CN113361413A - Mileage display area detection method, device, equipment and storage medium - Google Patents
Mileage display area detection method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN113361413A CN113361413A CN202110635677.2A CN202110635677A CN113361413A CN 113361413 A CN113361413 A CN 113361413A CN 202110635677 A CN202110635677 A CN 202110635677A CN 113361413 A CN113361413 A CN 113361413A
- Authority
- CN
- China
- Prior art keywords
- initial
- image
- detected
- standard
- region
- 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.)
- Granted
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 94
- 238000000034 method Methods 0.000 claims abstract description 67
- 238000012216 screening Methods 0.000 claims abstract description 42
- 238000012545 processing Methods 0.000 claims abstract description 24
- 238000004590 computer program Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 abstract description 24
- 239000013598 vector Substances 0.000 description 19
- 238000013528 artificial neural network Methods 0.000 description 10
- 238000004422 calculation algorithm Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 230000003287 optical effect Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000002372 labelling Methods 0.000 description 3
- 238000001914 filtration Methods 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 230000000717 retained effect Effects 0.000 description 2
- 239000013589 supplement Substances 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 210000000887 face Anatomy 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 210000003128 head Anatomy 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008707 rearrangement Effects 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 210000001525 retina Anatomy 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computational Linguistics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the application discloses a method, a device, equipment and a storage medium for detecting a mileage display area, wherein the method comprises the following steps: processing an image to be detected to obtain at least one initial region associated with a range display region in the image to be detected and initial characteristic data corresponding to the at least one initial region respectively; performing feature matching on at least one piece of initial feature data and standard feature data of a standard template picture associated with the image to be detected to obtain a feature matching result; performing primary screening on each initial region according to the feature matching result to obtain at least one target region; performing secondary screening on each target area to obtain a prediction area of the mileage display area; by the technical scheme, the process of target detection is optimized, and the recall rate of target detection is improved.
Description
Technical Field
The embodiment of the application relates to the technical field of target detection, in particular to a method, a device, equipment and a storage medium for detecting a mileage display area.
Background
Object detection is an important application of computer vision, for example, detecting a target object such as a human face, a vehicle, or a building from an image. In the solutions provided by the related art, usually, a target detection model is trained, and target detection is realized through the trained target detection model.
However, due to the influence of factors such as the lighting condition and the shooting angle, when the image to be recognized is input to the target detection model and the target in the image to be recognized is detected, even if the image to be recognized has the target to be detected, there may be a case that the recognition is failed or the recognition is wrong.
Therefore, there is a need for improvement in view of the problems in the prior art.
Disclosure of Invention
The application provides a method, a device, equipment and a storage medium for detecting a mileage display area, so as to optimize a target detection process and improve the recall rate of target detection.
In a first aspect, an embodiment of the present application provides a method for detecting a mileage display area, where the method includes:
processing an image to be detected to obtain at least one initial region associated with a range display region in the image to be detected and initial characteristic data corresponding to the at least one initial region respectively;
performing feature matching on at least one piece of initial feature data and standard feature data of a standard template picture associated with the image to be detected to obtain a feature matching result;
performing primary screening on each initial region according to the feature matching result to obtain at least one target region;
and carrying out secondary screening on each target area to obtain a prediction area of the mileage display area.
In a second aspect, an embodiment of the present application further provides a device for detecting a mileage display area, where the device includes:
the image processing module is used for processing an image to be detected to obtain at least one initial region associated with a range display region in the image to be detected and initial characteristic data respectively corresponding to the at least one initial region;
the matching module is used for performing feature matching on at least one piece of initial feature data and standard feature data of a standard template picture associated with the image to be detected to obtain a feature matching result;
the target area screening module is used for primarily screening each initial area according to the feature matching result to obtain at least one target area;
and the prediction area determining module is used for carrying out secondary screening on each target area to obtain the prediction area of the mileage display area.
In a third aspect, an embodiment of the present application further provides an electronic device, where the device includes:
one or more processors;
a storage device for storing one or more programs,
when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the mileage display area detection methods provided by the embodiments of the first aspect.
In a sixth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements any one of the mileage display area detection methods provided in the embodiments of the first aspect.
The method comprises the steps that an image to be detected is processed, so that at least one initial region associated with a range display region in the image to be detected and initial feature data corresponding to the at least one initial region are obtained; performing feature matching on at least one piece of initial feature data and standard feature data of a standard template picture associated with the image to be detected to obtain a feature matching result; performing primary screening on each initial region according to the feature matching result to obtain at least one target region; and carrying out secondary screening on each target area to obtain a prediction area of the mileage display area. Through the technical scheme, based on the characteristic matching result of the initial characteristic data and the standard characteristic data, the position areas possibly having the driving mileage areas are preliminarily screened, the position areas with the driving mileage areas more probably are screened out from the initial areas, the number of the initial areas is reduced, meanwhile, the initial areas can be screened in a purposeful and directional mode, the process of target detection is optimized, the recall rate of the target detection is improved, and meanwhile, the target detection speed is also improved.
Drawings
Fig. 1 is a flowchart of a method for detecting a mileage display area according to an embodiment of the present application;
fig. 2 is a flowchart of a method for detecting a mileage display area according to a second embodiment of the present application;
fig. 3 is a schematic diagram of a mileage display area detecting process provided in the second embodiment of the present application;
fig. 4 is a flowchart of a mileage display area detecting method provided in the third embodiment of the present application;
fig. 5 is a schematic diagram of a mileage display area detecting device according to a fourth embodiment of the present application;
fig. 6 is a schematic view of an electronic device provided in this application embodiment five.
Detailed Description
The present application will be described in further detail with reference to the following drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the application and are not limiting of the application. It should be further noted that, for the convenience of description, only some of the structures related to the present application are shown in the drawings, not all of the structures.
Before discussing exemplary embodiments in more detail, it should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although a flowchart may describe the steps as a sequential process, many of the steps can be performed in parallel, concurrently or simultaneously. In addition, the order of the steps may be rearranged. The process may be terminated when its operations are completed, but may have additional steps not included in the figure. The processes may correspond to methods, functions, procedures, subroutines, and the like.
Example one
Fig. 1 is a flowchart of a method for detecting a mileage display area according to an embodiment of the present application. The embodiment of the application can be suitable for detecting the mileage display area from the vehicle console image. The method can be executed by a mileage display area detecting device, which can be implemented by software and/or hardware and is specifically configured in an electronic device, which can be a mobile terminal or a fixed terminal.
