the designed control parameters, the driving optimization operation of the monocycle is executed, and the driving system of the monocycle is adjusted in real time to adapt to the control instructions, so that the driving performance optimization of the monocycle is realized, the driving stability and the driving safety are improved, and meanwhile, the efficient power output is maintained. And collecting feedback data of driving operation, such as actual running track, speed change and the like, adjusting control parameter design according to the feedback data, entering the optimization loop of the next round, and continuously improving the precision and response speed of the monocycle control system through continuous iterative optimization so as to adapt to complex and changeable track environments.
Preferably, step S1 comprises the steps of:
s11, acquiring an initial track image by using monitoring equipment built in a wheelbarrow, and generating initial track image data;
s12, performing spectral image analysis processing on the initial racetrack image to generate spectral image data;
s13, performing noise spectrum calculation processing on the spectral image data by using a noise spectrum calculation formula to generate noise spectrum data;
S14, carrying out correction spectrum data analysis according to the noise spectrum data to generate correction spectrum data;
Step S15, performing feature point selection processing on the spectral image data to generate spectral image feature point data;
s16, analyzing dynamic jitter data of the spectral image according to the characteristic point position data of the spectral image to generate dynamic jitter data;
S17, analyzing geometrical calibration data of the spectral image according to the dynamic jitter data to generate geometrical calibration data of the spectral image;
And S18, performing track image correction and reconstruction processing on the initial track image data according to the corrected spectrum data and the spectrum image geometric calibration data to generate track image data.
According to the invention, through initial track image acquisition, an original data basis is provided for subsequent image processing and analysis, and the starting point of data processing is ensured to have high-quality image information. The spectral image analysis processing can reveal the characteristics of the image under different spectrums, provide richer information for track identification, and is beneficial to improving the identification accuracy. Through noise spectrum calculation processing, noise components in the image are identified, a basis is provided for subsequent image noise reduction and definition improvement, and image quality is improved. And (3) carrying out correction spectrum data analysis according to the noise spectrum data, and adjusting image data corresponding to different spectrums to adapt to different illumination conditions and environmental changes, so that the environmental adaptability of the system is improved. And the spectral image data is subjected to characteristic point position selection processing, and the characteristic point position selection processing focuses on the information which is most critical to the track identification in the image, so that the efficiency and accuracy of data processing are improved. Dynamic variation in the image acquisition process is identified and measured through dynamic jitter data analysis of the spectrum image, and a strategy is provided for stable image acquisition. And according to the analysis of the geometric calibration data of the spectrum image, the geometric deviation of the image is adjusted, the geometric characteristics of the racetrack reflected by the image are ensured to be consistent with the actual conditions, and the use value of the image is improved. And the initial track image data is corrected and reconstructed by combining the corrected spectrum data and the spectrum image geometric calibration data, and the quality of the finally generated track image data is obviously improved, so that a high-quality visual basis is provided for subsequent road planning and navigation decisions.
As an example of the present invention, referring to fig. 2, a detailed implementation step flow diagram of step S1 in fig. 1 is shown, where in this embodiment, step S1 includes:
s11, acquiring an initial track image by using monitoring equipment built in a wheelbarrow, and generating initial track image data;
In the embodiment of the invention, multispectral monitoring is used for capturing detailed images of the racetrack under different spectrums, including racetrack images under infrared light and visible light, and camera parameters such as exposure time and ISO sensitivity are set to ensure that clear images can be obtained under various illuminations so as to generate initial racetrack image data.
S12, performing spectral image analysis processing on the initial racetrack image to generate spectral image data;
In the embodiment of the invention, the image is decomposed into components with different spectrums, such as RGB or a wider spectrum range, by applying a spectrum analysis technology, so that the content of the image is better analyzed, and a proper spectrum decomposition method and parameters are selected according to the spectrum distribution of the image to generate spectrum image data.
S13, performing noise spectrum calculation processing on the spectral image data by using a noise spectrum calculation formula to generate noise spectrum data;
In the embodiment of the invention, spectral frequency conversion is carried out on spectral image data, and noise spectrum caused by the influence of environmental factors and the like is calculated based on a preset standard spectral threshold value under each spectrum and converted spectral frequency calculation, so as to generate noise spectrum data.
S14, carrying out correction spectrum data analysis according to the noise spectrum data to generate correction spectrum data;
According to the embodiment of the invention, according to noise spectrum data, influence parameters caused by reasons such as environment and the like under different spectrums are analyzed, and correction spectrum data are established based on the influence parameters, so that pictures of acquired racetrack image data are unified, the establishment of a subsequent dynamic binarization threshold value, the identification of racetrack boundaries and the like are facilitated, and the correction spectrum data are generated.
Step S15, performing feature point selection processing on the spectral image data to generate spectral image feature point data;
in the embodiment of the invention, a feature point detection algorithm, such as SIFT, SURF or ORB, is adopted to extract key feature points from the spectral image data, and parameters specific to the algorithm, such as the scale, direction and the like of the feature points, are set so as to ensure the representativeness and the distinguishing degree of the feature points, so as to generate the spectral image feature point data.
S16, analyzing dynamic jitter data of the spectral image according to the characteristic point position data of the spectral image to generate dynamic jitter data;
In the embodiment of the invention, the dynamic jitter data analysis of the spectral image is carried out on the spectral image characteristic point position data by utilizing the image sequence analysis technology, the image jitter caused by the movement of a camera or a monocycle is evaluated, the spectral images acquired by the light-sensitive cameras corresponding to different spectrums under the monitoring equipment are possibly non-uniform, the movement track of the characteristic points in the image sequence is analyzed, the amplitude and the direction of the jitter are calculated, the dynamic jitter data are generated, and the dynamic jitter data of the spectral images acquired under different light-sensitive cameras are reflected.
S17, analyzing geometrical calibration data of the spectral image according to the dynamic jitter data to generate geometrical calibration data of the spectral image;
in the embodiment of the invention, geometric transformation technology such as affine transformation or perspective transformation is applied to calibrate the dithered image, correct geometric shape of the image is restored, parameters of geometric transformation under different spectrum images such as rotation angle, scaling and translation distance are determined according to dynamic dithering data, and spectrum image geometric calibration data is generated.
And S18, performing track image correction and reconstruction processing on the initial track image data according to the corrected spectrum data and the spectrum image geometric calibration data to generate track image data.
In the embodiment of the invention, the results of spectral correction and geometric correction are comprehensively applied, final correction reconstruction is carried out on the initial track image data so as to eliminate the influence of noise and jitter, the real situation of the track image is recovered, and the corrected different spectral images are aligned and reconstructed so as to generate the track image data.
Preferably, the noise spectrum calculation formula in step S13 is as follows:
;
In the method, in the process of the invention, Expressed as wavelength/>Noise spectrum data at time,/>Represented as spectral weight information,Expressed as wavelength/>Time-base raw spectral signal intensity,/>Expressed as/>Environmental impact spectrum intensity signal,/>Expressed as wavelength/>Spectral reflectance at time,/>Expressed as spectral angle of incidence,/>Reference spectral signal intensity expressed as feature point location,/>Expressed as the actual spectral signal intensity of the feature point.
