CN115223128A - Road rut congestion detection method and system based on neural network - Google Patents
Road rut congestion detection method and system based on neural network Download PDFInfo
- Publication number
- CN115223128A CN115223128A CN202210955092.3A CN202210955092A CN115223128A CN 115223128 A CN115223128 A CN 115223128A CN 202210955092 A CN202210955092 A CN 202210955092A CN 115223128 A CN115223128 A CN 115223128A
- Authority
- CN
- China
- Prior art keywords
- road surface
- deformation
- type
- road
- laser line
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/16—Measuring arrangements characterised by the use of optical techniques for measuring the deformation in a solid, e.g. optical strain gauge
- G01B11/165—Measuring arrangements characterised by the use of optical techniques for measuring the deformation in a solid, e.g. optical strain gauge by means of a grating deformed by the object
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/95—Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/766—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using regression, e.g. by projecting features on hyperplanes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention relates to the technical field of road engineering, and particularly discloses a road rut congestion detection method and a system based on a neural network, wherein the method comprises the steps of collecting a road surface image containing a laser line based on a preset collection frequency; inputting the road surface image into a trained type recognition model to obtain a road surface type; performing regression fitting on the laser line in the road surface image according to a preset fitting algorithm to obtain a coordinate point of the laser line in the road surface image; carrying out deformation analysis on the laser line based on the coordinate points to determine deformation parameters; and generating an identification record according to the deformation parameters and the road surface type. Compared with manual detection, the method can quickly and intelligently analyze the road surface deformation diseases, and workers only need to concentrate on driving, so that the working efficiency of the workers is greatly improved, and the safety risk is reduced; compared with the existing heavy detection equipment, the equipment cost and the operation cost are greatly reduced.
Description
Technical Field
The invention relates to the technical field of road engineering, in particular to a road rut congestion detection method and system based on a neural network.
Background
With the rapid development of road traffic construction in China, china has built a huge municipal, high-speed, medium-low road network, and a large amount of road facilities bring great challenges to pavement disease detection and pavement management and maintenance. The traditional road detection mode is manual patrol, manual measurement and recording of found diseases, and has the problems of low speed, multiple mistakes and omissions, incapability of large-area and high-frequency detection and the like. The management and maintenance unit with better economic conditions uses heavy detection equipment to regularly detect the road health, but the equipment is expensive, the maintenance cost is high, a vehicle needs to be specially modified, the equipment quantity is limited, the equipment cannot be used in a large area and at a high frequency, and the equipment is generally only used for annual detection of roads and cannot guide daily management and maintenance.
Disclosure of Invention
The present invention is directed to a road rut congestion detection method and system based on a neural network, so as to solve the problems in the background art.
In order to achieve the purpose, the invention provides the following technical scheme:
a neural network-based road track congestion detection method, the method comprising:
collecting a pavement image containing laser lines based on a preset collection frequency; the laser line is emitted by a preset line laser emitter;
inputting the road surface image into a trained type recognition model to obtain a road surface type; the pavement types comprise asphalt pavement, cement pavement and sand stone pavement;
performing regression fitting on the laser line in the road surface image according to a preset fitting algorithm to obtain a coordinate point of the laser line in the road surface image;
carrying out deformation analysis on the laser line based on the coordinate points to determine deformation parameters; the deformation analysis comprises straight line correlation analysis and curvature analysis;
and generating an identification record according to the deformation parameters and the road surface type.
As a further scheme of the invention: the step of collecting the road surface image containing the laser line based on the preset collection frequency comprises the following steps:
acquiring the motion parameters of the vehicle in real time, and adjusting the acquisition frequency according to the motion parameters; the motion parameters comprise motion speed and motion acceleration; the motion speed and the motion acceleration are both vectors.
As a further scheme of the invention: the step of carrying out deformation analysis on the laser line based on the coordinate points and determining deformation parameters comprises the following steps:
performing straight line fitting on the coordinate points according to a least square method, and calculating a linear correlation coefficient;
comparing the linear correlation coefficient with a preset coefficient threshold;
when the linear correlation coefficient reaches a preset coefficient threshold value, judging that the laser line deformation is normal;
when the linear correlation coefficient is smaller than a preset coefficient threshold value, judging that the laser line deformation is abnormal, and determining the deformation type according to the fluctuation direction; the deformation types include depressions and protrusions.