Referring to fig. 1, a method for detecting a mileage display area provided in an embodiment of the present application includes:
s110, processing the image to be detected to obtain initial characteristic data corresponding to at least one initial region and at least one initial region associated with the middle range display region in the image to be detected.
The image to be detected is an input vehicle center console image, a mileage display area, also called a mileage area, exists in the vehicle center console image, and the kilometers of the vehicle that has traveled since the vehicle leaves a factory can be known by reading the numbers in the mileage area. In general, the mileage of a vehicle is one of important indicators for measuring the breakage rate of the vehicle, and therefore, it is important to identify the mileage area of the vehicle in the console image.
The initial area refers to a position area which is determined to be possible to have a mileage display area, and the initial characteristic data is used for characterizing the characteristic condition of the initial area, and each initial area can determine the corresponding initial characteristic data.
In this embodiment, the image to be detected may be input to a pre-trained target detection model, so as to obtain initial feature data corresponding to at least one initial region and at least one initial region associated with the range display region in the image to be detected.
The target detection model may be a deep neural Network mileage detection model, typically a back propagation Network (back propagation neural Network) model, and is trained in a supervised learning manner based on pre-labeled vehicle sample data.
Specifically, vehicle sample data can be obtained by collecting center console pictures of various different vehicle types, vehicle series and annual money for manual labeling and positioning a travel mileage area in the center console pictures. For simplicity, four vertices of the mileage area can be located in the console picture, and the location area where the four vertices are located is determined as the mileage area.
In some embodiments, the target detection model may also be a deep neural network mileage detection model based on reinforcement learning, and according to a set target function and a set reward value function, the agent interacts with the environment in a set environment, and as the number of interactions increases, the model evolves continuously.
Alternatively, the initial Feature data may be obtained from an initial region based on SIFT-Invariant Feature Transform (Scale-Invariant Feature Transform). However, when the image quality of the image to be detected is not high, the acquired feature data may be inaccurate or incorrect.
Or optionally, in order to acquire initial feature data with higher accuracy from the initial region, the image to be processed may be input to a deep neural network mileage detection model, and a detection result output by a detection head in the target detection model and a feature vector in a last-layer feature map output by a multi-scale feature fusion structure in the target detection model are acquired.
The detection result includes a regression frame, a confidence level, and a coordinate point, where the regression frame is the initial region in this embodiment, and the feature vector in the last layer of feature map is the initial feature data in this embodiment. The outputs of the deep neural network mileage detection model are in one-to-one correspondence, namely, one point in the confidence coefficient corresponds to one frame in the regression frame, a group of coordinates in the coordinate points and the feature vector in the feature map.
It can be understood that compared with the SIFT features, the initial feature data determined by the deep neural network mileage detection model can better describe the feature conditions (such as size and rotation invariance) of the initial region, and the accuracy of the initial feature data obtained from the initial region is higher.
And S120, performing feature matching on at least one piece of initial feature data and standard feature data of a standard template picture associated with the image to be detected to obtain a feature matching result.
The standard template image refers to a vehicle center console image similar to the image to be detected, namely the type of the vehicle center console image. It can be understood that, for different types of vehicles, the acquired vehicle console images are also different, for example, the size and shape of the mileage area are different for different types of vehicle console images.
The standard feature data is obtained from a standard area in the standard template drawing, and the standard area refers to a position area which is determined that the mileage display area is likely to exist, and the standard feature data is used for characterizing the standard area.
Optionally, the performing feature matching on at least one of the initial feature data and the standard feature data of the standard template map associated with the image to be detected includes: and calculating the Hamming distance of at least one piece of initial characteristic data and the standard characteristic data of the standard template picture associated with the image to be detected.
It can be understood that by calculating the hamming distance between the standard feature data and the initial feature data and using the hamming distance as the evaluation criterion of the feature matching similarity, the difference between the features can be accurately measured without performing complicated calculation, which has high efficiency and accelerates the speed of feature matching, thereby improving the speed of target detection.
Optionally, feature matching may be performed on initial feature data corresponding to a part of the initial region in the initial region and standard feature data of a standard template drawing associated with the image to be detected, so as to obtain a feature matching result; of course, the initial feature data corresponding to all the initial regions may also be subjected to feature matching with the standard feature data of the standard template drawing associated with the image to be detected, so as to obtain a feature matching result.
For example, according to the confidence of the initial region, the initial feature data corresponding to the initial region with higher confidence and the standard feature data of the standard template drawing associated with the image to be detected may be subjected to feature matching to obtain a feature matching result, and of course, the initial feature data corresponding to the initial region meeting a certain confidence threshold and the standard feature data of the standard template drawing associated with the image to be detected may also be subjected to feature matching to obtain a feature matching result.
Optionally, the standard feature data and the initial feature data may be feature vectors obtained based on a deep neural network mileage detection model; and performing characteristic matching on at least one initial characteristic data and the standard characteristic data of the standard template picture associated with the image to be detected to obtain a characteristic matching result.
It can be understood that the SIFT feature matching method of the traditional graphics is easy to fail in matching when the size, the visual angle, the rotation angle, the illumination and other conditions of the image to be detected and the standard template image are different greatly, and the algorithm calculation amount is large, so that the target detection speed is influenced. Compared with the SIFT feature matching mode of the traditional graphics, the feature vector obtained based on the deep neural network mileage detection model can fully utilize the feature similarity, the size and the rotation invariance of the deep learning model, can accurately identify the image to be detected and the standard template image when the size, the visual angle and the rotation angle of the image are different greatly, and has higher robustness.
And S130, performing primary screening on each initial area according to the feature matching result to obtain at least one target area.
And the target area is a position area obtained after primary screening is carried out on each initial area. Of these target areas, there is a position area in which the mileage area is more likely to be traveled.
In this embodiment, according to the obtained feature matching result, the initial feature data may be screened, and an initial region where the initial feature data (by comparing with a preset threshold) matched with the standard template map feature associated with the image to be detected is located is retained, so that the initial region may be screened.
Specifically, if the hamming distance is calculated to perform the feature matching, the hamming distance in the feature matching result is not less than the preset distance threshold, and the matching is determined to be an erroneous matching, so as to perform filtering; and judging that the Hamming distance in the feature matching result is smaller than a preset distance threshold value, and determining that the matching is correct, thereby retaining. The preset distance threshold value can be set according to actual conditions and empirical values.
And S140, carrying out secondary screening on each target area to obtain a prediction area of the mileage display area.
The prediction area is a position area with a driving mileage area which is finally determined from the image to be detected.