The invention utilizes a noise spectrum calculation formula, whereinReflecting noise levels for different spectral wavelengths; /(I)Reflecting the weight information distributed under different spectrum irradiation; /(I)Reflecting the intensity of the clean spectral signal emitted by the light source without any external interference (e.g., ambient light); /(I)Representing the intensity of the influence of ambient light or other external light sources on the optical signal at that wavelength, such as sunlight or indoor illumination; /(I)Reflecting the ability of the object surface to reflect light at that wavelength, the magnitude of the reflectivity directly affecting the intensity of the spectral signal; /(I)The method is characterized in that the included angle between the light ray and the normal direction when the light ray is incident on the surface of an object influences the intensity and the distribution of a spectrum signal, the influence of the spectrum incident angle on the signal intensity is considered, the adjustment effect of the light ray incident angle on the spectrum reflectivity is reflected through geometric and trigonometric function processing, and the influence of external noise on the spectrum reflectivity is facilitated to be identified; /(I)/>Is the reference spectrum signal intensity and the actual spectrum signal intensity under the preset characteristic point position,/>Reflecting deviations caused by other external environmental influences. By calculating/>The relation between the signal intensity and the spectral reflectivity under the influence of the original spectral signal and the ambient light is reflected, the real situation of the spectral signal under the given environment is accurately estimated, and the influence of noise on the signal quality is reflected through the spectral weight information and the difference value between the reference spectral signal intensity and the actual spectral signal intensity under the characteristic point position, and the noise spectrum under different specific wavelengths is accurately identified, particularly in specific areas or wavelengths in a spectral image. The noise spectrum corresponding to different spectrum wavelengths can be accurately identified through the noise spectrum calculation formula, so that the noise spectrum can be used for image optimization and reconstruction in the subsequent step.
Preferably, step S2 comprises the steps of:
Step S21, data real-time reading is carried out on the track image data to generate read track image data, wherein the read track image data comprises read image sequence data and real-time track image data;
S22, performing gray level image conversion processing on the real-time track image data to generate gray level track image data;
s23, establishing an image analysis queue according to the read image sequence data, and transmitting the gray-scale racetrack image data to the image analysis queue to generate queue racetrack image data;
S24, carrying out dynamic binarization threshold analysis on the queue track image data to generate dynamic binarization threshold data;
s25, performing image binarization traversal on the queue track image data according to the dynamic binarization threshold value data to generate binarization track image data;
Step S26, performing track boundary detection processing on the binarized track image data by utilizing edge detection to generate track boundary data.
By reading the image data of the track in real time, the system can immediately acquire the latest track condition, including the image sequence data and the image data of the track in real time, which provides a basis for quick response to road surface changes. The real-time track image data is converted into the gray level image, so that the complexity of the image is simplified, the calculated amount is reduced for the subsequent image analysis and processing, and the processing speed and efficiency are improved. By establishing the image analysis queue, the system can process the image data in sequence, thereby ensuring the consistency and the order of the image processing, optimizing the flow of the image processing and improving the processing speed. Dynamic binarization threshold analysis is carried out on the racetrack image data in the queue, so that the binarization threshold can be adjusted according to the specific characteristics of the real-time image, and the adaptability to racetrack changes is improved. The images are subjected to binarization traversal processing by using the dynamically determined binarization threshold value, so that the racetrack and non-racetrack areas can be more accurately segmented, and a basis is provided for accurate identification of racetrack boundaries. The edge detection technology is used for detecting the track boundary of the binarized image, so that the track boundary can be accurately identified, and important basic data is provided for the subsequent track center line extraction and navigation planning.
As an example of the present invention, referring to fig. 3, a detailed implementation step flow diagram of step S2 in fig. 1 is shown, where in this embodiment, step S2 includes:
Step S21, data real-time reading is carried out on the track image data to generate read track image data, wherein the read track image data comprises read image sequence data and real-time track image data;
In the embodiment of the invention, the real-time data transmission setting of the built-in monitoring equipment (such as a camera) of the monocycle is configured, a high-speed data transmission technology such as USB 3.0, HDMI or wireless transmission technology is used for ensuring the rapid and real-time reading of image data, and a buffer technology such as a ring buffer area is used for managing the real-time data flow so as to avoid the data loss. The method comprises the steps of ensuring that track images can be captured and transmitted without delay, setting the acquisition frequency of image data to match the speed and processing capacity of a wheelbarrow, ensuring real-time performance and data continuity, receiving the real-time image data through a control system of the wheelbarrow, carrying out preliminary formatting processing, and recording the sequence data of the acquired real-time track images to generate read track image data.
S22, performing gray level image conversion processing on the real-time track image data to generate gray level track image data;
In the embodiment of the invention, a gray conversion algorithm is applied to the received real-time track image data, a color image is converted into a gray image, after gray conversion, the image is subjected to standardized processing, and the brightness and contrast of the image are adjusted, including histogram equalization and other technologies, so as to improve the recognizability of the track characteristics and generate the gray track image data.
S23, establishing an image analysis queue according to the read image sequence data, and transmitting the gray-scale racetrack image data to the image analysis queue to generate queue racetrack image data;
In the embodiment of the invention, an image analysis queue is created for managing and ordering the gray track image data to be analyzed, ensuring the sequence and the high efficiency of image analysis, adding the converted gray track image data into the image analysis queue according to the time sequence of acquisition, and setting a queue processing mechanism, such as a first-in first-out (FIFO), for optimizing the processing flow of the image data to generate the queue track image data.
S24, carrying out dynamic binarization threshold analysis on the queue track image data to generate dynamic binarization threshold data;
in the embodiment of the invention, the gray distribution of the racetrack image data in the queue is analyzed, the optimal binarization threshold value suitable for the current image set is determined, factors such as illumination change and shadow in the image are considered, the binarization threshold value is dynamically adjusted so as to meet the image processing requirements under different illumination conditions, and a statistical method, such as calculating the extreme value of an image gray histogram, is used as a reference for threshold value adjustment to generate dynamic binarization threshold value data.
S25, performing image binarization traversal on the queue track image data according to the dynamic binarization threshold value data to generate binarization track image data;
In the embodiment of the invention, a dynamic binarization threshold value is applied to carry out binarization processing on each frame of gray level race track image in a queue, each pixel value in the image is compared with the threshold value, the setting larger than the threshold value is white (race track) and the setting smaller than the threshold value is black (non-race track), and the image data after the binarization processing is stored to generate binarization race track image data for preparing boundary detection.
Step S26, performing track boundary detection processing on the binarized track image data by utilizing edge detection to generate track boundary data.
In the embodiment of the invention, the binarized track image is processed by utilizing an edge detection algorithm, such as Canny edge detection, a clear track boundary line is identified, an edge detection result is analyzed, a continuous boundary line segment is identified, the identified track boundary line data is marked and stored, track boundary data is generated, and a foundation is provided for subsequent track center line fitting.
Preferably, step S24 comprises the steps of:
performing extreme gray value analysis on the queue race track image data to respectively generate minimum gray value data and maximum gray value data;
when the queue track image data is updated, respectively carrying out extreme value gray value updating on the minimum gray value data and the maximum gray value data according to the updated queue track image data so as to obtain the updated minimum gray value data and the updated maximum gray value data;
And establishing a dynamic binarization threshold according to the updated minimum gray value data and the updated maximum gray value data to generate dynamic binarization threshold data.