As a further scheme of the invention: the step of generating an identification record according to the deformation parameter and the road surface type comprises the following steps:
inputting the road surface image containing the laser line into a trained facility recognition model, and recognizing a target object; the target comprises a well cover, a road foreign matter and a road deceleration strip;
when the target object is not identified, reading the deformation parameter and the road surface type, and generating a first identification record according to the deformation parameter and the road surface type;
and when the target object is identified, reading the deformation parameter and the road surface type, and generating a second identification record according to the deformation parameter and the road surface type.
As a further scheme of the invention: when the target object is not identified, the step of reading the deformation parameter and the road surface type and generating a first identification record according to the deformation parameter and the road surface type comprises the following steps:
when the target object is not identified, reading the deformation parameter and the road surface type;
when the pavement type is an asphalt pavement and the deformation type is a depression, calculating the length of the depression; when the length of the recess reaches a preset length threshold value, judging that the abnormal type is a track; when the length of the recess is smaller than a preset length threshold value, judging that the abnormal type is sink;
when the pavement type is the asphalt pavement and the deformation type is convex, judging that the abnormal type is a hug;
and when the road surface type is a cement road surface and the deformation type is a bulge, judging that the abnormal type is the bulge.
As a further scheme of the invention: when the target object is identified, the step of reading the deformation parameter and the road surface type and generating a second identification record according to the deformation parameter and the road surface type comprises the following steps:
when the identified target object is a well lid and the linear correlation coefficient exceeds a preset threshold value, judging that the abnormal type is the height difference of the well lid;
when the identified target object is a road foreign object and the linear correlation coefficient exceeds a preset threshold value, judging that the abnormal type is a dangerous foreign object;
and when the identified target object is a deceleration strip and the linear correlation coefficient exceeds a preset threshold value, judging that the abnormal type is a large-altitude-difference deceleration strip.
The technical scheme of the invention also provides a road rut congestion detecting system based on the neural network, and the system comprises:
the image acquisition module is used for acquiring a road surface image containing laser lines based on a preset acquisition frequency; the laser line is emitted by a preset line laser emitter;
the road surface type identification module is used for inputting the road surface image into a trained type identification model to obtain a road surface type; the pavement types comprise asphalt pavement, cement pavement and sand stone pavement;
the laser line identification module is used for performing regression fitting on the laser line in the road surface image according to a preset fitting algorithm to obtain a coordinate point of the laser line in the road surface image;
the parameter analysis module is used for carrying out deformation analysis on the laser line based on the coordinate points to determine deformation parameters; the deformation analysis comprises straight line correlation analysis and curvature analysis;
and the record generating module is used for generating an identification record according to the deformation parameter and the road surface type.
Compared with the prior art, the invention has the beneficial effects that: compared with manual detection, the method can quickly and intelligently analyze the road surface deformation diseases, and workers only need to concentrate on driving, so that the working efficiency of the workers is greatly improved, and the safety risk is reduced; compared with the existing heavy detection equipment, the equipment cost and the operation cost are greatly reduced.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention.
Fig. 1 is a device relationship diagram of a road track congestion detection method based on a neural network.
Fig. 2 is a flow chart of a road rut congestion detection method based on a neural network.
Fig. 3 is a schematic diagram of a road rut congestion detection system based on a neural network.
Fig. 4 is a schematic diagram of various installation combinations of the neural network-based road track congestion detection system.
Fig. 5 is a block diagram of the structure of a road rut congestion detection system based on a neural network.
Detailed Description
In order to make the technical problems, technical solutions and advantageous effects of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the invention.