In this embodiment, each target region may be secondarily screened by using a Non-Maximum Suppression (NMS) algorithm. The NMS algorithm is an algorithm for extracting a window with the highest score in the target detection result, and the window is the target area in this embodiment.
Specifically, in the process of target detection, a plurality of detection windows are generated for the same target, but in fact most of the contents in the windows are repeated, only one window is really needed, and each target area is processed through the NMS algorithm, so that each target can be detected only once, and a frame with the best detection effect is found from each target area, namely, a position area with a driving mileage area is finally determined from the image to be detected.
The method comprises the steps that an image to be detected is processed, so that at least one initial region associated with a range display region in the image to be detected and initial feature data corresponding to the at least one initial region are obtained; performing feature matching on at least one piece of initial feature data and standard feature data of a standard template picture associated with the image to be detected to obtain a feature matching result; performing primary screening on each initial region according to the feature matching result to obtain at least one target region; and carrying out secondary screening on each target area to obtain a prediction area of the mileage display area. Through the technical scheme, based on the characteristic matching result of the initial characteristic data and the standard characteristic data, the position areas possibly having the driving mileage areas are preliminarily screened, the position areas with the driving mileage areas more probably are screened out from the initial areas, the number of the initial areas is reduced, meanwhile, the initial areas can be screened in a purposeful and directional mode, the process of target detection is optimized, the recall rate of the target detection is improved, and meanwhile, the target detection speed is also improved.
Example two
Fig. 2 is a flowchart of a method for detecting a mileage display area according to a second embodiment of the present application, which is based on the above embodiments and optimizes the above schemes.
Further, the operation of processing the image to be detected to obtain at least one initial region associated with the intermediate range display region in the image to be detected and initial feature data corresponding to the at least one initial region respectively is refined into the operation of inputting the image to be detected to a target detection model based on a first confidence threshold value to obtain at least one first initial region associated with the intermediate range display region in the image to be detected; inputting the image to be detected into the target detection model based on a second confidence threshold value to obtain at least one second initial region associated with a range display region in the image to be detected and initial feature data corresponding to each second initial region; wherein the second confidence threshold is less than the first confidence threshold "; correspondingly, the operation of performing feature matching on at least one initial feature data and the standard feature data of the standard template drawing associated with the image to be detected to obtain a feature matching result is refined into the operation of performing feature matching on the initial feature data corresponding to at least one second initial region and the standard feature data of the standard template drawing associated with the image to be detected to obtain a feature matching result so as to clarify the processes of image processing and feature matching.
Wherein explanations of the same or corresponding terms as those of the above-described embodiments are omitted.
Referring to fig. 2, the method for detecting the mileage display area provided by the present embodiment includes:
s210, inputting the image to be detected into the target detection model based on the first confidence threshold value, and obtaining at least one first initial region associated with a range display region in the image to be detected.
The target detection model is a pre-trained deep neural network mileage detection model.
In this embodiment, the first confidence threshold may be set according to actual conditions and specific requirements. For example, a first confidence threshold may be set to 0.95, and a regression box with a confidence greater than 0.95 may be reserved as the first initial region according to the preset first confidence threshold.
S220, inputting the image to be detected into the target detection model based on a second confidence threshold value to obtain at least one second initial region associated with a range display region in the image to be detected and initial characteristic data corresponding to each second initial region; wherein the second confidence threshold is less than the first confidence threshold.
In this embodiment, similarly, the second confidence threshold may be set according to actual conditions and specific requirements. For example, the second confidence threshold may be set to 0.94, and a regression box with a confidence greater than 0.94 may be reserved as the second initial region according to the preset second confidence threshold.
Alternatively, the second confidence threshold may be set based on empirical values, such as may be determined through extensive experimentation.
In an alternative embodiment, the second confidence threshold is 0.009 below the first confidence threshold.
It should be noted that, since the second confidence threshold is smaller than the first confidence threshold, the number of the second initial regions obtained from the target detection model output is larger than the number of the first initial regions, i.e., the number of the initial regions obtained based on the smaller confidence threshold is larger.
And S230, performing feature matching on the initial feature data corresponding to the at least one second initial region and the standard feature data of the standard template drawing associated with the image to be detected to obtain a feature matching result.
In this embodiment, it is considered that the screening condition of the higher confidence threshold is more severe, and thus the number of the obtained first initial regions is small, and some position regions where mileage-traveled regions may exist may be omitted. Based on the method and the device, the initial characteristic data corresponding to the second initial areas with more quantity are subjected to characteristic matching with the standard characteristic data of the standard template drawing related to the image to be detected, so that the position areas where mileage areas possibly run are subjected to supplementary searching, and omission is avoided.
It can be understood that, due to the influence of factors such as the illumination condition and the shooting angle, when the image to be recognized is input into the target detection model and the target in the image to be recognized is detected, even if the image to be recognized has the target to be detected, there may be a case of recognition failure or recognition error, that is, the target object is detected only by the target detection model, and there may be a case of low recall rate. The method and the device fully utilize the characteristic relation between the image to be detected and the standard template drawing to discover and supplement the position area possibly having the mileage area, thereby improving the recall rate of target detection.
Optionally, the standard characteristic data of the standard template map is determined in the following manner: inputting the standard template graph to a target detection model based on a preset confidence threshold value to obtain at least one standard area; and acquiring standard characteristic data corresponding to each standard area.
The target detection model can also be a pre-trained deep neural network mileage detection model.
The preset confidence threshold may be the same as the first confidence threshold or the second confidence threshold; alternatively, the preset confidence threshold may also be different from the first confidence threshold or the second confidence threshold, and the setting of the preset confidence threshold may be determined according to an actual situation, which is not specifically limited in this embodiment of the present application.
In an alternative embodiment, the preset confidence threshold may be the same as the second confidence threshold.
Optionally, the inputting the standard template map into a target detection model based on a preset confidence threshold to obtain a standard region includes: inputting the standard template graph into a target detection model based on a preset confidence threshold value to obtain at least one candidate standard area and candidate standard characteristic data corresponding to each candidate standard area; taking the candidate standard area meeting the intersection ratio condition as the standard area; correspondingly, acquiring standard feature data corresponding to the standard area, including: and taking the candidate standard feature data corresponding to the standard area as the standard feature data.
The Intersection-over-Union condition is also a preset IoU (Intersection-over-Union) threshold condition, and the threshold value IoU may be set according to an actual situation.
In this embodiment, the finally obtained standard feature data may be controlled by setting a preset confidence threshold and a cross-over ratio condition.
