Image recognition-based traffic signboard accuracy testing method and system

Abstract

The application provides a traffic signboard accuracy testing method and system based on image recognition, which comprises the following steps: S1, based on high-precision map positioning, obtaining traffic signboard image samples to be tested under various scene types; S2, searching for corresponding traffic signboard true value data in a high-precision map database based on the position information of the image samples to be tested, and outputting a true value data table according to the searched traffic signboard true value data; S3, performing image recognition on each frame of image sample to be tested respectively, and outputting a test sample table according to the obtained traffic signboard recognition data; and S4, comparing the test sample table and the true value data table based on the id number of the traffic signboard, and obtaining the test accuracy of various image samples to be tested. The application solves the problem of low efficiency of image recognition traffic signboard accuracy testing, and provides preliminary scene analysis results for algorithm optimization.

Description

Technical Field

[0001] This invention relates to the field of high-precision map technology, and more specifically, to a method, system, electronic device, and storage medium for testing the accuracy of traffic signs based on image recognition. Background Technology

[0002] With the gradual application of autonomous driving functions in urban areas, urban traffic signs can provide intelligent vehicles with speed limit information and location information for vehicle positioning. Currently, image-based AI recognition algorithms are increasingly widely used in urban mapping, quickly identifying ground features for high-precision map creation. Compared to highway traffic signs, urban traffic signs are more varied (e.g., speed limit, low speed limit, speed limit lifted, height limit, weight limit, etc.), more numerous, and located in more complex scenarios. Traditional methods for testing the accuracy of traffic sign recognition mainly rely on manual visual inspection of images to check for missing or incorrect recognitions, resulting in low testing efficiency and failing to meet the testing needs of large-scale algorithm use. Therefore, it is necessary to research a testing method that can quickly test the recognition accuracy of the algorithm and provide preliminary scene analysis results for iterative optimization of image recognition algorithms. Summary of the Invention

[0003] This invention addresses the technical problems existing in the prior art by providing a method, system, electronic device, and storage medium for testing the accuracy of traffic signs based on image recognition, thereby improving the efficiency of traffic sign accuracy testing.

[0004] According to a first aspect of the present invention, a method for testing the accuracy of traffic signs based on image recognition is provided, comprising:

[0005] S1, based on high-precision map positioning, acquires the image samples of traffic signs in various scene types;

[0006] S2, based on the location information of the image sample to be tested, search for the corresponding traffic sign true value data in the high-precision map database, and output the true value data table according to the searched traffic sign true value data;

[0007] S3, perform image recognition on each frame of the image sample to be tested, and output the test sample table based on the obtained traffic sign recognition data;

[0008] S4. Based on the ID number of the traffic sign, compare the test sample table and the truth data table to obtain the test accuracy of various types of image samples to be tested.

[0009] Based on the above technical solution, the present invention can also be improved as follows.

[0010] Optionally, step S1 includes:

[0011] Based on high-precision map positioning, test image samples of traffic signs in various scene types are obtained, and each frame of the test image sample includes corresponding location information;

[0012] Labels are added to the image samples to be tested according to their attribute information, which includes at least location information, camera type, image resolution, and weather.

[0013] Optionally, step S2 includes:

[0014] Based on the location information of each image sample to be tested, several map sheet numbers (areaid) and road IDs spanning the high-precision map are obtained. Based on the map sheet number (areaid) and road ID, the corresponding traffic sign truth data is extracted from the high-precision map storage database (database1). Based on the extracted traffic sign truth data, a truth sample table is output for each traffic sign, and the truth sample table is distinguished by the traffic sign ID number.

[0015] Optionally, the ground truth data of the traffic signboard includes at least the traffic signboard ID, map sheet number (areaid), geographic location information (shape), traffic signboard type, and content value.

[0016] Optionally, step S3 includes:

[0017] Image recognition is performed on each frame of the image sample to be tested to obtain the traffic sign recognition data corresponding to the image sample. Based on the obtained traffic sign recognition data, a test sample table corresponding to each traffic sign is output, and the test sample table is distinguished by the traffic sign ID number.