The invention can accurately capture the brightness range of the image by analyzing the extreme gray values, namely the minimum gray value and the maximum gray value, in the image data of the queue track, provides a key reference basis for the subsequent dynamic binarization threshold setting, is beneficial to adapting to the scene of the color change of the track, and ensures that the effective identification of the track boundary is not limited by the reasons of illumination conditions, the color change of the track and the like. With the updating of the image data of the queue track, the minimum and maximum gray value data are updated in time, so that the calculation of the dynamic binary threshold value is ensured to be based on the latest image data, the real-time updating mechanism improves the adaptability of the system, and an algorithm can respond to the change of the real-time environment in real time, thereby improving the accuracy of track identification and the response speed of the system. Based on the updated extreme gray value data, a dynamic binarization threshold is established, so that the flexibility and effect of binarization processing are improved, the dynamic binarization threshold can be effectively adapted to illumination changes and track environments in different scenes, binarization errors caused by fixed threshold values are reduced, and the accuracy and reliability of track boundary detection are improved. In addition, the dynamic adjustment of the threshold value is also helpful for reducing the influence of background noise, clearly highlighting the track boundary, and providing a higher-quality image basis for track boundary detection and subsequent navigation decisions.
In the embodiment of the invention, gray value statistical analysis is carried out on each frame of racetrack image data in the queue, the minimum and maximum gray values of each frame of image are calculated, an image processing library such as OpenCV is used, the gray extreme value of each frame of image is rapidly obtained through a statistical function provided by the image processing library, the minimum and maximum gray values of each frame of image are recorded, and a time sequence extreme value database is established and used for storing the minimum gray value data and the maximum gray value data. Setting a monitoring mechanism for monitoring the update condition of the queue track image data, immediately triggering the update flow of the extreme value gray value once new image data is added into the queue, and recalculating the minimum gray value and the maximum gray value according to the newly added track image data, and updating the extreme value database to ensure that each analysis is performed based on the latest image data. According to the updated minimum gray value data and maximum gray value data, a dynamic calculation method is adopted to determine a binarized threshold value, an optimal threshold value is automatically determined by using an Otsu algorithm, or according to the average value of the minimum gray value and the maximum gray value as a threshold value, the dynamic adjustment of the threshold value is critical to adapt to the change of the environment in consideration of the illumination conditions of different images and the change of the track characteristics. The calculated dynamic binarization threshold value is applied to the track image data in the queue, binarization processing is carried out, the binarization result is evaluated, the threshold value setting is ensured to accurately divide the track and non-track areas, and fine adjustment is carried out if necessary, so that dynamic binarization threshold value data are generated.
Preferably, the track boundary line-filling processing according to the track image data in step S32 includes the following steps:
performing unidentified boundary racetrack image extraction processing according to racetrack image data to generate unidentified boundary racetrack image data;
Performing boundary tracking data analysis processing on unidentified boundary track image data based on track boundary data corresponding to track image data in an image analysis queue to generate boundary tracking data;
performing boundary fracture node identification processing on unidentified boundary racetrack image data based on the boundary tracking data to generate boundary fracture node data;
performing unrecognized boundary offset distance analysis processing according to the boundary fracture node data to generate unrecognized boundary offset distance data;
performing track boundary line interpolation processing on unidentified boundary track image data according to the boundary fracture node data and unidentified boundary offset distance data to generate track boundary line interpolation data;
and carrying out line filling smoothing treatment on the line filling interpolation data of the track boundary to generate line filling data of the track boundary.
The invention analyzes the track image data and extracts the track image without identifying the boundary, ensures the concentrated attention to the difficulty of identifying the track boundary, provides targeted input data for accurate line supplementing of the boundary, and effectively solves the problem of identifying the track boundary in the open circuit area and the crossroad area. The track boundary data in the image analysis queue is used for carrying out boundary tracking analysis, so that track boundary trend without identifying the boundary periphery can be accurately determined, important reference information is provided for the line supplementing process, line supplementing accuracy is improved, and line supplementing operation is ensured to accord with the trend of an actual track boundary. The specific position of the track boundary interruption can be clearly identified through the identification processing of the boundary fracture node, which is important for the subsequent boundary line filling operation, and the clear fracture node position is helpful for more accurately filling lines and reduces the possibility of error line filling. The offset distance of the unidentified boundary is analyzed, the boundary offset to be considered in the line supplementing operation can be determined, and the analysis ensures accurate butt joint of the line supplementing interpolation, so that the line supplementing operation can be closer to the actual track boundary. And (3) carrying out line supplementing interpolation processing on the unidentified boundary based on the boundary fracture node data and the unidentified boundary offset distance data, so that the track boundary can be effectively reconstructed, and especially under the condition that boundary information is incomplete or partially lost. Such patch interpolation ensures the continuity of the track boundary, providing complete boundary information for accurate navigation of the wheelbarrow. The track boundary data after the line is supplemented is subjected to smooth processing, so that acute angles and unnatural curves possibly generated in the line supplementing process can be eliminated, the track boundary is smoother and more natural, the visual quality of the track boundary is improved, and a more accurate and stable foundation is provided for path planning of the wheelbarrow.
In the embodiment of the invention, the image segmentation technology, such as threshold processing, is used for extracting the data of the track image of the unidentified boundary area to generate unidentified boundary track image data, including the track image data without boundaries such as the open circuit area, the crossroad area and the like. And applying a boundary tracking algorithm to unidentified boundary track image data based on track boundary data corresponding to the track image data in the image analysis queue, such as a contour-based tracking method, determining the trend and the direction of the track boundary data corresponding to the track image data in the image analysis queue, generating boundary tracking data according to the result of boundary tracking, recording the trend information of the boundary, and providing a basis for subsequent line supplementing processing. And analyzing the boundary tracking data, identifying broken nodes in the track boundary, namely, the positions of sudden breaks of the boundary, marking and recording the identified broken nodes of the boundary, and generating the broken nodes of the boundary. For each boundary fracture node, calculating the distance from the boundary fracture node to the nearest identified boundary, recording the offset distance of each fracture node as the offset distance, wherein the offset distance is used for judging how the patch should extend to most naturally connect the fractured boundaries, and generating unidentified boundary offset distance data. And selecting a proper interpolation algorithm according to the overall trend and the local characteristic of the track boundary, wherein linear interpolation is suitable for a region with a straight boundary line, spline interpolation (such as Bezier curve or B spline curve) is suitable for a region with a meandering boundary line, and executing the selected interpolation algorithm according to boundary fracture node data and unidentified boundary offset distance data to generate a complementary line. When the line is repaired, the trend of the adjacent identified boundary is considered to ensure that the line is naturally fused into the whole track boundary, the result after the line is repaired is recorded, and the track boundary line repairing interpolation data is generated, including the information of the starting point, the ending point, the trend and the like of the line repairing. And selecting a proper smoothing algorithm, such as Gaussian blur or moving average, to smooth the interpolation data of the patch, ensuring seamless connection of the patch and the original track boundary, eliminating any possible acute angle or unnatural transition, generating the patch data of the track boundary, and verifying the naturalness and accuracy of the patch. The line filling and smoothing process is manually adjusted or re-performed as necessary to achieve the best effect.