Example 1
Fig. 1 shows an equipment relationship diagram of a road rut congestion detection method based on a neural network, and fig. 2 shows a flow chart of the road rut congestion detection method based on the neural network, in an embodiment of the present invention, a road rut congestion detection method based on the neural network includes:
collecting a pavement image containing laser lines based on a preset collection frequency; the laser line is emitted by a preset line laser emitter;
in an example of the technical scheme of the invention, hardware equipment for executing the method is an intelligent inspection vehicle, equipment such as a high-definition camera, a central industrial personal computer and a line laser emitter are installed on the inspection vehicle, relevant equipment is started, the high-definition camera acquires images according to instructions of the central industrial personal computer in the driving process of the inspection vehicle, a plurality of neural network algorithms deployed on the central industrial personal computer respectively identify the road surface type, coordinate points [ u, v ] of laser lines in a fitting image, identify whether well covers or road surface foreign bodies appear near the laser lines, perform linear and curvature analysis on the laser line fitting coordinate points [ u, v ] by using a linear discrimination algorithm, judge whether deformation type diseases appear on the road surface or whether well covers, foreign bodies and deceleration zones with large deformation appear according to an identification result, and store the identification result locally and transmit the identification result back to a cloud end through a mobile network.
Wherein, install equipment such as high definition camera, central industrial computer, line laser emitter on the inspection car, equipment fixing position is as shown in fig. 3 and fig. 4.
(1) The high-definition camera is used for collecting road surface and laser line images. The effective pixels of the high-definition camera are more than or equal to 200 ten thousand pixels, a central industrial personal computer or a vehicle-mounted power supply can be used, and the high-definition camera can continuously acquire high-definition images in the process of inspection of the advancing vehicle.
(2) The line laser transmitter can emit a straight line visible light, and the length of the light line can cover the width of a lane. The power of the laser transmitter is more than or equal to 10mW, and a central industrial personal computer or vehicle-mounted power supply can be used.
(3) In order to make the change of the laser line at the road surface deformation position more obvious in the collected image, the included angle formed by the laser emitter, the road surface projection point and the camera is required to be ensured to be as large as possible, and the relative positions of the laser emitter and the high-definition camera mounted on the vehicle can be in various combination modes, as shown in fig. 4, but not limited to fig. 4.
Taking fig. 4 (a) as an example, the high-definition camera is installed at the bottom of the tail of the vehicle and shoots towards the advancing direction of the vehicle, the line laser emitter is installed in the middle of the chassis of the vehicle or at the position of the chassis of the vehicle close to the head of the vehicle and emits line laser towards the back, and the chassis of the vehicle and clear laser line imaging can be seen in the image collected by the camera.
Taking fig. 4 (b) as an example, the high-definition camera is installed in the middle of the vehicle chassis or at a position close to the head of the vehicle, and shoots an image backward, the line laser sensor is installed on the trunk door and emits line laser toward the road surface, and the vehicle chassis and clear laser line imaging can be seen in the image collected by the camera.
Taking fig. 4 (c) as an example, the high-definition camera is installed on a trunk door of the vehicle and shoots towards the advancing direction of the vehicle, the line laser transmitter is installed in the middle of the chassis of the vehicle or at a position of the chassis of the vehicle close to the head of the vehicle and emits line laser towards the rear, and the road surface and clear laser line imaging can be seen in the image collected by the camera.
Taking fig. 4 (d) as an example, the high-definition camera is installed at the bottom of the tail of the vehicle and shoots towards the rear of the vehicle, the line laser emitter is installed on the trunk door and emits line laser towards the road surface, and the image collected by the camera is the road behind the vehicle and the clear laser line image. The installation mode can collect laser lines and images of surrounding roads at the same time, and can help maintenance personnel to find the positions of disease points quickly.
Inputting the road surface image into a trained type recognition model to obtain a road surface type; the pavement types comprise asphalt pavement, cement pavement and sand stone pavement;
the type recognition model is a trained neural network model, after the preprocessed collected image is input, the road surface types can be divided into asphalt, cement and sand, and the classification accuracy rate is more than or equal to 95%.
Performing regression fitting on the laser line in the road surface image according to a preset fitting algorithm to obtain a coordinate point of the laser line in the road surface image;
the fitting algorithm is a laser line coordinate point fitting algorithm, and is a trained neural network model, after a preprocessed collected image is input, coordinate points [ u, v ] of a laser line on the image can be fitted, the number of the generated coordinate points is more than or equal to 10, and the change condition of the laser line on the image can be fitted well. When the deformation of the road surface is too large or peripheral light interference causes that laser line imaging in the image is not obvious, interruption occurs or the color changes, the neural network can resist the interference and estimate the possible trend of the laser line in the interrupted or unobvious area.