[0018] Optionally, step S4 includes:

[0019] Build the test database test_database and import all truth sample tables and all test sample tables;

[0020] SQL execution is performed on the test database test_database to match the truth sample table and the test sample table with the same traffic sign ID. The consistency between the traffic sign truth data in the truth sample table and the traffic sign recognition data in the test sample table is used to determine whether the image is correctly recognized, and the correctly recognized traffic sign ID and the incorrectly recognized traffic sign ID are obtained. The traffic sign IDs that are not matched are recorded as the missing traffic sign IDs.

[0021] The correct traffic sign IDs, incorrect traffic sign IDs, and missing traffic sign IDs are statistically analyzed to obtain the traffic sign recognition accuracy rate.

[0022] Based on the label information of the image sample corresponding to the correctly identified traffic sign ID, the recognition accuracy of traffic signs for each scene type is statistically analyzed.

[0023] Optionally, step S4, the process of determining whether the image is correctly recognized, includes:

[0024] Set preset distance and height thresholds, and obtain test sample tables and ground truth sample tables matched according to traffic sign IDs;

[0025] Calculate the distance L1 between the traffic sign projection onto the ground in the test sample table and the distance L2 between the traffic sign projection onto the ground in the true sample table. Subtract the distance L1 from the distance L2 and compare the difference with the distance threshold. If the distance difference is less than the distance threshold, proceed to the next step; otherwise, determine that the image has not been correctly recognized.

[0026] Calculate the height dimension H1 of the traffic sign in the test sample table and the height dimension H2 of the traffic sign in the true value sample table respectively. Subtract the height dimension H1 from the height dimension H2 and compare the height difference with the height threshold. If the height difference is less than the height threshold, proceed to the next step of judgment; otherwise, determine that the image is not correctly recognized.

[0027] Compare the type and content of the traffic signs in the test sample table and the true sample table respectively. If the type and content are the same, the image recognition is considered correct; otherwise, the image is considered not recognized correctly.

[0028] According to a second aspect of the present invention, a traffic sign accuracy testing system based on image recognition is provided, comprising:

[0029] The sample acquisition module, based on high-precision map positioning, acquires test image samples of traffic signs in various scene types;

[0030] The truth value acquisition module searches for the corresponding traffic sign truth value data in the high-precision map database based on the location information of the image sample to be tested, and outputs a truth value data table based on all the traffic sign truth value data found.

[0031] The image recognition module performs image recognition on each frame of the image sample to be tested, and outputs a test sample table based on the traffic sign recognition data of each frame.

[0032] The comparison module compares the test sample table and the truth data table based on the ID number of the traffic sign to obtain the test accuracy of various types of image samples to be tested.

[0033] According to a third aspect of the present invention, an electronic device is provided, including a memory and a processor, wherein the processor is configured to execute a computer management program stored in the memory to implement the steps of the above-described image recognition-based traffic sign accuracy testing method.

[0034] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, on which a computer management program is stored, wherein the computer management program, when executed by a processor, implements the steps of the above-described method for testing the accuracy of traffic signs based on image recognition.

[0035] This invention provides a method, system, electronic device, and storage medium for testing the accuracy of traffic signs based on image recognition. It involves collecting test image samples of traffic signs in various scene types based on high-precision map positioning and performing image recognition on them. The method constructs a scene database of test image samples using the image recognition results and the ground truth information of traffic signs in the high-precision map database corresponding to each test image sample. By comparing the image recognition results of traffic signs at the same location in the high-precision map with the ground truth information of the traffic signs, this invention solves the problems of insufficient test samples and low testing efficiency for image recognition traffic signs, and provides preliminary scene analysis results for algorithm optimization. Attached Figure Description

[0036] Figure 1 A flowchart of a traffic sign accuracy testing method based on image recognition provided by the present invention;

[0037] Figure 2 A flowchart of a traffic sign accuracy testing method based on image recognition for a specific implementation scenario;

[0038] Figure 3 A block diagram of a traffic sign accuracy testing system based on image recognition provided by the present invention;

[0039] Figure 4 A schematic diagram of the hardware structure of a possible electronic device provided by the present invention;

[0040] Figure 5 This is a schematic diagram of the hardware structure of a possible computer-readable storage medium provided by the present invention. Detailed Implementation

[0041] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0042] Figure 1 A flowchart of a traffic sign accuracy testing method based on image recognition provided by this invention is shown below. Figure 1 As shown, the method includes:

[0043] S1, based on high-precision map positioning, acquires the image samples of traffic signs in various scene types;

[0044] S2, based on the location information of the image sample to be tested, search for the corresponding traffic sign true value data in the high-precision map database, and output the true value data table according to the searched traffic sign true value data;

[0045] S3, perform image recognition on each frame of the image sample to be tested, and output the test sample table based on the obtained traffic sign recognition data;

[0046] S4. Based on the ID number of the traffic sign, compare the test sample table and the truth data table to obtain the test accuracy of various types of image samples to be tested.