Preferably, step S4 comprises the steps of:
Step S41, performing preliminary track center line fitting processing on track boundary center line data and track boundary patch center line data to generate preliminary track center line data;
S42, performing obstacle characteristic node analysis according to the binarized racetrack image data to generate obstacle characteristic node data;
s43, performing detour path node calculation processing of the monocycle by using a graph theory algorithm and obstacle characteristic node data to generate detour path nodes;
S44, performing detour path curvature analysis processing based on the detour path nodes to generate detour path curvature data;
Step S45, carrying out smooth adjustment processing on the detour path according to the detour path curvature data and the detour path node data to generate smooth detour path data;
step S46, performing obstacle region central line optimization adjustment processing on the preliminary track central line data according to the smooth detour path data to generate track central line data;
and step S47, performing single wheel lane road planning processing according to the track central line data to generate planning single wheel lane road data.
The invention generates a rough track center line through preliminary fitting treatment of track boundary and line supplementing center line data, provides a foundation for follow-up finer path planning and optimization, ensures continuity and integrity of the track center line, and lays a foundation for stable running of the wheelbarrow. By analyzing obstacle characteristic nodes in the binarized racetrack image data, the positions of obstacles on the racetrack are accurately identified, key information is provided for the establishment of obstacle detouring strategies, the wheelbarrow is facilitated to effectively avoid obstacles, and the driving safety is improved. The round path nodes are calculated by combining the graph theory algorithm with the obstacle characteristic node data, a round strategy based on actual obstacle distribution is provided for the wheelbarrow, the accuracy and the practicability of path planning are improved, and the wheelbarrow is ensured to be capable of efficiently navigating in a complex environment. The curvature of the detour path is analyzed, the feasibility and the safety of the path are evaluated, the stability of the wheelbarrow when detouring around obstacles is ensured, and the curvature analysis result is crucial to the adjustment of the path design so as to avoid unstable running caused by the too-sharp path. The detour path is smoothly regulated according to the curvature data and the node information, a smoother detour path is generated, abrupt direction changes in the running process of the monocycle are reduced, the driving comfort and safety are improved, and the detour path is optimized, so that the real running requirement is met. The part of the track center line in the obstacle area is optimally adjusted by combining the smooth detour path data, so that the accuracy of the track center line is further refined, the wheelbarrow can be ensured to more accurately follow the track center line to detour the obstacle, and the quality of path planning and the running efficiency of the wheelbarrow are improved. The method integrates all the optimization and analysis results into an actual road plan, provides an optimization path considering obstacles, track boundaries and driving safety, and ensures that the wheelbarrow can realize efficient and safe autonomous navigation in a changeable environment.
As an example of the present invention, referring to fig. 4, a detailed implementation step flow diagram of step S4 in fig. 1 is shown, where step S4 includes:
Step S41, performing preliminary track center line fitting processing on track boundary center line data and track boundary patch center line data to generate preliminary track center line data;
In the embodiment of the invention, the track boundary central line data and the track boundary line-supplementing central line data are integrated, so that all track information is ensured to be considered, the track boundary central line data and the track boundary line-supplementing central line data are subjected to data sorting by considering the time sequences of different images, a proper fitting algorithm such as polynomial fitting or spline curve fitting is selected, and central line fitting is performed according to the overall trend and local change of the track, so as to generate preliminary track central line data.
S42, performing obstacle characteristic node analysis according to the binarized racetrack image data to generate obstacle characteristic node data;
In the embodiment of the invention, the binary track image data is analyzed, the obstacle characteristics in the image, such as the edge and the shape of an object, are identified, the obstacle data is identified through the pre-designed obstacle identification characteristics, key characteristic nodes of the obstacle, such as boundary points or salient points of the obstacle, are extracted, the nodes are used for subsequent path node calculation, the extracted obstacle characteristic nodes are recorded, and the obstacle characteristic node data is generated.
S43, performing detour path node calculation processing of the monocycle by using a graph theory algorithm and obstacle characteristic node data to generate detour path nodes;
in an embodiment of the present invention, graph theory algorithms, such as Dijkstra or Dijkstra, are utilized The searching algorithm calculates the detour path node of the monocycle based on the track central line data and the obstacle characteristic node data, determines key nodes on the detour path, including a starting point, an ending point and a turning point, ensures that the path detours the obstacle and optimizes the driving distance and time as much as possible, records the calculated detour path node, generates detour path node data, and provides a basis for path curvature analysis.
S44, performing detour path curvature analysis processing based on the detour path nodes to generate detour path curvature data;
In the embodiment of the invention, the trend of the path is determined according to the nodes of the detour path, a curvature analysis method, such as a method based on angle change or curve fitting, is applied, the curvature of each section on the detour path is calculated, the curvature data of the detour path is generated according to the result of curvature analysis, the bending degree of the path is provided, and the steering control and the speed adjustment of the wheelbarrow are reasonably adjusted, so that the wheelbarrow can pass through the obstacle at the optimal turning angle.
Step S45, carrying out smooth adjustment processing on the detour path according to the detour path curvature data and the detour path node data to generate smooth detour path data;
In the embodiment of the invention, the curvature data of the detour path and the node data of the detour path are comprehensively considered, sharp turns or unnecessary turning back in the path are identified, and a proper path smoothing algorithm, such as Bezier curve technology, is selected for adjusting the path which is over-bent or turned back in the detour, so that the path is smoother and more natural. And regulating the bypass path by using a selected smoothing algorithm, ensuring the natural transition of the path at the steering point, avoiding the abrupt change angle and generating smooth bypass path data.
Step S46, performing obstacle region central line optimization adjustment processing on the preliminary track central line data according to the smooth detour path data to generate track central line data;
In the embodiment of the invention, a central line offset area caused by obstacles in a track is identified according to smooth detour path data, a central line optimization strategy is designed, the width of the track, the position and the size of the obstacles and the running characteristics of a monocycle are considered, the central line of the track is subjected to necessary adjustment, the smooth detour path data and the preliminary central line data of the track are connected at the connection position, further central line adjustment is optimized, the transition smoothness of the central line of the track in the obstacle area is ensured and the actual running requirement is met, the central line data of the track after adjustment and optimization is recorded, and the optimal detour path reflecting all the obstacle areas is ensured.
And step S47, performing single wheel lane road planning processing according to the track central line data to generate planning single wheel lane road data.
In the embodiment of the invention, according to the optimized track central line data, preparing to carry out final road planning, taking the dynamics characteristics and the safe driving requirements of the monocycles into consideration, and applying a road planning algorithm, such asThe search algorithm or dynamic programming algorithm generates an actual travel route of the wheelbarrow based on the track centerline data for use in planning the wheelbarrow road and planning the wheelbarrow road including specific coordinates, turning points, speeds, etc. of the travel route. And (3) evaluating the practicability and safety of the planning result, adjusting the planning route if necessary, ensuring that the final road planning result meets the actual driving requirement, and simultaneously considering the efficiency and safety.