Carrying out deformation analysis on the laser line based on the coordinate points to determine deformation parameters; the deformation analysis comprises straight line correlation analysis and curvature analysis;
and generating an identification record according to the deformation parameters and the road surface type.
As a preferred embodiment of the technical solution of the present invention, the step of acquiring the road surface image containing the laser line based on the preset acquisition frequency includes:
acquiring the motion parameters of the vehicle in real time, and adjusting the acquisition frequency according to the motion parameters; the motion parameters comprise motion speed and motion acceleration; the motion speed and the motion acceleration are both vectors.
In one example of the technical scheme of the invention, the frequency of the acquired images of the high-definition camera can be adjusted in various ways, and can be dynamically adjusted according to the vehicle speed, wherein the higher the vehicle speed is, the higher the acquisition frequency is; a rotary encoder can be additionally arranged on the wheel, and when the vehicle advances for a certain distance, the central industrial personal computer sends an image acquisition instruction.
The central industrial personal computer comprises a fourth generation or fifth generation mobile communication technology, an information acquisition system, a high-precision positioning system and a miniature image processor. The function is to send an image acquisition command to the high-definition camera according to the running speed or the advancing distance of the vehicle, and continuously receive, process and upload acquired data at the same time. The mounting position is the vehicle interior.
As a preferred embodiment of the technical solution of the present invention, the step of performing deformation analysis on the laser line based on the coordinate points and determining deformation parameters includes:
performing straight line fitting on the coordinate points according to a least square method, and calculating a linear correlation coefficient;
comparing the linear correlation coefficient with a preset coefficient threshold;
when the linear correlation coefficient reaches a preset coefficient threshold value, judging that the laser line deformation is normal;
when the linear correlation coefficient is smaller than a preset coefficient threshold value, judging that the laser line deformation is abnormal, and determining the deformation type according to the fluctuation direction; the deformation types include depressions and protrusions.
In an example of the technical scheme of the invention, a least square method can be used for carrying out linear fitting on [ u, v ], and calculating a linear correlation coefficient R, if the linear correlation coefficient R is higher than a certain threshold value, the road surface is judged not to have obvious deformation damage, and if the linear correlation coefficient R is smaller than the certain threshold value, the laser line at the position has obvious fluctuation.
As a preferred embodiment of the technical solution of the present invention, the step of generating an identification record according to the deformation parameter and the road surface type includes:
inputting the road surface image containing the laser line into a trained facility identification model, and identifying a target object; the target comprises a well cover, a road foreign body and a road deceleration strip;
when the target object is not identified, reading the deformation parameter and the road surface type, and generating a first identification record according to the deformation parameter and the road surface type;
and when the target object is identified, reading the deformation parameter and the road surface type, and generating a second identification record according to the deformation parameter and the road surface type.
Further, when the target object is not identified, the step of reading the deformation parameter and the road surface type and generating the first identification record according to the deformation parameter and the road surface type includes:
when the target object is not identified, reading the deformation parameter and the road surface type;
when the pavement type is an asphalt pavement and the deformation type is a depression, calculating the length of the depression; when the length of the recess reaches a preset length threshold value, judging that the abnormal type is a track; when the length of the recess is smaller than a preset length threshold value, judging that the abnormal type is sink;
when the pavement type is asphalt pavement and the deformation type is convex, judging that the abnormal type is hug;
and when the road surface type is a cement road surface and the deformation type is a bulge, judging that the abnormal type is the bulge.
Specifically, the step of reading the deformation parameter and the road surface type and generating the second identification record according to the deformation parameter and the road surface type when the target object is identified comprises:
when the identified target object is a well lid and the linear correlation coefficient exceeds a preset threshold value, judging that the abnormal type is the height difference of the well lid;
when the identified target object is a road foreign object and the linear correlation coefficient exceeds a preset threshold value, judging that the abnormal type is a dangerous foreign object;
and when the identified target object is a deceleration strip and the linear correlation coefficient exceeds a preset threshold value, judging that the abnormal type is a large-altitude-difference deceleration strip.