In an optional embodiment, by reasonably setting the preset confidence level threshold and the intersection ratio condition, the number of the standard feature data can be controlled within a preset number range under the condition of not reducing the accuracy of target detection, so as to reduce the occupation of a storage space and improve the target detection speed.
Specifically, regression boxes with confidence higher than a preset confidence threshold value can be found first, IoU between the regression boxes and pre-labeled real boxes is calculated, a confidence point set larger than IoU threshold value is reserved, corresponding feature vectors of the points in the feature map are saved, the number of the feature vectors of each template map is smaller than 20, and the length of each feature vector is 256.
And S240, performing primary screening on each initial area according to the feature matching result to obtain at least one target area.
Wherein each initial region comprises a first initial region and a second initial region.
In this embodiment, each second initial region may be preliminarily screened according to the feature matching result, instead of the first initial region, because the first initial region is a position region determined with a higher confidence, which is generally considered as a position region where there is a greater probability of a mileage region being traveled.
Specifically, according to the feature matching result, the process of performing preliminary screening on each initial region may include: for each feature vector in the standard template graph, sequentially performing feature matching on the feature vector and each feature vector of the first initial region; and reserving the corresponding first initial region with the highest matching degree of the feature vectors in the feature matching result, and finally, screening to obtain position regions with the quantity consistent with that of the feature vectors of the standard template graph, namely obtaining the target region after primarily screening the second initial region.
In this embodiment, the second initial regions in which the feature matching is successful in the feature matching result can be retained, so that effective identification of the position regions where mileage areas may be traveled is realized, the situation that target regions are omitted under a higher confidence coefficient is avoided, and the purpose of screening each second initial region is achieved.
Optionally, the performing preliminary screening on each initial region according to the feature matching result to obtain at least one target region includes: and combining the successfully matched second initial region and the first initial region to obtain at least one target region.
Wherein, the first initial region can be understood as a normal recognition result obtained based on the target detection model.
In this embodiment, the second initial area in which the feature data is successfully matched is used as a supplement to the normal recognition result, the position area in which the mileage area is likely to be traveled is expanded into the target area, and the recall rate of target detection can be increased from 95% to 96% during the test.
It can be understood that by combining the successfully matched second initial region with the first initial region, the common sense result is supplemented, and the recall rate of target detection is improved.
It should be noted that, if there is no standard template associated with the image to be detected, the embodiment of the present application may directly perform secondary screening based on the first initial region to obtain the predicted region of the mileage display region.
And S250, carrying out secondary screening on each target area to obtain a prediction area of the mileage display area.
Referring to a schematic diagram of a mileage display area detection process shown in fig. 3, the target detection process is mainly implemented by two modules, namely a standard feature vector offline calculation module and a central console mileage area online detection module. The target detection model selects a Retina face algorithm (a single-stage target detection algorithm special for detecting human faces) model, and uses a high-resolution network as a backbone network, and the target detection model outputs a confidence coefficient, a regression frame, a coordinate point and a last layer of feature map of the backbone network, wherein the outputs are in one-to-one correspondence; the data quantity of the feature vectors stored in each standard template graph is not more than n x 256 floating point numbers, namely the number of the feature vectors in each standard template graph is less than n (for example, n can be 20), and the length of each feature vector is 256, so that the storage space occupation and the reading time are greatly reduced.
Specifically, fig. 3 exemplarily shows a console mileage detection method based on deep learning feature similarity matching, and an implementation process of the method mainly includes two stages: the first stage is common confidence screening under higher confidence, and x positions can be obtained; and the second stage calculates the most similar characteristic vector under lower confidence, merges the corresponding second detection result (n results are selected from the y results) with the highest matching degree of the characteristic vector in the standard template graph in the characteristic matching result into the first stage result, and then processes the expanded first stage result (n + x results) by adopting an NMS algorithm to obtain a final detection result, thereby improving the recall rate of target detection.
On the basis of the embodiment, the target detection process is determined, and the image to be detected is input into a target detection model based on a first confidence threshold value to obtain at least one first initial region associated with a middle range display region in the image to be detected; inputting the image to be detected into the target detection model based on a second confidence threshold value to obtain at least one second initial region associated with a range display region in the image to be detected and initial feature data corresponding to each second initial region; wherein the second confidence threshold is less than the first confidence threshold; and performing characteristic matching on the initial characteristic data corresponding to at least one second initial region and the standard characteristic data of the standard template drawing associated with the image to be detected. By the technical scheme, the initial regions with different numbers can be obtained based on the first confidence coefficient threshold value and the second confidence coefficient threshold value, the initial feature data corresponding to at least one second initial region under the lower confidence coefficient is subjected to feature matching with the standard feature data of the standard template image associated with the image to be detected, the second initial region with successful feature matching is reserved, the position region where the mileage region possibly runs is effectively identified, the situation that the target region is missed under the higher confidence coefficient is avoided, the target detection process is optimized by performing secondary screening on each target region, and the recall rate of target detection is improved.
EXAMPLE III
Fig. 4 is a flowchart of a method for detecting a mileage display area according to a third embodiment of the present application, where the present embodiment is an optimization of the foregoing scheme based on the foregoing embodiments.
Further, adding operation of acquiring vehicle attribute data of a target vehicle corresponding to the image to be detected; wherein the vehicle attribute data comprises vehicle type, vehicle series and annual fee; and according to the vehicle attribute data, selecting a standard template drawing from at least one candidate standard template drawing associated with the target vehicle so as to perfect the selection process of the standard template drawing associated with the image to be detected.
Wherein explanations of the same or corresponding terms as those of the above-described embodiments are omitted.
Referring to fig. 4, the method for detecting the mileage display area provided in the present embodiment includes:
s310, inputting the image to be detected to a target detection model based on the first confidence threshold value, and obtaining at least one first initial region associated with a range display region in the image to be detected.
S320, inputting the image to be detected into the target detection model based on a second confidence threshold value to obtain at least one second initial region associated with a range display region in the image to be detected and initial characteristic data corresponding to each second initial region; wherein the second confidence threshold is less than the first confidence threshold.
S330, acquiring vehicle attribute data of a target vehicle corresponding to the image to be detected; the vehicle attribute data includes a vehicle type, a vehicle series, and a year money.
In this embodiment, the target vehicle is a vehicle in the image to be detected, and the image to be detected may be identified by an image processing technique, so as to identify vehicle attribute data of the target vehicle, including a vehicle type, a vehicle series, and a year money.