Preferably, step S44 includes the steps of:
acquiring wheelbarrow size parameters and wheelbarrow test dynamics data;
Carrying out detour path node optimization processing according to preset wheelbarrow safety margin data and wheelbarrow size parameters, and generating detour path optimization node data;
carrying out barrier region dynamics data extraction on the monocycle test dynamics data according to the barrier characteristic node data to generate barrier region dynamics data;
And carrying out detour path curvature analysis processing according to the obstacle region dynamics data and the detour path node data, and generating detour path curvature data.
The method for calculating the detour path by using the monocycles provides basic information for detour path planning by collecting the size and the dynamics data of the monocycles, ensures the practicability and the feasibility of path planning, and is crucial for accurately calculating the detour path by knowing the physical size and the dynamics performance of the monocycles, especially in the scene requiring fine operation to avoid obstacles. The optimization processing of the path nodes is carried out by considering the safety margin and the size parameter of the wheelbarrow, so that the safety and the stability of the wheelbarrow when the wheelbarrow bypasses obstacles are ensured, and the collision or dangerous condition of the bypass path is avoided. The dynamic data aiming at the specific obstacle area is extracted, so that the path planning can be more finely adapted to the specific characteristics of different obstacles, the specific dynamic data extraction is favorable for more accurate power and steering adjustment when the obstacle is bypassed, and the bypassing efficiency and the safety are improved. By comprehensively considering the dynamics data of the obstacle area and the nodes of the detour path, the curvature analysis processing can generate a detour path which accords with the dynamics performance of the monocycle and adapts to the terrain characteristics, so that the space layout of the path is considered, the dynamic response of the monocycle is considered, and the smoothness and the performability of the detour action are ensured.
In the embodiment of the invention, the size parameters of the monocycle and the dynamic data of the monocycle test are obtained, for example, the size parameters of the monocycle, including length, width, height and the like, are recorded in advance, the parameters directly influence the capability of the monocycle to bypass obstacles, and the dynamic data of each node, including acceleration, braking distance, steering radius and the like, are collected in the dynamic test before the monocycle. According to the running environment and actual conditions of the wheelbarrow, a safety margin is designed in advance, the safety margin considers the overstretched safety margin of the wheelbarrow in different speed states, so that enough safety distance is ensured between the wheelbarrow and the obstacle, and the wheelbarrow size parameter and the safety margin data are utilized to optimize nodes of a detour path, so that the feasibility and the safety of the path are ensured. And extracting the dynamics data of the obstacle region from the test dynamics data of the monocycle according to the obstacle characteristic node data, analyzing the obstacle characteristic node data, determining the position, the size and the possible influence of the obstacle on the driving path of the monocycle, and extracting the dynamics data related to the feature and the position of the obstacle from the test dynamics data of the monocycle so as to generate the dynamics data of the obstacle region. According to the data of steering capability, braking distance, acceleration performance and the like of the monocycle in the dynamics data of the obstacle area at different speeds, analyzing the influence of the position of the obstacle on the driving of the monocycle, further obtaining the node needing to be steered, according to the node needing to be steered, applying a curvature calculation formula, such as a circular arc curvature formula, to calculate the curvature radius of each steering node, and according to each steering node and the corresponding curvature radius, carrying out detour path curvature calculation, and finally obtaining detour path curvature data.
Preferably, the wheelbarrow driving optimization operation includes a wheelbarrow driving operation and a wheelbarrow driving iterative optimization operation, and step S5 includes the steps of:
s51, performing image width center data analysis processing according to the racetrack image data to generate image width center data;
step S52, carrying out path offset distance analysis according to the image width center data and the planned single wheel lane road data to generate path offset distance data;
Step S53, designing wheelbarrow control parameters according to the path offset distance data and the planning wheelbarrow road data so as to obtain wheelbarrow control parameters;
Step S54, executing a wheelbarrow driving operation based on the wheelbarrow control parameters;
step S55, driving feedback data acquisition is carried out on driving operation of the wheelbarrow, and driving feedback data are generated;
Step S56, establishing a self-adaptive adjustment mapping relation of driving feedback data and monocycle control parameters based on a preset reinforcement learning model so as to generate an initial control parameter adjustment model;
s57, performing model training treatment on the initial control parameter adjustment model according to the driving feedback data to generate a control parameter adjustment model;
And S58, carrying out control parameter correction optimization processing on the monocycle control parameters based on the control parameter adjustment model, generating optimized monocycle control parameters, and feeding back the optimized monocycle control parameters to the step S54 to execute monocycle driving iterative optimization operation.
The invention provides accurate reference for determining the central line of the track by analyzing the width center of the track image, is the basis for carrying out accurate path deviation analysis, and is beneficial to accurately positioning the actual center of the track by knowing the center point of the image so as to optimize the running track of the wheelbarrow. The difference between the image width center data and the planned wheelbarrow road data is analyzed, the path offset distance is accurately calculated, and the accurate offset distance data enables the control system to correct the driving direction in time and keep the wheelbarrow to drive along the optimal path. The accurate design of the control parameters of the wheelbarrow is carried out based on the path offset distance data, so that the wheelbarrow can pass through the track at the optimal gesture and speed, the running stability and efficiency of the wheelbarrow are directly influenced, and the key for realizing efficient driving is realized. The actual driving operation of the wheelbarrow is performed according to the carefully designed control parameters, so that the wheelbarrow can accurately and stably run according to the preset path. And collecting feedback data generated in the driving operation, providing basis for subsequent control parameter adjustment and optimization, wherein the feedback data is important for understanding the performance of the wheelbarrow in actual operation, and is helpful for identifying and solving the existing problems. The self-adaptive adjustment mapping relation between the driving feedback data and the control parameters is established by utilizing a pre-designed reinforcement learning model, so that the control system can learn and adapt to environmental changes, and the control strategy is automatically adjusted to optimize the running performance. And through model training treatment, the control parameter adjustment model is continuously optimized, and the accuracy and the adaptability of the model are improved. The trained model can more accurately adjust control parameters according to feedback data, and the running performance of the wheelbarrow is improved. And (3) based on the optimized control parameter adjustment model, carrying out correction optimization processing on the control parameters, further improving the driving performance of the wheelbarrow, and continuously optimizing the driving path and speed of the wheelbarrow through iterative optimization, so that higher driving efficiency and safety are finally achieved.
As an example of the present invention, referring to fig. 5, a detailed implementation step flow diagram of step S5 in fig. 1 is shown, where step S5 includes:
s51, performing image width center data analysis processing according to the racetrack image data to generate image width center data;
in the embodiment of the invention, an image processing technology is applied to extract central line data of image width from track image data to generate central data of image width.
Step S52, carrying out path offset distance analysis according to the image width center data and the planned single wheel lane road data to generate path offset distance data;
in the embodiment of the invention, path offset distance calculation is performed according to the image width center data and the planned single wheel lane road data, and the cross section offset between the image width center and the planned single wheel lane road data at the same moving distance position is calculated to generate the path offset distance data.
Step S53, designing wheelbarrow control parameters according to the path offset distance data and the planning wheelbarrow road data so as to obtain wheelbarrow control parameters;
In the embodiment of the invention, correction offset control parameters in the wheelbarrow control parameters are analyzed according to the path offset distance data and are used for correcting the offset error of the wheelbarrow driving path, and parameters such as speed, steering and the like of the wheelbarrow in a planned road are set by planning wheelbarrow road data, and the control parameters are integrated to obtain the wheelbarrow control parameters.