The method for judging whether the road surface has deformation diseases or not or whether well covers, foreign bodies and deceleration strips with large deformation exist is obtained by comprehensively judging the combination of a road surface type analysis result, a laser line fitting coordinate point linear analysis result and an image target identification result:
(1) And (4) carrying out linear correlation analysis on the [ u, v ], such as using a least square method, carrying out linear fitting on the [ u, v ], calculating a linear correlation coefficient R, and judging that the road surface has no obvious deformation damage if the linear correlation coefficient R is higher than a certain threshold value.
(2) If the linear correlation coefficient R is lower than a certain threshold value and the manhole cover and the foreign matter are not identified near the laser line by utilizing the neural network, judging the type of the deformation disease according to whether the deformation direction meter of the laser line continuously appears or not by combining the road surface type, for example:
a. if the pavement type is asphalt pavement, the laser line inter-line zone is sunken downwards and appears continuously, and the pavement is judged to be a track;
b. if the pavement type is the asphalt pavement, the laser wire wheels sink downwards but do not continuously appear, and the pavement is judged to sink;
c. if the pavement type is asphalt pavement, the laser line is raised upwards, and the pavement is judged to be raised and upheaved;
d. if the road surface type is a cement road surface, the laser line is convex upwards, and the road surface is determined to be arched.
(3) If the linear correlation coefficient R is lower than a certain threshold value, and the well lid and the foreign matter are identified near the laser line by utilizing the neural network, the type of the target object is combined for judgment:
a. if the identified target object well lid exceeds a certain threshold value, judging that the well lid height difference exists;
b. if the identified target object is a road foreign object, judging that the linear correlation coefficient R exceeds a certain threshold value, and judging that the target object is a dangerous foreign object;
c. if the identified target object is a deceleration strip, the linear correlation coefficient R exceeds a certain threshold value, and the deceleration strip with a large height difference is determined.
As a preferred embodiment of the technical solution of the present invention, if a road surface with a deformation disease or a manhole cover height difference is identified, the central industrial personal computer records and stores related data in the system, and uploads the related data to the cloud server, where the data content includes information such as an acquired image, a result of the determination, position information of the vehicle, a traveling speed of the vehicle, and a heading angle of the vehicle. After the data are uploaded to the cloud server, the cloud server records the data into a database and sends messages to the relevant platform and the business system. And the analysis result is uploaded to a cloud server from a central industrial personal computer, and a power-off continuous transmission mode is used for dealing with the network fluctuation condition in the travelling process of the patrol vehicle.
The collection of the road surface deformation condition can be realized by means of the light sensor, the road surface deformation condition is analyzed by means of an artificial intelligence algorithm, rapid and accurate identification can be realized, and the analysis result can be recorded to the local and uploaded to a cloud server. Specifically, the method comprises the following steps:
(1) Compared with manual detection, the method and the system can be used for rapidly and intelligently analyzing the road surface deformation diseases, and workers only need to be concentrated in driving, so that the working efficiency of the workers is greatly improved, and the safety risk is reduced.
(2) Compared with the existing heavy detection equipment, the method and the system of the patent mainly rely on the light weight sensor, can cope with external interference by using an intelligent algorithm, and greatly reduce the equipment cost and the operation cost. By means of edge calculation and intelligent algorithm, real-time calculation of road surface deformation becomes possible, and the routing inspection efficiency of road surface deformation diseases is obviously improved.
(3) The intelligent algorithm and the judgment logic can solve the problems of road surface deformation and the like caused by facilities such as laser line terminals, well covers and the like due to external light interference, unclear laser line imaging and overlarge road surface deformation, and the system analysis has higher fine granularity and better robustness.
(4) In the aspect of maintenance benefit, the method and the system can realize high-frequency and large-range detection of the road surface deformation diseases, and can track the deformation diseases by combining a high-precision GPS. Meanwhile, the position relation of the laser emitter and the high-definition camera can be adjusted according to the actual vehicle condition and the service requirement, and surrounding road facilities can be collected by the image to help maintenance personnel to position the deformation position.
Example 2
Fig. 5 is a block diagram of a structure of a system for detecting road rut congestion based on a neural network, in an embodiment of the present invention, the system 10 includes:
the image acquisition module 11 is configured to acquire a road surface image containing a laser line based on a preset acquisition frequency; the laser line is emitted by a preset line laser emitter;
the road surface type recognition module 12 is used for inputting the road surface image into a trained type recognition model to obtain a road surface type; the pavement types comprise asphalt pavement, cement pavement and sand stone pavement;
the laser line identification module 13 is configured to perform regression fitting on the laser line in the road surface image according to a preset fitting algorithm to obtain a coordinate point of the laser line in the road surface image;
the parameter analysis module 14 is configured to perform deformation analysis on the laser line based on the coordinate points, and determine deformation parameters; the deformation analysis comprises straight line correlation analysis and curvature analysis;
and the record generating module 15 is configured to generate an identification record according to the deformation parameter and the road surface type.
The above description is intended to be illustrative of the preferred embodiment of the present invention and should not be taken as limiting the invention, but rather, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention.
Claims (7)
1. A road rut congestion detection method based on a neural network is characterized by comprising the following steps:
collecting a pavement image containing laser lines based on a preset collection frequency; the laser line is emitted by a preset line laser emitter;
inputting the road surface image into a trained type recognition model to obtain a road surface type; the pavement types comprise asphalt pavement, cement pavement and sand stone pavement;
performing regression fitting on the laser line in the road surface image according to a preset fitting algorithm to obtain a coordinate point of the laser line in the road surface image;
carrying out deformation analysis on the laser line based on the coordinate points to determine deformation parameters; the deformation analysis comprises straight line correlation analysis and curvature analysis;
and generating an identification record according to the deformation parameters and the road surface type.
2. The neural network-based road rut congestion detection method according to claim 1, wherein the step of collecting the road surface image containing the laser line based on the preset collection frequency comprises:
acquiring motion parameters of a vehicle in real time, and adjusting acquisition frequency according to the motion parameters; the motion parameters comprise motion speed and motion acceleration; the motion speed and the motion acceleration are both vectors.
3. The neural network-based road rut congestion detection method of claim 1, wherein said step of performing deformation analysis on the laser line based on the coordinate points and determining deformation parameters comprises:
performing straight line fitting on the coordinate points according to a least square method, and calculating a linear correlation coefficient;
comparing the linear correlation coefficient with a preset coefficient threshold;
when the linear correlation coefficient reaches a preset coefficient threshold value, judging that the laser line deformation is normal;
when the linear correlation coefficient is smaller than a preset coefficient threshold value, judging that the laser line deformation is abnormal, and determining the deformation type according to the fluctuation direction; the deformation types include depressions and protrusions.
4. The neural network-based road rut congestion detection method according to claim 1, wherein the step of generating an identification record according to the deformation parameter and the road surface type comprises:
inputting the road surface image containing the laser line into a trained facility recognition model, and recognizing a target object; the target comprises a well cover, a road foreign body and a road deceleration strip;
when the target object is not identified, reading the deformation parameter and the road surface type, and generating a first identification record according to the deformation parameter and the road surface type;
and when the target object is identified, reading the deformation parameter and the road surface type, and generating a second identification record according to the deformation parameter and the road surface type.
5. The neural network-based road rut congestion detection method according to claim 4, wherein the step of reading the deformation parameter and the road surface type when the target object is not identified and generating the first identification record according to the deformation parameter and the road surface type comprises:
when the target object is not identified, reading the deformation parameter and the road surface type;
when the pavement type is an asphalt pavement and the deformation type is a depression, calculating the length of the depression; when the length of the recess reaches a preset length threshold value, judging that the abnormal type is a track; when the length of the recess is smaller than a preset length threshold value, judging that the abnormal type is sink;
when the pavement type is asphalt pavement and the deformation type is convex, judging that the abnormal type is hug;
and when the road surface type is a cement road surface and the deformation type is a bulge, judging that the abnormal type is the bulge.
6. The neural network-based road rut congestion detection method according to claim 4, wherein the step of reading the deformation parameter and the road surface type when the target object is identified, and generating the second identification record according to the deformation parameter and the road surface type comprises:
when the identified target object is a well lid and the linear correlation coefficient exceeds a preset threshold value, judging that the abnormal type is the height difference of the well lid;
when the identified target object is a road foreign object and the linear correlation coefficient exceeds a preset threshold value, judging that the abnormal type is a dangerous foreign object;
and when the identified target object is a deceleration strip and the linear correlation coefficient exceeds a preset threshold value, judging that the abnormal type is a large-altitude-difference deceleration strip.
7. A neural network-based road rut hugging detection system, the system comprising:
the image acquisition module is used for acquiring a road surface image containing laser lines based on a preset acquisition frequency; the laser line is emitted by a preset line laser emitter;
the road surface type recognition module is used for inputting the road surface image into a trained type recognition model to obtain a road surface type; the pavement types comprise asphalt pavement, cement pavement and sand stone pavement;
the laser line identification module is used for performing regression fitting on the laser line in the road surface image according to a preset fitting algorithm to obtain a coordinate point of the laser line in the road surface image;
the parameter analysis module is used for carrying out deformation analysis on the laser line based on the coordinate points to determine deformation parameters; the deformation analysis comprises straight line correlation analysis and curvature analysis;
and the record generating module is used for generating an identification record according to the deformation parameter and the road surface type.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210955092.3A CN115223128A (en) | 2022-08-10 | 2022-08-10 | Road rut congestion detection method and system based on neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210955092.3A CN115223128A (en) | 2022-08-10 | 2022-08-10 | Road rut congestion detection method and system based on neural network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115223128A true CN115223128A (en) | 2022-10-21 |
Family
ID=83615043
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210955092.3A Pending CN115223128A (en) | 2022-08-10 | 2022-08-10 | Road rut congestion detection method and system based on neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115223128A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116704446A (en) * | 2023-08-04 | 2023-09-05 | 武汉工程大学 | Real-time detection method and system for foreign matters on airport runway pavement |
-
2022
- 2022-08-10 CN CN202210955092.3A patent/CN115223128A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116704446A (en) * | 2023-08-04 | 2023-09-05 | 武汉工程大学 | Real-time detection method and system for foreign matters on airport runway pavement |
CN116704446B (en) * | 2023-08-04 | 2023-10-24 | 武汉工程大学 | Real-time detection method and system for foreign matters on airport runway pavement |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Liu et al. | A review of applications of visual inspection technology based on image processing in the railway industry | |
CN111540201B (en) | Vehicle queuing length real-time estimation method and system based on roadside laser radar | |
CN103714363B (en) | A kind of motor vehicle exhaust smoke video identification system | |
US20150178572A1 (en) | Road surface condition classification method and system | |
CN102759347B (en) | Online in-process quality control device and method for high-speed rail contact networks and composed high-speed rail contact network detection system thereof | |
CN102426785B (en) | Traffic flow information perception method based on contour and local characteristic point and system thereof | |
CN114003849B (en) | Multi-lane non-contact type automatic calculation method and system for vehicle axle number | |
CN109238756A (en) | Lorry operation troubles motion image detection equipment and detection method | |
CN110806411A (en) | Unmanned aerial vehicle rail detecting system based on line structure light | |
CN115223128A (en) | Road rut congestion detection method and system based on neural network | |
CN114387785A (en) | Safety management and control method and system based on intelligent highway and storable medium | |
CN114715168A (en) | Vehicle yaw early warning method and system under road marking missing environment | |
CN114624726A (en) | Wheel axle identification system and wheel axle identification method | |
Chen et al. | A Novel Background Filtering Method with Automatic Parameter Adjustment for Real-Time Roadside LiDAR Sensing System | |
CN117409584B (en) | Novel smart city traffic management method and system based on carbon neutralization | |
CN116631187B (en) | Intelligent acquisition and analysis system for case on-site investigation information | |
Singh et al. | Road pothole detection from smartphone sensor data using improved LSTM | |
CN115655119A (en) | Method for detecting scope and size of road line laser rut subsidence congestion | |
KR102097673B1 (en) | System and method for obtaining information of vehicle using side image thereof | |
CN115372932A (en) | System and method for evaluating driveability of monorail crane and predicting roadway risk | |
CN114462647A (en) | Highway maintenance management system based on detect cloud platform | |
CN115326131A (en) | Intelligent analysis method and system for unmanned mine road conditions | |
CN113674329A (en) | Vehicle driving behavior detection method and system | |
CN113919571A (en) | Dangerous chemical transportation premium estimation method based on edge calculation | |
CN216049690U (en) | Online rolling stock wheel rim thickness measuring instrument |
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 |