[0047] Understandably, given the shortcomings in the background technology, this invention proposes an image recognition-based method for testing the accuracy of traffic signs. This method involves collecting test image samples of traffic signs in various scene types based on high-precision map positioning and performing image recognition. It then constructs a scene database of test image samples using the image recognition results and the ground truth information of traffic signs in the high-precision map database corresponding to each test image sample. By comparing the image recognition results of traffic signs at the same location on the high-precision map with the ground truth information of the traffic signs, this method solves the problems of limited test samples and low testing efficiency for image recognition accuracy of traffic signs, and provides preliminary scene analysis results for algorithm optimization.

[0048] In one possible embodiment, step S1 includes:

[0049] Based on high-precision map positioning, test image samples of traffic signs in various scene types are obtained, and each frame of the test image sample includes corresponding location information;

[0050] Labels are added to the image samples to be tested according to their attribute information, which includes at least location information (region), camera type, image resolution, and weather.

[0051] Understandably, for the collected traffic sign images of various scene types, the scene type can be classified through their label information, which facilitates the subsequent statistical analysis of the image recognition accuracy under each scene type.

[0052] In one possible embodiment, step S2 includes:

[0053] Based on the location information of each image sample to be tested, several map sheet numbers (areaid) and road IDs spanning the high-precision map are obtained. Based on the map sheet number (areaid) and road ID, the corresponding traffic sign truth data is extracted from the high-precision map storage database (database1). Based on the extracted traffic sign truth data, a truth sample table is output for each traffic sign, and the truth sample table is distinguished by the traffic sign ID number.

[0054] It is understandable that by mapping the image sample to the map sheet number (areaid) and road ID in the high-precision map, the traffic sign information in the corresponding map sheet of the high-precision map storage database (database1) can be accurately retrieved. This information is the pre-stored true value data of the traffic sign, and each traffic sign has a unique ID number.

[0055] In one possible embodiment, the traffic sign truth data includes at least the traffic sign ID, map sheet number (areaid), geographic location information (shape), traffic sign type, and content value.

[0056] In one possible embodiment, step S3 includes:

[0057] Image recognition is performed on each frame of the image sample to be tested to obtain the traffic sign recognition data corresponding to the image sample. Based on the obtained traffic sign recognition data, a test sample table corresponding to each traffic sign is output, and the test sample table is distinguished by the traffic sign ID number.

[0058] It is understandable that the traffic sign ID is the identification information of the traffic sign. Each traffic sign has a unique ID. Based on the ID of the traffic sign, the matching relationship between the test sample table and the true sample table can be found.

[0059] In one possible embodiment, such as Figure 2 As shown, step S4 includes:

[0060] Build the test database test_database and import all truth sample tables and all test sample tables;

[0061] SQL execution is performed on the test database test_database to match the truth sample table and the test sample table with the same traffic sign ID. The consistency between the traffic sign truth data in the truth sample table and the traffic sign recognition data in the test sample table is used to determine whether the image is correctly recognized, and the correctly recognized traffic sign ID and the incorrectly recognized traffic sign ID are obtained. The traffic sign IDs that are not matched are recorded as the missing traffic sign IDs.

[0062] The correct traffic sign IDs, incorrect traffic sign IDs, and missing traffic sign IDs are statistically analyzed to obtain the traffic sign recognition accuracy rate.

[0063] Based on the label information of the image sample corresponding to the correctly identified traffic sign ID, the recognition accuracy of traffic signs for each scene type is statistically analyzed.

[0064] It is understandable that the test database test_database is built by using all truth sample tables and all test sample tables, and the truth sample tables and test sample tables can be continuously supplemented in the future. By iteratively optimizing the test database through multiple tests, the accuracy and efficiency of the tests can be improved.