Step S54, executing a wheelbarrow driving operation based on the wheelbarrow control parameters;
In the embodiment of the invention, the control parameters are transmitted to the terminal based on the monocycle, and the terminal controls the corresponding monocycle device through the monocycle control parameters, so that the monocycle driving operation is executed.
Step S55, driving feedback data acquisition is carried out on driving operation of the wheelbarrow, and driving feedback data are generated;
In the embodiment of the invention, feedback data including speed change, steering angle adjustment, path offset and the like are acquired in real time in the execution process of the driving operation of the wheelbarrow, and driving feedback data are generated.
Step S56, establishing a self-adaptive adjustment mapping relation of driving feedback data and monocycle control parameters based on a preset reinforcement learning model so as to generate an initial control parameter adjustment model;
In an embodiment of the present invention, a suitable reinforcement learning algorithm, such as Q-learning, deep Q-Network (DQN) or other algorithms suitable for continuous motion space, such as Proximal Policy Optimization (PPO), is selected as a basis.
The state space is defined as the key parameters of the position, speed, path offset distance and the like of the wheelbarrow relative to the track, and the action space is defined as the adjustable wheelbarrow control parameters such as steering angle, acceleration or deceleration.
The reward function is designed to evaluate the driving performance of the wheelbarrow, and should reflect the accuracy of path tracking, smoothness and safety of driving, initialize the reinforcement learning model using preset parameters and structures, and prepare to accept training data to generate the initial control parameter adjustment model.
S57, performing model training treatment on the initial control parameter adjustment model according to the driving feedback data to generate a control parameter adjustment model;
in the embodiment of the invention, the driving feedback data of the monocycle is collected through the simulation environment or the actual driving test, including state conversion, actions taken and corresponding rewards, a reasonable training period is set according to the data quantity and the model complexity, so that the model is ensured to have enough learning process to improve the performance, the reinforcement learning model is trained under supervision, model parameters are continuously updated to maximize accumulated rewards, the performance of the training model is periodically evaluated, the learning process is ensured to be effective, and the model can accurately predict the optimal adjustment mode of the control parameters.
And S58, carrying out control parameter correction optimization processing on the monocycle control parameters based on the control parameter adjustment model, generating optimized monocycle control parameters, and feeding back the optimized monocycle control parameters to the step S54 to execute monocycle driving iterative optimization operation.
In the embodiment of the invention, a trained model is used, optimal control parameter adjustment is predicted according to the current state of the monocycle so as to improve running performance, the optimal control parameter predicted by the model is applied to the monocycle, a driving system is adjusted in real time so as to correct path deviation or improve running efficiency, feedback data obtained in actual running is input into the model again, continuous learning and optimization are carried out, a closed loop is formed in the process, and the model can be continuously adapted to new running environment and challenges, so that driving operation of the monocycle is continuously optimized.
The present disclosure provides an image processing-based single-wheel lane road searching and planning system for executing the image processing-based single-wheel lane road searching and planning method, which includes:
the track image data acquisition module is used for acquiring track image data by using monitoring equipment arranged in the wheelbarrow to generate track image data;
The track boundary analysis module is used for carrying out dynamic binarization threshold analysis processing on the track image data to generate dynamic binarization threshold data; performing track boundary data analysis processing on the track image data according to the dynamic binarization threshold data to generate track boundary data;
the track boundary data acquisition module is used for analyzing the track boundary central line data according to the track boundary data when the track boundary data are detected, so as to generate track boundary central line data; when the track boundary data is not detected, carrying out track boundary line supplementing processing according to the track image data to generate track boundary line supplementing data; analyzing the track boundary line-supplementing central line data according to the track boundary line-supplementing data to generate track boundary line-supplementing central line data;
The single wheel lane road planning module is used for performing track center line fitting optimization processing according to track boundary center line data and track line supplementing boundary center line data to generate track center line data; carrying out single wheel lane route planning processing according to the track central line data to generate planned single wheel lane route data;
The wheelbarrow driving module is used for carrying out path offset distance analysis according to the racetrack image data and the planned wheelbarrow road data to generate path offset distance data; designing wheelbarrow control parameters according to the path offset distance data and the planned wheelbarrow road data to obtain wheelbarrow control parameters; and executing the wheelbarrow driving optimization operation based on the wheelbarrow control parameters.
The method has the advantages that the method can accurately identify and process the driving paths of the wheelbarrow in the wheelbarrow track open-circuit area, the obstacle area and the like, and plan the whole driving paths of the wheelbarrow by carrying out line supplementing processing on the unidentified road boundaries, so that the efficiency of the wheelbarrow road searching and planning is high, the driving path deviation of the wheelbarrow can be effectively corrected, the wheelbarrow control parameters are continuously optimized and adjusted through the reinforcement learning model, the driving path of the wheelbarrow accords with the planning path, and the accuracy of the subsequent wheelbarrow road searching and planning is higher.
The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.
The foregoing is only a specific embodiment of the invention to enable those skilled in the art to understand or practice the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (7)
1. The image processing-based single wheel lane route searching and planning method is characterized by comprising the following steps of:
S1, performing track image data acquisition by using monitoring equipment built in a wheelbarrow to generate track image data;
Step S2, including:
Step S21, data real-time reading is carried out on the track image data to generate read track image data, wherein the read track image data comprises read image sequence data and real-time track image data;
S22, performing gray level image conversion processing on the real-time track image data to generate gray level track image data;
s23, establishing an image analysis queue according to the read image sequence data, and transmitting the gray-scale racetrack image data to the image analysis queue to generate queue racetrack image data;
S24, carrying out dynamic binarization threshold analysis on the queue track image data to generate dynamic binarization threshold data;
s25, performing image binarization traversal on the queue track image data according to the dynamic binarization threshold value data to generate binarization track image data;
s26, performing track boundary detection processing on the binarized track image data by utilizing edge detection to generate track boundary data;
step S3, including:
S31, when track boundary data is detected, track boundary central line data analysis is carried out according to the track boundary data, and track boundary central line data are generated;
s32, when the track boundary data is not detected, performing track boundary line supplementing processing according to the track image data to generate track boundary line supplementing data; analyzing the track boundary line-supplementing central line data according to the track boundary line-supplementing data to generate track boundary line-supplementing central line data;
Wherein, the step S32 of performing the track boundary line filling process according to the track image data includes: performing unidentified boundary racetrack image extraction processing according to racetrack image data to generate unidentified boundary racetrack image data; performing boundary tracking data analysis processing on unidentified boundary track image data based on track boundary data corresponding to track image data in an image analysis queue to generate boundary tracking data; performing boundary fracture node identification processing on unidentified boundary racetrack image data based on the boundary tracking data to generate boundary fracture node data; performing unrecognized boundary offset distance analysis processing according to the boundary fracture node data to generate unrecognized boundary offset distance data; performing track boundary line interpolation processing on unidentified boundary track image data according to the boundary fracture node data and unidentified boundary offset distance data to generate track boundary line interpolation data; performing line-filling smoothing on the line-filling interpolation data of the track boundary to generate line-filling data of the track boundary;
s4, performing track center line fitting optimization processing according to track boundary center line data and track patch boundary center line data to generate track center line data; carrying out single wheel lane route planning processing according to the track central line data to generate planned single wheel lane route data;
Step S5, including:
s51, performing image width center data analysis processing according to the racetrack image data to generate image width center data;
step S52, carrying out path offset distance analysis according to the image width center data and the planned single wheel lane road data to generate path offset distance data;
Step S53, designing wheelbarrow control parameters according to the path offset distance data and the planning wheelbarrow road data so as to obtain wheelbarrow control parameters;
Step S54, executing a wheelbarrow driving operation based on the wheelbarrow control parameters;
step S55, driving feedback data acquisition is carried out on driving operation of the wheelbarrow, and driving feedback data are generated;
Step S56, establishing a self-adaptive adjustment mapping relation of driving feedback data and monocycle control parameters based on a preset reinforcement learning model so as to generate an initial control parameter adjustment model;
s57, performing model training treatment on the initial control parameter adjustment model according to the driving feedback data to generate a control parameter adjustment model;
And S58, carrying out control parameter correction optimization processing on the monocycle control parameters based on the control parameter adjustment model, generating optimized monocycle control parameters, and feeding back the optimized monocycle control parameters to the step S54 to execute monocycle driving iterative optimization operation.