In some embodiments, a manual labeling mode may also be adopted to determine vehicle attribute data of the image to be detected corresponding to the target vehicle, for example, the vehicle attribute data is stored in association with the image to be detected in a keyword field mode, and when the image to be detected is read, the vehicle attribute data of the image to be detected corresponding to the target vehicle can be automatically obtained.
S340, according to the vehicle attribute data, selecting a standard template drawing from at least one candidate standard template drawing associated with the target vehicle.
In this embodiment, if the vehicle attribute data of the target vehicle corresponding to the image to be detected is predetermined in a manual labeling manner, and the vehicle attribute data of the vehicle standard template map in the candidate standard template maps is predetermined, the standard template map associated with the image to be detected may be determined in a field matching manner.
It can be understood that the characteristics of the images of vehicles of the same type are more similar, so that the target detection of the images to be detected is facilitated, and the accuracy of the target detection can be effectively improved.
In some embodiments, if no vehicle attribute data has been previously determined for each image, the image to be detected may be identified by image processing techniques to identify vehicle attribute data for the vehicle from the image. However, in the image recognition method, there is a case where the recognition fails, for example, the annual fee of the vehicle cannot be specifically recognized.
Thus, in some embodiments, it may also be provided to determine a similar standard template map associated with the image to be detected in the event that the annual payment for the vehicle cannot be accurately identified.
Optionally, the selecting a standard template drawing from at least one candidate standard template drawing associated with the target vehicle according to the vehicle attribute data includes: and if at least one candidate standard template drawing related to the target vehicle does not comprise a candidate standard template drawing corresponding to the annual fund of the target vehicle, selecting a candidate standard template drawing of the adjacent annual fund or no annual fund of the target vehicle as the standard template drawing.
In this embodiment, a comprehensive candidate standard template drawing without a annuity can be determined according to the requirement and the actual situation for selection, for example, features of vehicle template drawings of the same vehicle type and vehicle series within a set time can be fused to construct a corresponding candidate standard template drawing without a annuity. In addition, when the candidate standard template drawing corresponding to the yearly style is matched, the candidate standard template drawing adjacent to the yearly style can be selected as the standard template drawing, and because the types of the vehicle template drawings adjacent to the yearly style are relatively close, the candidate standard template drawing adjacent to the yearly style can be directly selected.
It can be understood that, in addition to strictly matching the vehicle type, the vehicle series and the annual fee, the vehicle type and the vehicle series can be matched, and the standard template drawing associated with the image to be detected is determined from the candidate standard template drawings, so that the flexibility of the standard template drawing determination process is improved.
And S350, performing characteristic matching on at least one piece of initial characteristic data and the standard characteristic data of the standard template picture associated with the image to be detected.
And S360, performing primary screening on each initial area according to the feature matching result to obtain at least one target area.
And S370, carrying out secondary screening on each target area to obtain a prediction area of the mileage display area.
On the basis of the embodiment, the determining process of the standard template graph associated with the image to be detected is clear, and vehicle attribute data of a target vehicle corresponding to the image to be detected is obtained; wherein the vehicle attribute data comprises vehicle type, vehicle series and annual fee; and selecting a standard template drawing from at least one candidate standard template drawing associated with the target vehicle according to the vehicle attribute data. Through the technical scheme, the standard template picture associated with the image to be detected is determined, the characteristic matching can be performed according to the determined standard characteristic data of the standard template picture, the data support is provided for performing the characteristic matching on the standard characteristic data of the standard template picture associated with the image to be detected and the initial characteristic data, and the process of target detection is optimized.
Example four
Fig. 5 is a schematic structural diagram of a mileage display area detecting device according to a fourth embodiment of the present application. Referring to fig. 5, an embodiment of the present application provides a device for detecting a mileage display area, including: an image processing module 410, a matching module 420, a target region screening module 430, and a prediction region determination module 440.
The image processing module 410 is configured to process an image to be detected to obtain at least one initial region associated with a home range display region in the image to be detected and initial feature data corresponding to the at least one initial region;
the matching module 420 is configured to perform feature matching on at least one piece of initial feature data and standard feature data of a standard template drawing associated with the image to be detected, so as to obtain a feature matching result;
a target area screening module 430, configured to perform preliminary screening on each initial area according to the feature matching result to obtain at least one target area;
and a prediction area determining module 440, configured to perform secondary screening on each target area to obtain a prediction area of the mileage display area.
The method comprises the steps that an image to be detected is processed, so that at least one initial region associated with a range display region in the image to be detected and initial feature data corresponding to the at least one initial region are obtained; performing feature matching on at least one piece of initial feature data and standard feature data of a standard template picture associated with the image to be detected to obtain a feature matching result; performing primary screening on each initial region according to the feature matching result to obtain at least one target region; and carrying out secondary screening on each target area to obtain a prediction area of the mileage display area. Through the technical scheme, based on the characteristic matching result of the initial characteristic data and the standard characteristic data, the position areas possibly having the driving mileage areas are preliminarily screened, the position areas with the driving mileage areas more probably are screened out from the initial areas, the number of the initial areas is reduced, meanwhile, the initial areas can be screened in a purposeful and directional mode, the process of target detection is optimized, the recall rate of the target detection is improved, and meanwhile, the target detection speed is also improved.
Further, the image processing module 410 includes:
the image processing first unit is used for inputting the image to be detected into a target detection model based on a first confidence coefficient threshold value to obtain at least one first initial region associated with a middle range display region in the image to be detected;
the image processing second unit is used for inputting the image to be detected into the target detection model based on a second confidence threshold value to obtain at least one second initial region associated with a middle range display region in the image to be detected and initial feature data corresponding to each second initial region; wherein the second confidence threshold is less than the first confidence threshold;
accordingly, the matching module 420 includes:
and the characteristic matching unit is used for carrying out characteristic matching on the initial characteristic data corresponding to at least one second initial region and the standard characteristic data of the standard template picture associated with the image to be detected to obtain a characteristic matching result.
Further, the target area filtering module 430 includes:
and the target area screening unit is used for combining the successfully matched second initial area and the first initial area to obtain at least one target area.
Further, the matching module 420 includes:
and the Hamming distance calculating unit is used for calculating the Hamming distance between at least one piece of initial characteristic data and the standard characteristic data of the standard template picture associated with the image to be detected.
Further, the apparatus further comprises: a standard feature data acquisition module, the standard feature data acquisition module comprising:
the standard area determining unit is used for inputting the standard template graph to a target detection model based on a preset confidence threshold value to obtain at least one standard area;
and the standard characteristic data acquisition unit is used for acquiring the standard characteristic data corresponding to each standard area.
Further, the apparatus further comprises: a template map determination module, the template map determination module comprising:
the attribute data acquisition unit is used for acquiring vehicle attribute data of a target vehicle corresponding to the image to be detected; wherein the vehicle attribute data comprises vehicle type, vehicle series and annual fee;
and the template drawing determining unit is used for selecting a standard template drawing from at least one candidate standard template drawing associated with the target vehicle according to the vehicle attribute data.
Further, the template map determination unit includes:
and the template map determining subunit is used for selecting the candidate standard template map which is adjacent to the annuity or has no annuity of the target vehicle as the standard template map if at least one candidate standard template map associated with the target vehicle does not contain the candidate standard template map corresponding to the annuity of the target vehicle.
The mileage display area detection device provided by the embodiment of the application can execute the mileage display area detection method provided by any embodiment of the application, and has the corresponding functional modules and beneficial effects of the execution method.
EXAMPLE five
Fig. 6 is a structural diagram of an electronic device according to a fifth embodiment of the present application. FIG. 6 illustrates a block diagram of an exemplary electronic device 512 suitable for use in implementing embodiments of the present application. The electronic device 512 shown in fig. 6 is only an example and should not bring any limitation to the functions and the scope of use of the embodiments of the present application.
As shown in fig. 6, the electronic device 512 is in the form of a general purpose computing device. Components of the electronic device 512 may include, but are not limited to: one or more processors or processing units 516, a system memory 528, and a bus 518 that couples the various system components including the system memory 628 and the processing unit 516.
The system memory 528 may include computer system readable media in the form of volatile memory, such as Random Access Memory (RAM)530 and/or cache memory 532. The electronic device 512 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 534 may be used to read from and write to non-removable, nonvolatile magnetic media (not shown in FIG. 6, and commonly referred to as a "hard drive"). Although not shown in FIG. 6, a magnetic disk drive for reading from and writing to a removable, nonvolatile magnetic disk (e.g., a "floppy disk") and an optical disk drive for reading from or writing to a removable, nonvolatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 518 through one or more data media interfaces. System memory 528 may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the application.
A program/utility 540 having a set (at least one) of program modules 542 may be stored, for example, in system memory 528, such program modules 542 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which examples or some combination thereof may comprise an implementation of a network environment. The program modules 542 generally perform the functions and/or methods of the embodiments described herein.
The electronic device 512 may also communicate with one or more external devices 514 (e.g., keyboard, pointing device, display 524, etc.), with one or more devices that enable a user to interact with the electronic device 512, and/or with any devices (e.g., network card, modem, etc.) that enable the electronic device 512 to communicate with one or more other computing devices. Such communication may occur via input/output (I/O) interfaces 522. Also, the electronic device 512 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network, such as the internet) via the network adapter 520. As shown, the network adapter 520 communicates with the other modules of the electronic device 512 via the bus 518. It should be appreciated that although not shown in FIG. 6, other hardware and/or software modules may be used in conjunction with the electronic device 512, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.
The processing unit 516 executes various functional applications and data processing by running at least one of other programs stored in the system memory 528, for example, implementing any one of the mileage display area detection methods provided in the embodiments of the present application.
EXAMPLE six
An embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements a method for detecting a mileage display area, the method including:
processing an image to be detected to obtain at least one initial region associated with a range display region in the image to be detected and initial characteristic data corresponding to the at least one initial region respectively; performing feature matching on at least one piece of initial feature data and standard feature data of a standard template picture associated with the image to be detected to obtain a feature matching result; performing primary screening on each initial region according to the feature matching result to obtain at least one target region; and carrying out secondary screening on each target area to obtain a prediction area of the mileage display area.
From the above description of the embodiments, it is obvious for those skilled in the art that the present application can be implemented by software and necessary general hardware, and certainly can be implemented by hardware, but the former is a better embodiment in many cases. Based on such understanding, the technical solutions of the present application may be embodied in the form of a software product, which may be stored in a computer-readable storage medium, such as a floppy disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a FLASH Memory (FLASH), a hard disk or an optical disk of a computer, and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device) to execute the methods described in the embodiments of the present application.
It should be noted that, in the embodiment of the mileage display area detecting device, the units and modules included in the embodiment are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, specific names of the functional units are only used for distinguishing one functional unit from another, and are not used for limiting the protection scope of the application.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present application and the technical principles employed. It will be understood by those skilled in the art that the present application is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the application. Therefore, although the present application has been described in more detail with reference to the above embodiments, the present application is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present application, and the scope of the present application is determined by the scope of the appended claims.
Claims (10)
1. A mileage display area detecting method is characterized by comprising:
processing an image to be detected to obtain at least one initial region associated with a range display region in the image to be detected and initial characteristic data corresponding to the at least one initial region respectively;
performing feature matching on at least one piece of initial feature data and standard feature data of a standard template picture associated with the image to be detected to obtain a feature matching result;
performing primary screening on each initial region according to the feature matching result to obtain at least one target region;
and carrying out secondary screening on each target area to obtain a prediction area of the mileage display area.
2. The method according to claim 1, wherein the processing the image to be detected to obtain at least one initial region associated with a range display region in the image to be detected and initial feature data corresponding to the at least one initial region respectively comprises:
inputting the image to be detected into a target detection model based on a first confidence threshold value to obtain at least one first initial region associated with a range display region in the image to be detected;
inputting the image to be detected into the target detection model based on a second confidence threshold value to obtain at least one second initial region associated with a range display region in the image to be detected and initial feature data corresponding to each second initial region; wherein the second confidence threshold is less than the first confidence threshold;
correspondingly, the performing feature matching on the at least one initial feature data and the standard feature data of the standard template drawing associated with the image to be detected includes:
and performing characteristic matching on the initial characteristic data corresponding to at least one second initial region and the standard characteristic data of the standard template drawing associated with the image to be detected.
3. The method of claim 2, wherein the pre-screening each of the initial regions to obtain at least one target region according to the feature matching result comprises:
and combining the successfully matched second initial region and the first initial region to obtain at least one target region.
4. The method according to claim 1, wherein the feature matching at least one of the initial feature data with the standard feature data of the standard template map associated with the image to be detected comprises:
and calculating the Hamming distance of at least one piece of initial characteristic data and the standard characteristic data of the standard template picture associated with the image to be detected.
5. The method of claim 1, wherein the standard feature data of the standard template graph is determined by:
inputting the standard template graph to a target detection model based on a preset confidence threshold value to obtain at least one standard area;
and acquiring standard characteristic data corresponding to each standard area.
6. The method according to claim 1, characterized in that the standard template map associated with the image to be detected is determined in the following way:
acquiring vehicle attribute data of a target vehicle corresponding to the image to be detected; wherein the vehicle attribute data comprises vehicle type, vehicle series and annual fee;
and selecting a standard template drawing from at least one candidate standard template drawing associated with the target vehicle according to the vehicle attribute data.
7. The method of claim 6, wherein selecting a standard template map from at least one candidate standard template map associated with the target vehicle based on the vehicle attribute data comprises:
and if at least one candidate standard template drawing related to the target vehicle does not comprise a candidate standard template drawing corresponding to the annual fund of the target vehicle, selecting a candidate standard template drawing of the adjacent annual fund or no annual fund of the target vehicle as the standard template drawing.
8. A mileage display area detecting device characterized by comprising:
the image processing module is used for processing an image to be detected to obtain at least one initial region associated with a range display region in the image to be detected and initial characteristic data respectively corresponding to the at least one initial region;
the matching module is used for performing feature matching on at least one piece of initial feature data and standard feature data of a standard template picture associated with the image to be detected to obtain a feature matching result;
the target area screening module is used for primarily screening each initial area according to the matching result to obtain at least one target area;
and the prediction area determining module is used for carrying out secondary screening on each target area to obtain the prediction area of the mileage display area.
9. An electronic device, comprising:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement a method of mileage display area detection as recited in any one of claims 1-7.
10. A computer-readable storage medium on which a computer program is stored, the program, when executed by a processor, implementing a mileage display area detecting method as recited in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110635677.2A CN113361413B (en) | 2021-06-08 | 2021-06-08 | Mileage display area detection method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110635677.2A CN113361413B (en) | 2021-06-08 | 2021-06-08 | Mileage display area detection method, device, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113361413A true CN113361413A (en) | 2021-09-07 |
| CN113361413B CN113361413B (en) | 2024-06-18 |
Family
ID=77533039
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110635677.2A Active CN113361413B (en) | 2021-06-08 | 2021-06-08 | Mileage display area detection method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113361413B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114241011A (en) * | 2022-02-22 | 2022-03-25 | 阿里巴巴达摩院(杭州)科技有限公司 | Target detection method, device, equipment and storage medium |
Citations (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140376769A1 (en) * | 2013-06-20 | 2014-12-25 | Xerox Corporation | Method for detecting large size and passenger vehicles from fixed cameras |
| CN105957345A (en) * | 2016-06-08 | 2016-09-21 | 南京航空航天大学 | Processing method for vehicle driving data |
| CN106960207A (en) * | 2017-04-26 | 2017-07-18 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of car steering position gauge field multipointer instrument automatic recognition system and method based on template matches |
| WO2018188453A1 (en) * | 2017-04-11 | 2018-10-18 | 腾讯科技（深圳）有限公司 | Method for determining human face area, storage medium, and computer device |
| CN109255336A (en) * | 2018-09-29 | 2019-01-22 | 南京理工大学 | Arrester recognition methods based on crusing robot |
| CN109409385A (en) * | 2018-10-16 | 2019-03-01 | 南京鑫和汇通电子科技有限公司 | A kind of pointer instrument automatic identifying method |
| CN109559300A (en) * | 2018-11-19 | 2019-04-02 | 上海商汤智能科技有限公司 | Image processing method, electronic equipment and computer readable storage medium |
| CN109766878A (en) * | 2019-04-11 | 2019-05-17 | 深兰人工智能芯片研究院（江苏）有限公司 | A kind of method and apparatus of lane detection |
| CN109815868A (en) * | 2019-01-15 | 2019-05-28 | 腾讯科技（深圳）有限公司 | A kind of image object detection method, device and storage medium |
| CN110146121A (en) * | 2019-05-24 | 2019-08-20 | 安徽扬远信息科技有限公司 | A kind of instrument board Performance Evaluation control system |
| CN110222780A (en) * | 2019-06-12 | 2019-09-10 | 北京百度网讯科技有限公司 | Object detecting method, device, equipment and storage medium |
| CN110245544A (en) * | 2018-09-26 | 2019-09-17 | 浙江大华技术股份有限公司 | A kind of method and device of determining dead ship condition |
| CN110378258A (en) * | 2019-07-04 | 2019-10-25 | 上海眼控科技股份有限公司 | A kind of vehicle seat information detecting method and equipment based on image |
| CN110517259A (en) * | 2019-08-30 | 2019-11-29 | 北京百度网讯科技有限公司 | A kind of detection method, device, equipment and the medium of product surface state |
| CN110991305A (en) * | 2019-11-27 | 2020-04-10 | 厦门大学 | Airplane detection method under remote sensing image and storage medium |
| CN111239158A (en) * | 2020-03-13 | 2020-06-05 | 苏州鑫睿益荣信息技术有限公司 | Automobile instrument panel detection system and detection method based on machine vision |
| US20200193225A1 (en) * | 2017-04-28 | 2020-06-18 | Toyota Motor Europe | System and method for detecting objects in a digital image, and system and method for rescoring object detections |
| CN111401424A (en) * | 2020-03-10 | 2020-07-10 | 北京迈格威科技有限公司 | Target detection method, device and electronic system |
| CN111523414A (en) * | 2020-04-13 | 2020-08-11 | 绍兴埃瓦科技有限公司 | Face recognition method and device, computer equipment and storage medium |
| CN111709416A (en) * | 2020-05-15 | 2020-09-25 | 珠海亿智电子科技有限公司 | License plate positioning method, device and system and storage medium |
| CN111768392A (en) * | 2020-06-30 | 2020-10-13 | 创新奇智(广州)科技有限公司 | Target detection method and device, electronic equipment and storage medium |
| CN111832557A (en) * | 2020-06-04 | 2020-10-27 | 北京百度网讯科技有限公司 | Power grid inspection method and device, electronic equipment and storage medium |
| CN112580657A (en) * | 2020-12-23 | 2021-03-30 | 陕西天诚软件有限公司 | Self-learning character recognition method |
| CN112639872A (en) * | 2020-04-24 | 2021-04-09 | 华为技术有限公司 | Method and device for difficult mining in target detection |
| US20210224998A1 (en) * | 2018-11-23 | 2021-07-22 | Tencent Technology (Shenzhen) Company Limited | Image recognition method, apparatus, and system and storage medium |
-
2021
- 2021-06-08 CN CN202110635677.2A patent/CN113361413B/en active Active
Patent Citations (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140376769A1 (en) * | 2013-06-20 | 2014-12-25 | Xerox Corporation | Method for detecting large size and passenger vehicles from fixed cameras |
| CN105957345A (en) * | 2016-06-08 | 2016-09-21 | 南京航空航天大学 | Processing method for vehicle driving data |
| WO2018188453A1 (en) * | 2017-04-11 | 2018-10-18 | 腾讯科技（深圳）有限公司 | Method for determining human face area, storage medium, and computer device |
| CN106960207A (en) * | 2017-04-26 | 2017-07-18 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of car steering position gauge field multipointer instrument automatic recognition system and method based on template matches |
| US20200193225A1 (en) * | 2017-04-28 | 2020-06-18 | Toyota Motor Europe | System and method for detecting objects in a digital image, and system and method for rescoring object detections |
| CN110245544A (en) * | 2018-09-26 | 2019-09-17 | 浙江大华技术股份有限公司 | A kind of method and device of determining dead ship condition |
| CN109255336A (en) * | 2018-09-29 | 2019-01-22 | 南京理工大学 | Arrester recognition methods based on crusing robot |
| CN109409385A (en) * | 2018-10-16 | 2019-03-01 | 南京鑫和汇通电子科技有限公司 | A kind of pointer instrument automatic identifying method |
| CN109559300A (en) * | 2018-11-19 | 2019-04-02 | 上海商汤智能科技有限公司 | Image processing method, electronic equipment and computer readable storage medium |
| US20210224998A1 (en) * | 2018-11-23 | 2021-07-22 | Tencent Technology (Shenzhen) Company Limited | Image recognition method, apparatus, and system and storage medium |
| CN109815868A (en) * | 2019-01-15 | 2019-05-28 | 腾讯科技（深圳）有限公司 | A kind of image object detection method, device and storage medium |
| CN109766878A (en) * | 2019-04-11 | 2019-05-17 | 深兰人工智能芯片研究院（江苏）有限公司 | A kind of method and apparatus of lane detection |
| CN110146121A (en) * | 2019-05-24 | 2019-08-20 | 安徽扬远信息科技有限公司 | A kind of instrument board Performance Evaluation control system |
| CN110222780A (en) * | 2019-06-12 | 2019-09-10 | 北京百度网讯科技有限公司 | Object detecting method, device, equipment and storage medium |
| CN110378258A (en) * | 2019-07-04 | 2019-10-25 | 上海眼控科技股份有限公司 | A kind of vehicle seat information detecting method and equipment based on image |
| CN110517259A (en) * | 2019-08-30 | 2019-11-29 | 北京百度网讯科技有限公司 | A kind of detection method, device, equipment and the medium of product surface state |
| CN110991305A (en) * | 2019-11-27 | 2020-04-10 | 厦门大学 | Airplane detection method under remote sensing image and storage medium |
| CN111401424A (en) * | 2020-03-10 | 2020-07-10 | 北京迈格威科技有限公司 | Target detection method, device and electronic system |
| CN111239158A (en) * | 2020-03-13 | 2020-06-05 | 苏州鑫睿益荣信息技术有限公司 | Automobile instrument panel detection system and detection method based on machine vision |
| CN111523414A (en) * | 2020-04-13 | 2020-08-11 | 绍兴埃瓦科技有限公司 | Face recognition method and device, computer equipment and storage medium |
| CN112639872A (en) * | 2020-04-24 | 2021-04-09 | 华为技术有限公司 | Method and device for difficult mining in target detection |
| CN111709416A (en) * | 2020-05-15 | 2020-09-25 | 珠海亿智电子科技有限公司 | License plate positioning method, device and system and storage medium |
| CN111832557A (en) * | 2020-06-04 | 2020-10-27 | 北京百度网讯科技有限公司 | Power grid inspection method and device, electronic equipment and storage medium |
| CN111768392A (en) * | 2020-06-30 | 2020-10-13 | 创新奇智(广州)科技有限公司 | Target detection method and device, electronic equipment and storage medium |
| CN112580657A (en) * | 2020-12-23 | 2021-03-30 | 陕西天诚软件有限公司 | Self-learning character recognition method |
Non-Patent Citations (1)
| Title |
|---|
| 李健 等: "汽车里程表读数自动识别***的研究与实现", 《新型工业化》, vol. 6, no. 4, pages 1 - 7 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114241011A (en) * | 2022-02-22 | 2022-03-25 | 阿里巴巴达摩院(杭州)科技有限公司 | Target detection method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113361413B (en) | 2024-06-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11100320B2 (en) | Image recognition method and apparatus | |
| CN113936198B (en) | Low-beam laser radar and camera fusion method, storage medium and device | |
| CN108428248B (en) | Vehicle window positioning method, system, equipment and storage medium | |
| CN103383732B (en) | Image processing method and device | |
| CN111723865B (en) | Method, apparatus and medium for evaluating performance of image recognition model and attack method | |
| CN112906823B (en) | Target object recognition model training method, recognition method and recognition device | |
| CN113240716B (en) | Twin network target tracking method and system with multi-feature fusion | |
| CN110084230B (en) | Image-based vehicle body direction detection method and device | |
| CN111461113A (en) | Large-angle license plate detection method based on deformed plane object detection network | |
| CN111680546A (en) | Attention detection method, attention detection device, electronic equipment and storage medium | |
| CN113361413B (en) | Mileage display area detection method, device, equipment and storage medium | |
| JP2015103188A (en) | Image analysis device, image analysis method, and image analysis program | |
| CN116964588A (en) | Target detection method, target detection model training method and device | |
| CN115861981A (en) | Driver fatigue behavior detection method and system based on video attitude invariance | |
| CN110097108A (en) | Recognition methods, device, equipment and the storage medium of non-motor vehicle | |
| CN111860512B (en) | Vehicle identification method, device, electronic equipment and computer readable storage medium | |
| CN111753812A (en) | Text recognition method and equipment | |
| CN111753719A (en) | Fingerprint identification method and device | |
| CN114120305B (en) | Training method of text classification model, and text content recognition method and device | |
| CN117274132A (en) | Multi-scale self-encoder generation method, electronic device and storage medium | |
| CN112559340A (en) | Picture testing method, device, equipment and storage medium | |
| CN112559342A (en) | Method, device and equipment for acquiring picture test image and storage medium | |
| CN106909936B (en) | Vehicle detection method based on double-vehicle deformable component model | |
| US20230326237A1 (en) | Method, device, electronic device and non-transitory storage medium for fingerprint comparison | |
| US20230401691A1 (en) | Image defect detection method, electronic device and readable storage medium |
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 |