[0065] In one possible embodiment, step S4, the process of determining whether the image is correctly recognized, includes:

[0066] Set preset distance and height thresholds, and obtain test sample tables and ground truth sample tables matched according to traffic sign IDs;

[0067] Calculate the distance L1 between the traffic sign projection onto the ground in the test sample table and the distance L2 between the traffic sign projection onto the ground in the true sample table. Subtract the distance L1 from the distance L2 and compare the difference with the distance threshold. If the distance difference is less than the distance threshold, proceed to the next step; otherwise, determine that the image has not been correctly recognized.

[0068] Calculate the height dimension H1 of the traffic sign in the test sample table and the height dimension H2 of the traffic sign in the true value sample table respectively. Subtract the height dimension H1 from the height dimension H2 and compare the height difference with the height threshold. If the height difference is less than the height threshold, proceed to the next step of judgment; otherwise, determine that the image is not correctly recognized.

[0069] Compare the type and content of the traffic signs in the test sample table and the true sample table respectively. If the type and content are the same, the image recognition is considered correct; otherwise, the image is considered not recognized correctly.

[0070] Understandably, testing accuracy can be improved by selecting test samples and ground truth samples that meet the testing requirements for both the distance and height of traffic signs.

[0071] To better understand the technical solution of the present invention, the following is an example. Figure 2 The following example illustrates a specific implementation scenario.

[0072] 1. Preparation of sample data for the image to be tested

[0073] 1) Prepare sample image data of various traffic signs along the vehicle's journey. The locations for collecting the sample image data are covered by a high-precision map. Simultaneously, the location information of the sampling points can be automatically recorded. The collected sample data will be named log1, log2, log3, etc.

[0074] 2) Classify the scene according to the location information (region), camera type, image resolution, weather, etc. of the image sample data to be tested, such as: log1(Wuhan, camera a, resolution 1920*1080, sunny, etc.).

[0075] 2. Preparation of test database and ground truth sample data

[0076] Based on the acquisition trajectory of the image sample data to be tested, the map sheet number (areaid) and road ID corresponding to each acquisition point and the map it crosses are obtained (map storage is based on map sheet number). The high-precision map storage database (database1) is accessed to extract high-precision map traffic sign data (id, areaid, shape, type, content, etc.) as ground truth data for the traffic signs. Here, the traffic sign id is a unique primary key, areaid is the map sheet number, shape is the geographic location information, type is the category, and content is the value, thus obtaining the ground truth sample table `table_log1` for the traffic signs. Similarly, the ground truth sample tables `table_log2`, `table_log3`, etc., corresponding to the other traffic signs are obtained.

[0077] Build a test database named test_database and import each true value sample table into the test_database.

[0078] 3. Extract image recognition results

[0079] 1) Perform image recognition based on the sample data of the image to be tested, log1, log2, log3... respectively, and save the image recognition results in database database2.

[0080] 2) Extract the image recognition result data (id, areaid, shape, type, content, etc.) of a single traffic sign to obtain the test sample table table_log1_ai for that single traffic sign. Similarly, obtain the test sample tables table_log2_ai, table_log3_ai, etc. for the remaining traffic signs.

[0081] 3) Import all test sample tables into the test database test_database.

[0082] 4. Recognition accuracy test

[0083] 1) Set distance and height thresholds. When the distance and height in the traffic sign ground truth data and the traffic sign image recognition data are less than the corresponding thresholds, and the traffic sign type and value are the same, the image is considered to be correctly recognized. Otherwise, the image is considered not recognized or recognized incorrectly.

[0084] 2) Based on the built-in functions of the database test_database, SQL is executed on the true value sample table table_log1 and the test sample table table_log1_ai of a single traffic sign with the same ID to obtain the correctly identified traffic sign ID and the traffic sign ID that was incorrectly identified or omitted, and statistics are performed.

[0085] 3) By using the traffic sign ID and the scene label information of the corresponding image sample data log, the recognition accuracy of each category scene with different cameras, weather, resolution, sign type, etc. can be statistically obtained, providing the R&D team with more information and data support.

[0086] The log of the above-mentioned image sample data can be continuously increased, and the test database can be continuously used for iterative testing of the image recognition algorithm.

[0087] In the past, image recognition algorithms were tested through manual verification. The technical solution of this invention solves the problem of low efficiency in testing the accuracy of image recognition traffic signs by constructing a sample scene database, and provides preliminary scene analysis results for algorithm optimization.