2. The image processing-based solitary lane route planning method as claimed in claim 1 wherein step S1 comprises the steps of:
s11, acquiring an initial track image by using monitoring equipment built in a wheelbarrow, and generating initial track image data;
s12, performing spectral image analysis processing on the initial racetrack image to generate spectral image data;
s13, performing noise spectrum calculation processing on the spectral image data by using a noise spectrum calculation formula to generate noise spectrum data;
S14, carrying out correction spectrum data analysis according to the noise spectrum data to generate correction spectrum data;
Step S15, performing feature point selection processing on the spectral image data to generate spectral image feature point data;
s16, analyzing dynamic jitter data of the spectral image according to the characteristic point position data of the spectral image to generate dynamic jitter data;
S17, analyzing geometrical calibration data of the spectral image according to the dynamic jitter data to generate geometrical calibration data of the spectral image;
And S18, performing track image correction and reconstruction processing on the initial track image data according to the corrected spectrum data and the spectrum image geometric calibration data to generate track image data.
3. The method for planning road search of a monowheel road based on image processing as set forth in claim 2, wherein the noise spectrum calculation formula in step S13 is as follows:
;
In the method, in the process of the invention, Expressed as wavelength/>Noise spectrum data at time,/>Expressed as spectral weight information,/>Expressed as wavelength/>Time-base raw spectral signal intensity,/>Expressed as/>Environmental impact spectrum intensity signal,/>Expressed as wavelength/>Spectral reflectance at time,/>Expressed as spectral angle of incidence,/>Reference spectral signal intensity expressed as feature point location,/>Expressed as the actual spectral signal intensity of the feature point.
4. The image processing-based solitary lane route planning method as claimed in claim 1 wherein step S24 comprises the steps of:
performing extreme gray value analysis on the queue race track image data to respectively generate minimum gray value data and maximum gray value data;
when the queue track image data is updated, respectively carrying out extreme value gray value updating on the minimum gray value data and the maximum gray value data according to the updated queue track image data so as to obtain the updated minimum gray value data and the updated maximum gray value data;
And establishing a dynamic binarization threshold according to the updated minimum gray value data and the updated maximum gray value data to generate dynamic binarization threshold data.
5. The image processing-based solitary lane route planning method as claimed in claim 1 wherein step S4 comprises the steps of:
Step S41, performing preliminary track center line fitting processing on track boundary center line data and track boundary patch center line data to generate preliminary track center line data;
S42, performing obstacle characteristic node analysis according to the binarized racetrack image data to generate obstacle characteristic node data;
s43, performing detour path node calculation processing of the monocycle by using a graph theory algorithm and obstacle characteristic node data to generate detour path nodes;
S44, performing detour path curvature analysis processing based on the detour path nodes to generate detour path curvature data;
Step S45, carrying out smooth adjustment processing on the detour path according to the detour path curvature data and the detour path node data to generate smooth detour path data;
step S46, performing obstacle region central line optimization adjustment processing on the preliminary track central line data according to the smooth detour path data to generate track central line data;
and step S47, performing single wheel lane road planning processing according to the track central line data to generate planning single wheel lane road data.
6. The method of image processing-based solitary lane route planning as claimed in claim 5 wherein step S44 comprises the steps of:
acquiring wheelbarrow size parameters and wheelbarrow test dynamics data;
Carrying out detour path node optimization processing according to preset wheelbarrow safety margin data and wheelbarrow size parameters, and generating detour path optimization node data;
carrying out barrier region dynamics data extraction on the monocycle test dynamics data according to the barrier characteristic node data to generate barrier region dynamics data;
And carrying out detour path curvature analysis processing according to the obstacle region dynamics data and the detour path node data, and generating detour path curvature data.