[0088] Figure 3 A structural diagram of a traffic sign accuracy testing system based on image recognition is provided for an embodiment of the present invention, as shown below. Figure 3 As shown, a traffic sign accuracy testing system based on image recognition includes a sample acquisition module, a ground truth acquisition module, an image recognition module, and a comparison module, wherein:

[0089] The sample acquisition module, based on high-precision map positioning, acquires test image samples of traffic signs in various scene types;

[0090] The truth value acquisition module searches for the corresponding traffic sign truth value data in the high-precision map database based on the location information of the image sample to be tested, and outputs a truth value data table based on all the traffic sign truth value data found.

[0091] The image recognition module performs image recognition on each frame of the image sample to be tested, and outputs a test sample table based on the traffic sign recognition data of each frame.

[0092] The comparison module compares the test sample table and the truth data table based on the ID number of the traffic sign to obtain the test accuracy of various types of image samples to be tested.

[0093] It is understood that the traffic sign accuracy testing system based on image recognition provided by this invention corresponds to the traffic sign accuracy testing method based on image recognition provided in the foregoing embodiments. The relevant technical features of the traffic sign accuracy testing system based on image recognition can be referred to the relevant technical features of the traffic sign accuracy testing method based on image recognition, and will not be repeated here.

[0094] Please see Figure 4 , Figure 4 This is a schematic diagram illustrating an embodiment of the electronic device provided in this invention. For example... Figure 4 As shown, this embodiment of the invention provides an electronic device 400, including a memory 410, a processor 420, and a computer program 411 stored in the memory 410 and executable on the processor 420. When the processor 420 executes the computer program 411, it performs the following steps:

[0095] S1, based on high-precision map positioning, acquires the image samples of traffic signs in various scene types;

[0096] S2, based on the location information of the image sample to be tested, search for the corresponding traffic sign true value data in the high-precision map database, and output the true value data table according to the searched traffic sign true value data;

[0097] S3, perform image recognition on each frame of the image sample to be tested, and output the test sample table based on the obtained traffic sign recognition data;

[0098] S4. Based on the ID number of the traffic sign, compare the test sample table and the truth data table to obtain the test accuracy of various types of image samples to be tested.

[0099] Please see Figure 5 , Figure 5 This is a schematic diagram illustrating an embodiment of a computer-readable storage medium provided by the present invention. (See diagram below.) Figure 5 As shown, this embodiment provides a computer-readable storage medium 500 on which a computer program 511 is stored. When the computer program 511 is executed by a processor, it performs the following steps:

[0100] S1, based on high-precision map positioning, acquires the image samples of traffic signs in various scene types;

[0101] S2, based on the location information of the image sample to be tested, search for the corresponding traffic sign true value data in the high-precision map database, and output the true value data table according to the searched traffic sign true value data;

[0102] S3, perform image recognition on each frame of the image sample to be tested, and output the test sample table based on the obtained traffic sign recognition data;

[0103] S4. Based on the ID number of the traffic sign, compare the test sample table and the truth data table to obtain the test accuracy of various types of image samples to be tested.

[0104] This invention provides a method, system, electronic device, and storage medium for testing the accuracy of traffic signs based on image recognition. It involves collecting test image samples of traffic signs in various scene types based on high-precision map positioning and performing image recognition on them. The method constructs a scene database of test image samples using the image recognition results and the ground truth information of traffic signs in the high-precision map database corresponding to each test image sample. By comparing the image recognition results of traffic signs at the same location in the high-precision map with the ground truth information of the traffic signs, this invention solves the problems of insufficient test samples and low testing efficiency for image recognition accuracy of traffic signs, and provides preliminary scene analysis results for algorithm optimization.

[0105] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0106] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0107] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0108] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0109] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0110] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0111] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for testing the accuracy of traffic signs based on image recognition, characterized in that, include: S1, based on high-precision map positioning, acquires the image samples of traffic signs in various scene types; S2, based on the location information of the image sample to be tested, search for the corresponding traffic sign true value data in the high-precision map database, and output the true value data table according to the searched traffic sign true value data; S3, perform image recognition on each frame of the image sample to be tested, and output the test sample table based on the obtained traffic sign recognition data; S4. Based on the ID number of the traffic sign, compare the test sample table and the truth data table to obtain the test accuracy of various types of image samples to be tested.

2. The method for testing the accuracy of traffic signs based on image recognition according to claim 1, characterized in that, Step S1 includes: Based on high-precision map positioning, test image samples of traffic signs in various scene types are obtained, and each frame of the test image sample includes corresponding location information; Labels are added to the image samples to be tested according to their attribute information, which includes at least location information, camera type, image resolution, and weather.

3. The method for testing the accuracy of traffic signs based on image recognition according to claim 1, characterized in that, Step S2 includes: Based on the location information of each image sample to be tested, several map sheet numbers (areaid) and road IDs spanning the high-precision map are obtained. Based on the map sheet number (areaid) and road ID, the corresponding traffic sign truth data is extracted from the high-precision map storage database (database1). Based on the extracted traffic sign truth data, a truth sample table is output for each traffic sign, and the truth sample table is distinguished by the traffic sign ID number.

4. The method for testing the accuracy of traffic signs based on image recognition according to claim 3, characterized in that, The ground truth data of the traffic signboard includes at least the traffic signboard ID, map sheet number (areaid), geographic location information (shape), traffic signboard type, and content value.

5. The method for testing the accuracy of traffic signs based on image recognition according to claim 1, characterized in that, Step S3 includes: Image recognition is performed on each frame of the image sample to be tested to obtain the traffic sign recognition data corresponding to the image sample. Based on the obtained traffic sign recognition data, a test sample table corresponding to each traffic sign is output, and the test sample table is distinguished by the traffic sign ID number.

6. The method for testing the accuracy of traffic signs based on image recognition according to claim 2, characterized in that, Step S4 includes: Build the test database test_database and import all truth sample tables and all test sample tables; SQL execution is performed on the test database test_database to match the truth sample table and the test sample table with the same traffic sign ID. The consistency between the traffic sign truth data in the truth sample table and the traffic sign recognition data in the test sample table is used to determine whether the image is correctly recognized, and the correctly recognized traffic sign ID and the incorrectly recognized traffic sign ID are obtained. The traffic sign IDs that are not matched are recorded as the missing traffic sign IDs. The correct traffic sign IDs, incorrect traffic sign IDs, and missing traffic sign IDs are statistically analyzed to obtain the traffic sign recognition accuracy rate. Based on the label information of the image sample corresponding to the correctly identified traffic sign ID, the recognition accuracy of traffic signs for each scene type is statistically analyzed.

7. The method for testing the accuracy of traffic signs based on image recognition according to claim 6, characterized in that, Step S4, the process of determining whether the image is correctly recognized, includes: Set preset distance and height thresholds, and obtain test sample tables and ground truth sample tables matched according to traffic sign IDs; Calculate the distance L1 between the traffic sign projection onto the ground in the test sample table and the distance L2 between the traffic sign projection onto the ground in the true sample table. Subtract the distance L1 from the distance L2 and compare the difference with the distance threshold. If the distance difference is less than the distance threshold, proceed to the next step; otherwise, determine that the image has not been correctly recognized. Calculate the height dimension H1 of the traffic sign in the test sample table and the height dimension H2 of the traffic sign in the true value sample table respectively. Subtract the height dimension H1 from the height dimension H2 and compare the height difference with the height threshold. If the height difference is less than the height threshold, proceed to the next step of judgment; otherwise, determine that the image is not correctly recognized. Compare the type and content of the traffic signs in the test sample table and the true sample table respectively. If the type and content are the same, the image recognition is considered correct; otherwise, the image is considered not recognized correctly.

8. A traffic sign accuracy testing system based on image recognition, characterized in that, include: The sample acquisition module, based on high-precision map positioning, acquires test image samples of traffic signs in various scene types; The truth value acquisition module searches for the corresponding traffic sign truth value data in the high-precision map database based on the location information of the image sample to be tested, and outputs a truth value data table based on all the traffic sign truth value data found. The image recognition module performs image recognition on each frame of the image sample to be tested, and outputs a test sample table based on the traffic sign recognition data of each frame. The comparison module compares the test sample table and the truth data table based on the ID number of the traffic sign to obtain the test accuracy of various types of image samples to be tested.

9. An electronic device, characterized in that, The system includes a memory and a processor, wherein the processor is used to execute computer management programs stored in the memory to implement the steps of the image recognition-based traffic sign accuracy testing method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, It stores a computer management program, which, when executed by a processor, implements the steps of the image recognition-based traffic sign accuracy testing method as described in any one of claims 1-7.

Images