7. An image processing-based single wheel lane road search planning system for performing the image processing-based single wheel lane road search planning method according to any one of claims 1 to 6, comprising:
the track image data acquisition module is used for acquiring track image data by using monitoring equipment arranged in the wheelbarrow to generate track image data;
The track boundary analysis module is used for reading the track image data in real time to generate read track image data, wherein the read track image data comprises read image sequence data and real-time track image data; carrying out gray level image conversion processing on the real-time race track image data to generate gray level race track image data; establishing an image analysis queue according to the read image sequence data, and transmitting the gray level racetrack image data to the image analysis queue to generate queue racetrack image data; performing dynamic binarization threshold analysis on the queue track image data to generate dynamic binarization threshold data; performing image binarization traversal on the queue track image data according to the dynamic binarization threshold value data to generate binarization track image data; performing track boundary detection processing on the binarized track image data by utilizing edge detection to generate track boundary data;
The track boundary data acquisition module is used for analyzing the track boundary central line data according to the track boundary data when the track boundary data are detected, so as to generate track boundary central line data; when the track boundary data is not detected, carrying out track boundary line supplementing processing according to the track image data to generate track boundary line supplementing data; analyzing the track boundary line-supplementing central line data according to the track boundary line-supplementing data to generate track boundary line-supplementing central line data; the track boundary line supplementing process according to the track image data comprises the following steps: performing unidentified boundary racetrack image extraction processing according to racetrack image data to generate unidentified boundary racetrack image data; performing boundary tracking data analysis processing on unidentified boundary track image data based on track boundary data corresponding to track image data in an image analysis queue to generate boundary tracking data; performing boundary fracture node identification processing on unidentified boundary racetrack image data based on the boundary tracking data to generate boundary fracture node data; performing unrecognized boundary offset distance analysis processing according to the boundary fracture node data to generate unrecognized boundary offset distance data; performing track boundary line interpolation processing on unidentified boundary track image data according to the boundary fracture node data and unidentified boundary offset distance data to generate track boundary line interpolation data; performing line-filling smoothing on the line-filling interpolation data of the track boundary to generate line-filling data of the track boundary;
The single wheel lane road planning module is used for performing track center line fitting optimization processing according to track boundary center line data and track line supplementing boundary center line data to generate track center line data; carrying out single wheel lane route planning processing according to the track central line data to generate planned single wheel lane route data;
The monocycle driving module is used for carrying out image width center data analysis processing according to the racetrack image data to generate image width center data; carrying out path offset distance analysis according to the image width center data and the planned single wheel lane road data to generate path offset distance data; designing wheelbarrow control parameters according to the path offset distance data and the planned wheelbarrow road data to obtain wheelbarrow control parameters; executing a wheelbarrow driving operation based on the wheelbarrow control parameters; driving feedback data acquisition is carried out on driving operation of the monocycle, and driving feedback data are generated; establishing a self-adaptive adjustment mapping relation of driving feedback data and monocycle control parameters based on a preset reinforcement learning model to generate an initial control parameter adjustment model; performing model training treatment on the initial control parameter adjustment model according to the driving feedback data to generate a control parameter adjustment model; and (3) carrying out control parameter correction optimization processing on the monocycle control parameters based on the control parameter adjustment model, generating optimized monocycle control parameters, and feeding back the optimized monocycle control parameters to the step S54 to execute monocycle driving iterative optimization operation.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410424071.8A CN118010060B (en) | 2024-04-10 | 2024-04-10 | Image processing-based single wheel lane route searching and planning method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410424071.8A CN118010060B (en) | 2024-04-10 | 2024-04-10 | Image processing-based single wheel lane route searching and planning method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN118010060A CN118010060A (en) | 2024-05-10 |
CN118010060B true CN118010060B (en) | 2024-05-31 |
Family
ID=90959785
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202410424071.8A Active CN118010060B (en) | 2024-04-10 | 2024-04-10 | Image processing-based single wheel lane route searching and planning method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN118010060B (en) |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102662402A (en) * | 2012-06-05 | 2012-09-12 | 北京理工大学 | Intelligent camera tracking car model for racing tracks |
CN106055745A (en) * | 2016-05-20 | 2016-10-26 | 浙江大学 | Method of establishing linear CCD four-wheeler simulation model based on MATLAB |
CN110132288A (en) * | 2019-05-08 | 2019-08-16 | 南京信息工程大学 | A kind of minicar vision navigation method on wide road surface |
CN111731324A (en) * | 2020-05-29 | 2020-10-02 | 徐帅 | Control method and system for guiding AGV intelligent vehicle based on vision |
CN112101128A (en) * | 2020-08-21 | 2020-12-18 | 东南大学 | Unmanned formula racing car perception planning method based on multi-sensor information fusion |
CN112578673A (en) * | 2020-12-25 | 2021-03-30 | 浙江科技学院 | Perception decision and tracking control method for multi-sensor fusion of formula-free racing car |
CN112731925A (en) * | 2020-12-21 | 2021-04-30 | 浙江科技学院 | Conical barrel identification and path planning and control method for unmanned formula racing car |
CN114332647A (en) * | 2021-12-31 | 2022-04-12 | 合肥工业大学 | River channel boundary detection and tracking method and system for unmanned ship |
CN114863387A (en) * | 2022-03-31 | 2022-08-05 | 中山大学 | Tracking intelligent vehicle track identification and classification method |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11774571B2 (en) * | 2019-12-30 | 2023-10-03 | Wipro Limited | Method and system for navigating autonomous ground vehicle using radio signal and vision sensor |
-
2024
- 2024-04-10 CN CN202410424071.8A patent/CN118010060B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102662402A (en) * | 2012-06-05 | 2012-09-12 | 北京理工大学 | Intelligent camera tracking car model for racing tracks |
CN106055745A (en) * | 2016-05-20 | 2016-10-26 | 浙江大学 | Method of establishing linear CCD four-wheeler simulation model based on MATLAB |
CN110132288A (en) * | 2019-05-08 | 2019-08-16 | 南京信息工程大学 | A kind of minicar vision navigation method on wide road surface |
CN111731324A (en) * | 2020-05-29 | 2020-10-02 | 徐帅 | Control method and system for guiding AGV intelligent vehicle based on vision |
CN112101128A (en) * | 2020-08-21 | 2020-12-18 | 东南大学 | Unmanned formula racing car perception planning method based on multi-sensor information fusion |
CN112731925A (en) * | 2020-12-21 | 2021-04-30 | 浙江科技学院 | Conical barrel identification and path planning and control method for unmanned formula racing car |
CN112578673A (en) * | 2020-12-25 | 2021-03-30 | 浙江科技学院 | Perception decision and tracking control method for multi-sensor fusion of formula-free racing car |
CN114332647A (en) * | 2021-12-31 | 2022-04-12 | 合肥工业大学 | River channel boundary detection and tracking method and system for unmanned ship |
CN114863387A (en) * | 2022-03-31 | 2022-08-05 | 中山大学 | Tracking intelligent vehicle track identification and classification method |
Non-Patent Citations (1)
Title |
---|
hao zhu,等.A Path Planning Algorithm Based on Fusing Lane and Obstacle Map.《2014 IEEE 17th International Conference on Intelligent Transportation Systems (ITSC)》.2014,全文. * |
Also Published As
Publication number | Publication date |
---|---|
CN118010060A (en) | 2024-05-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11482014B2 (en) | 3D auto-labeling with structural and physical constraints | |
Mu et al. | Lane detection based on object segmentation and piecewise fitting | |
Kim | Robust lane detection and tracking in challenging scenarios | |
US7894666B2 (en) | Image processing method and apparatus, image sensing apparatus, and program | |
US8280165B2 (en) | System and method for segmenting foreground and background in a video | |
JP4159794B2 (en) | Image processing apparatus and method | |
JP4157620B2 (en) | Moving object detection apparatus and method | |
CN110135302B (en) | Method, device, equipment and storage medium for training lane line recognition model | |
Engel et al. | Deep object tracking on dynamic occupancy grid maps using rnns | |
CN111428759A (en) | Data fusion method, electronic device and storage medium | |
JPH10307923A (en) | Face parts extraction device and face direction detection device | |
CN103413312A (en) | Video target tracking method based on neighborhood components analysis and scale space theory | |
CN115205636A (en) | Image target detection method, system, equipment and storage medium | |
CN116597122A (en) | Data labeling method, device, electronic equipment and storage medium | |
CN117746343B (en) | Personnel flow detection method and system based on contour map | |
Stević et al. | Vision-based extrapolation of road lane lines in controlled conditions | |
Sanberg et al. | Extending the stixel world with online self-supervised color modeling for road-versus-obstacle segmentation | |
CN118010060B (en) | Image processing-based single wheel lane route searching and planning method and system | |
JP2002175534A (en) | Method for detecting road white line | |
CN110837760A (en) | Target detection method, training method and device for target detection | |
JP4749879B2 (en) | Face discrimination method, apparatus, and program | |
JP2011170890A (en) | Face detecting method, face detection device, and program | |
Bhupathi et al. | An augmented sliding window technique to improve detection of curved lanes in autonomous vehicles | |
Sang et al. | A Robust Lane Detection Algorithm Adaptable to Challenging Weather Conditions | |
JP2007025766A (en) | Face detecting method, face detection device, and program |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |