Traffic sign detection method and system based on image recognition

Abstract

The invention relates to the technical field of traffic sign detection, and particularly discloses a traffic sign detection method and system based on image recognition, and the system comprises a motion platform which can flexibly move in a specified detection area, and generally employs a vehicle as a remote sensing platform; the data acquisition end is mounted on the motion platform and is responsible for acquiring real-time position information of the motion platform and related data of traffic signs; the preprocessing module is used for analyzing the position information, collected in real time, of the motion platform, controlling a data collection end to collect traffic sign data according to the position information of the motion platform, and preprocessing and analyzing the collected data; the traffic sign can be rapidly and fully detected based on a positioning technology, an image recognition technology and a three-dimensional scanning technology, comprehensive analysis is carried out according to a detection result and traffic sign historical maintenance data, and different suggestions for traffic sign maintenance in a specified detection area are generated.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic sign detection, and in particular to a traffic sign detection method and system based on image recognition. Background Art

[0002] Road traffic signs are road facilities set up by national transportation departments on both sides of the road. They use signs with text or special symbols to convey guidance, restrictions, warnings or instruction information. They play the role of directing vehicles and conveying traffic regulations. However, in actual use, traffic signs may be unclear, lost, blocked, etc. due to various reasons. These problems may bring safety hazards to road traffic. Therefore, staff are required to regularly inspect traffic signs and promptly maintain traffic signs in abnormal conditions to reduce safety hazards in road traffic.

[0003] In the prior art, the process of detecting traffic signs is mostly carried out by staff regularly patrolling the roads to check whether the road traffic signs are intact and record any damage or missing conditions. The patrol results are then fed back to the maintenance department, and staff are arranged to repair and maintain abnormal road traffic signs. This road sign detection method is simple and direct to operate, but a lot of manpower and material resources are wasted during the detection process, which is inefficient, and the detection results may also be inaccurate due to deviations or errors caused by the subjectivity of the staff. Summary of the invention

[0004] The purpose of the present invention is to provide a traffic sign detection method and system based on image recognition to solve the following technical problems:

[0005] How to accurately and quickly detect traffic signs based on image recognition technology.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A traffic sign detection system based on image recognition, the system comprising:

[0008] Motion platform, which can move flexibly within the designated detection area;

[0009] The data collection terminal is installed on the motion platform and is responsible for collecting the real-time location information of the motion platform and relevant data of traffic signs;

[0010] The preprocessing module is used to analyze the real-time collected motion platform position information, control the collection of traffic sign data, and perform preprocessing analysis on the collected data to determine whether the traffic sign status is abnormal;

[0011] A data storage module is used to store traffic sign related data information and historical maintenance record data within a specified detection area;

[0012] Comprehensive analysis module, used to conduct comprehensive analysis of all traffic signs in the designated detection area and generate corresponding warning signals;

[0013] The early warning module executes the corresponding early warning signal.

[0014] Furthermore, the data collection terminal includes:

[0015] A state acquisition module, comprising a laser scanner and a CCD camera, wherein the laser scanner is used to acquire spatial data of traffic signs, and the CCD camera is used to acquire image data of traffic signs;

[0016] The position acquisition module is used to collect the position information of the motion platform.

[0017] Furthermore, the process of preprocessing and analyzing the collected motion platform position data and traffic sign data includes:

[0018] Step 1: Analyze the position information of the motion platform in real time. When the motion platform moves to the preset detection point area, control the laser scanner and CCD camera to collect data of the traffic signs corresponding to the detection point area.

[0019] Step 2: Analyze the traffic sign spatial data collected by the laser scanner to determine whether the traffic sign spatial state is abnormal;

[0020] Step 3: Analyze the traffic sign image data collected by the CCD camera to determine whether the display status of the traffic sign is abnormal.

[0021] Furthermore, the process of analyzing the traffic sign spatial data collected by the laser scanner includes:

[0022]

[0023] The spatial data deviation coefficient Q of traffic signs is obtained by analyzing and calculating formula (1)-(2) s ;

[0024] Among them, E s is the displacement parameter of the traffic sign, ΔR s is the tilt angle of the traffic sign, R std is the allowable tilt angle of the traffic sign, ω1 and ω2 are the first weight coefficients, N is the number of analysis reference points preset on the traffic sign, i∈[1,N], L i is the distance between the i-th analysis reference point and the preset comparison reference point, L i,stdis the standard distance between the i-th analysis reference point and the preset comparison reference point;

[0025] The spatial data deviation coefficient Q of the traffic sign s Critical coefficient Q of deviation from preset spatial data risk Make a comparison;

[0026] If Q s ≥Q risk , it is determined that the spatial state of the traffic sign is abnormal, and the traffic sign is marked as a first abnormal state;

[0027] If Q s risk , proceed to step 3.

[0028] Furthermore, the process of analyzing the traffic sign image data collected by the CDD camera includes:

[0029]

[0030] The bad evaluation parameter Z of the traffic sign is obtained by analyzing and calculating formula (3): s ;

[0031] Where M is the number of factors affecting traffic sign quality, j∈[1,M], W j is the bad evaluation value of the jth quality influencing factor, is the weight coefficient corresponding to the jth quality influencing factor;

[0032] The bad evaluation parameter Z of the traffic sign s and the preset critical parameter Z risk Make a comparison;

[0033] If Z s ≥Z risk , it is determined that the traffic sign display state is abnormal, and the traffic sign is marked as a second abnormal state.

[0034] Furthermore, the process of comprehensively analyzing all traffic signs within the designated detection area includes:

[0035]

[0036] The evaluation parameter A of all traffic sign status in the specified detection area is obtained by analyzing and calculating formulas (4)-(6): s ;

[0037] Among them, C k is the average value of abnormal parameters of traffic signs in normal state, m is the number of traffic signs marked as the second abnormal state, n is the number of traffic signs marked as the first abnormal state, B is the number of all traffic signs in the specified detection area, D​s is the average usage time of all traffic signs in the specified detection area, D std is the standard usage time of the traffic sign, τ1, τ2, τ3 are the second weight coefficients, k∈[1,Bnm], Q k is the spatial data deviation coefficient of the kth normal traffic sign, Z k is the bad evaluation parameter of the kth normal traffic sign, σ1 and σ2 are the third weight coefficients, l∈[1,B], D l is the usage time of the lth traffic sign;

[0038] All traffic sign status evaluation parameters A s And the preset status evaluation warning parameter A risk Make a comparison;

[0039] If A s ≥A risk , generating a first warning signal;

[0040] If A s risk , generating a second warning signal.

[0041] Furthermore, the system further comprises:

[0042] The detection control module reads and analyzes the historical maintenance record data of traffic signs in the specified detection area and generates a recommended time for executing the traffic sign detection task.

[0043] Furthermore, the process of generating a recommended duration for executing the traffic sign detection task includes:

[0044]

[0045] The recommended execution time T of the traffic sign detection task is obtained by analyzing and calculating formula (7): s ;

[0046] Among them, G std is the preset number of traffic sign changes within the specified detection area, G s is the average number of traffic signs changed in a single time in the specified detection area, X std X is the preset number of times to detect traffic signs in the specified detection area. s T is the cumulative number of traffic sign detections in the specified detection area. std The standard duration for performing traffic sign detection tasks is preset.

[0047] A traffic sign detection method based on image recognition, the method being used in a traffic sign detection system based on image recognition, the method comprising:

[0048] ​S1, control the motion platform to move within the specified detection area, and collect the real-time position information of the motion platform through the data acquisition terminal;

[0049] S2. Analyze the position information of the motion platform through the preprocessing module, and control the data collection end to collect traffic sign related data;

[0050] S3, preprocessing and analyzing the collected traffic sign related data through the preprocessing module to determine whether the status of each traffic sign is abnormal;

[0051] S4. Perform a comprehensive analysis on all traffic signs in the designated detection area through a comprehensive analysis module to generate a corresponding warning signal.

[0052] Beneficial effects of the present invention:

[0053] (1) The present invention analyzes the collected data through the preprocessing module to determine whether the corresponding traffic sign status is abnormal, and stores the analysis results and the collected data in the data storage module. Then, the comprehensive analysis module reads the data information stored in the data storage module, performs a comprehensive analysis on all traffic signs in the specified detection area, and generates a corresponding warning signal according to the analysis result. The warning module executes the warning signal. Different warning signals are different suggestions for the maintenance of traffic signs in the specified detection area. Traffic signs can be quickly and fully detected based on positioning technology, image recognition and three-dimensional scanning technology, avoiding possible mistakes or subjective deviations in the manual detection process, saving manpower and material resources. In addition, a comprehensive analysis is performed based on the detection results and the historical maintenance data of traffic signs to generate different suggestions for the maintenance of traffic signs in the specified detection area, which is convenient for the detection and maintenance process of traffic signs, and further saves manpower and material resources.

[0054] (2) The present invention controls the laser scanner and CCD camera to collect data on the traffic signs corresponding to the detection point area through real-time analysis of the position information of the motion platform, thereby effectively avoiding the problem of missed detection that may occur during the traffic sign detection process and ensuring the detection of all traffic signs. Then, the data scanned by the laser scanner is modeled and analyzed to determine whether the corresponding traffic signs have any abnormalities in space. Then, the traffic signs that do not have any abnormalities in space are further analyzed. By performing convolutional neural network analysis on the image information collected by the CCD camera, it is determined whether the corresponding traffic signs have any abnormalities in display. This preprocessing analysis method can quickly and effectively analyze and determine the status of traffic signs, classify and mark abnormal traffic signs, and facilitate subsequent maintenance of traffic signs. In addition, the spatial status of traffic signs is first analyzed, which can reduce the computing power required in the traffic sign analysis process to a certain extent.

[0055] (3) The present invention conducts a comprehensive analysis of the quality of all traffic signs to determine whether the traffic signs in the detection area need to be uniformly replaced. If not, targeted maintenance is performed based on the abnormal status of the signs. This can further save manpower and material resources during the traffic sign detection process. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The present invention will be further described below in conjunction with the accompanying drawings.

[0057] Figure 1 It is a schematic block diagram and step flow chart of a traffic sign detection system based on image recognition proposed by the present invention;

[0058] Figure 2 This is a flow chart of the steps of a traffic sign detection method based on image recognition proposed by the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] See also Figure 1 As shown, in one embodiment, a traffic sign detection system based on image recognition is provided, the system comprising:

[0061] Motion platform, which can move flexibly within the designated detection area, usually uses a vehicle as a remote sensing platform;

[0062] The data collection terminal is installed on the motion platform and is responsible for collecting the real-time location information of the motion platform and relevant data of traffic signs;

[0063] The preprocessing module is used to analyze the real-time collected motion platform position information, control the data collection terminal to collect traffic sign data according to the motion platform position information, and preprocess and analyze the collected data to determine whether the traffic sign status is abnormal;

[0064] A data storage module is used to store traffic sign related data information and historical maintenance record data within a specified detection area;

[0065] Comprehensive analysis module, used to conduct comprehensive analysis of all traffic signs in the designated detection area and generate corresponding warning signals;

[0066] The early warning module executes the corresponding early warning signal.

[0067] The data collection end comprises:

[0068] A state acquisition module, comprising a laser scanner and a CCD camera, wherein the laser scanner is used to acquire spatial data of traffic signs, and the CCD camera is used to acquire image data of traffic signs;

[0069] The position acquisition module is used to collect the position information of the motion platform, generally using GNSS technology.

[0070] Through the above technical scheme, this embodiment provides a traffic sign detection system based on image recognition, the system includes a motion platform that can move flexibly in a specified detection area, the data acquisition terminal and the preprocessing module are both installed on the motion platform, during the movement of the motion platform in the specified detection area, the preprocessing module continuously analyzes the motion platform position information collected by the position acquisition module, when the motion platform moves to a preset detection point area, the preprocessing module controls the state acquisition module to collect data on the traffic sign corresponding to the detection point, the collected data includes the spatial data and image data of the traffic sign, the preprocessing module analyzes the collected data, determines whether the corresponding traffic sign state is abnormal, and stores the analysis results and the collected data in the data storage module, then the comprehensive analysis module reads the data information stored in the data storage module, performs a comprehensive analysis on all traffic signs in the specified detection area, generates a corresponding warning signal according to the analysis results, and the warning module executes the warning signal, Different warning signals represent different suggestions for the maintenance of traffic signs in a designated detection area. For example, one warning signal may be a suggestion to uniformly replace all traffic signs in a designated detection area. The reason for this may be that the proportion of abnormal traffic signs in the designated detection area is high and the cumulative usage time of all traffic signs is long. Another warning signal may be a suggestion to repair and maintain all traffic signs with abnormal markings. The reason for this may be that the proportion of abnormal traffic signs in the designated detection area is low and the cumulative usage time of all traffic signs is short. This detection method can quickly and fully detect traffic signs based on positioning technology, image recognition and three-dimensional scanning technology, avoiding possible mistakes or subjective biases in the manual detection process, saving manpower and material resources, and generating different suggestions for the maintenance of traffic signs in the designated detection area based on a comprehensive analysis of the detection results and historical maintenance data of traffic signs, which is convenient for the traffic sign detection and maintenance process, and further saves manpower and material resources.

[0071] In one embodiment, the process of preprocessing and analyzing the collected position data of the motion platform and the traffic sign data includes:

[0072] Step 1: Analyze the position information of the motion platform in real time. When the motion platform moves to the preset detection point area, control the laser scanner and CCD camera to collect data of the traffic signs corresponding to the detection point area.

[0073] Step 2: Analyze the traffic sign spatial data collected by the laser scanner to determine whether the traffic sign spatial state is abnormal;

[0074] Step 3: Analyze the traffic sign image data collected by the CCD camera to determine whether the display status of the traffic sign is abnormal.

[0075] The process of analyzing the spatial data of traffic signs collected by laser scanners includes:

[0076]

[0077] The spatial data deviation coefficient Q of traffic signs is obtained by analyzing and calculating formula (1)-(2) s ;

[0078] Among them, E s is the displacement parameter of the traffic sign, ΔR s is the inclination angle of the traffic sign, which is the difference between the angle of the traffic sign when it is detected and the angle when it is installed. It can be obtained by modeling and analyzing the traffic sign after scanning it with a laser scanner. std is the allowable tilt angle of the traffic sign, which is obtained by experience. ω1 and ω2 are the first weight coefficients, which are obtained by experience. N is the number of analysis reference points preset on the traffic sign. The reference points can be preset according to the traffic sign displacement experiment. For example, a circular traffic sign generally has a certain number of reference points preset at equal intervals on its circular edge, and a rectangular traffic sign generally has a certain number of reference points preset on its rectangular edge. i∈[1,N], L i is the distance between the ith analysis reference point and the preset comparison reference point, which can be preset during the installation of the traffic sign, such as a point on the traffic sign column, a point on the base, or a point on the road centerline. i,std is the standard distance between the ith analysis reference point and the preset comparison reference point, wherein the standard distance is the initial distance between the ith analysis reference point and the preset comparison reference point obtained by scanning and modeling with a laser scanner when the traffic sign is installed;

[0079] The spatial data deviation coefficient Q of the traffic sign s Critical coefficient Q of deviation from preset spatial data risk Make a comparison;

[0080] If Q s ≥Q risk, it is determined that the spatial state of the traffic sign is abnormal, and the traffic sign is marked as a first abnormal state, where the first abnormal state indicates that the traffic sign may be displaced or deflected in space, affecting its indication effect on moving vehicles;

[0081] If Q s risk , proceed to step 3.

[0082] The process of analyzing traffic sign image data collected by the CDD camera includes:

[0083]

[0084] The bad evaluation parameter Z of the traffic sign is obtained by analyzing and calculating formula (3): s ;

[0085] Where M is the number of factors affecting the quality of traffic signs, including but not limited to clarity, reflective performance, color characteristics, etc., j∈[1,M], W j is the poor evaluation value of the jth quality influencing factor. The worse the quality corresponding to each quality influencing factor obtained by analyzing the collected traffic sign image data, the higher the corresponding poor evaluation value. Specifically, it can be obtained by analyzing, comparing and evaluating the collected traffic sign image data and the standard traffic sign image data according to the convolutional neural network. For example, the image data clarity of the standard traffic sign is 600PI, and the image data clarity of the collected traffic sign is 500PI. At this time, the poor evaluation value of the image data clarity of the traffic sign can be preset to 100. is the weight coefficient corresponding to the jth quality influencing factor, which can be obtained based on the experimental analysis of traffic sign quality influencing factors;

[0086] The bad evaluation parameter Z of the traffic sign s and the preset critical parameter Z risk Make a comparison;

[0087] If Z s ≥Z risk , it is determined that the display state of the traffic sign is abnormal, and the traffic sign is marked as a second abnormal state. The second abnormal state indicates that the traffic sign may not meet the display standards, affecting its indication effect on moving vehicles.

[0088] ​Through the above technical solution, this embodiment provides a method for preprocessing and analyzing traffic signs. Through real-time analysis of the position information of the motion platform, when the motion platform moves to the preset detection point area, the laser scanner and the CCD camera are controlled to collect data for the traffic signs corresponding to the detection point area, which can effectively avoid the problem of missed detection that may occur in the process of traffic sign detection, and ensure the detection of all traffic signs. Then, the data scanned by the laser scanner is modeled and analyzed to determine whether the corresponding traffic sign is abnormal in space. If it is abnormal, the traffic sign is marked as a first abnormal state, and then the traffic signs that are not abnormal in space are further analyzed. By performing convolutional neural network analysis on the image information collected by the CCD camera, it is determined whether the corresponding traffic sign is abnormal in display. If it is abnormal, the traffic sign is marked as a second abnormal state. This preprocessing and analysis method can quickly and effectively analyze and judge the state of traffic signs, classify and mark abnormal traffic signs, and facilitate subsequent maintenance of traffic signs. In addition, the spatial state of traffic signs is first analyzed, which can reduce the computing power requirements in the process of traffic sign analysis to a certain extent.

[0089] In one embodiment, the process of comprehensively analyzing all traffic signs in a designated detection area includes:

[0090]

[0091] The evaluation parameter A of all traffic sign status in the specified detection area is obtained by analyzing and calculating formulas (4)-(6): s ;

[0092] Among them, C k is the average value of abnormal parameters of traffic signs in normal state, m is the number of traffic signs marked as the second abnormal state, which can be obtained by statistically analyzing the number of second abnormal state markers generated in the preprocessing module, n is the number of traffic signs marked as the first abnormal state, which can be obtained by statistically analyzing the number of first abnormal state markers generated in the preprocessing module, B is the number of all traffic signs in the specified detection area, which can be obtained by statistically analyzing the number of all traffic signs in the specified detection area, and D s is the average usage time of all traffic signs in the specified detection area, D std is the standard usage time of the traffic sign, which is obtained by experience. τ1, τ2, τ3 are the second weight coefficients, which are obtained by experience. k∈[1,Bnm], Q k is the spatial data deviation coefficient of the kth normal traffic sign, Z k is the bad evaluation parameter of the kth normal traffic sign, σ1 and ρ2 are the third weight coefficients, which are obtained based on experience, l∈[1,B], D lis the usage time of the lth traffic sign, obtained according to the traffic sign usage history record;

[0093] All traffic sign status evaluation parameters A s And the preset status evaluation warning parameter A risk Make a comparison;

[0094] If A s ≥A risk , indicating that the overall quality of all traffic signs in the detection area is poor, generating the first warning signal, and recommending that all traffic signs in the detection area be uniformly replaced;

[0095] If A s risk , indicating that the overall quality of all traffic signs in the detection area is good, generating a second warning signal, and recommending targeted maintenance of the traffic signs in the detection area based on the abnormal status of the marks.

[0096] Through the above technical solution, this embodiment provides a method for comprehensively analyzing all traffic signs in a specified detection area and generating corresponding warning signals. By comprehensively analyzing the quality of all traffic signs, it is determined whether the traffic signs in the detection area need to be uniformly replaced. If not, targeted maintenance is performed based on the abnormal status of the marks. This can further save manpower and material resources during the traffic sign detection process.

[0097] The system further comprises:

[0098] The detection control module reads and analyzes the historical maintenance record data of traffic signs in the specified detection area and generates a recommended time for executing the traffic sign detection task.

[0099] The process of generating the recommended duration for executing the traffic sign detection task includes:

[0100]

[0101] The recommended execution time T of the traffic sign detection task is obtained by analyzing and calculating formula (7): s ;

[0102] Among them, G std is the preset number of traffic signs to be replaced in the specified detection area, which is obtained based on experience. s is the average number of traffic signs replaced in a single time in the specified detection area, obtained based on historical maintenance data statistics, X std is the preset number of times to detect traffic signs in the specified detection area, obtained based on experience, X s is the cumulative number of traffic sign detections in the specified detection area, obtained based on historical maintenance data statistics, T std ​The standard duration for performing traffic sign detection tasks is preset based on experience.

[0103] Through the above technical solution, this embodiment provides a method for generating a recommended duration for executing a traffic sign task in a specified detection area. Through this method, the duration of the next detection can be intelligently generated according to the status of the traffic sign. Compared with the regular detection in the prior art, the waste of manpower and material resources in the process of traffic sign detection can be further reduced. At the same time, when the overall quality of the traffic sign is reduced, the traffic sign can also be detected in a timely manner.

[0104] See also Figure 2 As shown, in one embodiment, a traffic sign detection method based on image recognition is used in a traffic sign detection system based on image recognition, and the method includes:

[0105] S1, control the motion platform to move within the specified detection area, and collect the real-time position information of the motion platform through the data acquisition terminal;

[0106] S2. Analyze the position information of the motion platform through the preprocessing module, and control the data collection end to collect traffic sign related data;

[0107] S3, preprocessing and analyzing the collected traffic sign related data through the preprocessing module to determine whether the status of each traffic sign is abnormal;

[0108] S4. Perform a comprehensive analysis on all traffic signs in the designated detection area through a comprehensive analysis module to generate a corresponding warning signal.

[0109] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A traffic sign detection system based on image recognition, characterized in that: The system comprises: Motion platform, which can move flexibly within the designated detection area; The data collection terminal is installed on the motion platform and is responsible for collecting the real-time location information of the motion platform and relevant data of traffic signs; The preprocessing module is used to analyze the real-time collected motion platform position information, control the collection of traffic sign data, and perform preprocessing analysis on the collected data to determine whether the traffic sign status is abnormal; A data storage module is used to store traffic sign related data information and historical maintenance record data within a specified detection area; Comprehensive analysis module, used to conduct comprehensive analysis of all traffic signs in the designated detection area and generate corresponding warning signals; The early warning module executes the corresponding early warning signal.

2. A traffic sign detection system based on image recognition according to claim 1, characterized in that: The data collection end comprises: A state acquisition module, comprising a laser scanner and a CCD camera, wherein the laser scanner is used to acquire spatial data of traffic signs, and the CCD camera is used to acquire image data of traffic signs; The position acquisition module is used to collect the position information of the motion platform.

3. A traffic sign detection system based on image recognition according to claim 2, characterized in that: The process of preprocessing and analyzing the collected motion platform position data and traffic sign data includes: Step 1: Analyze the position information of the motion platform in real time. When the motion platform moves to the preset detection point area, control the laser scanner and CCD camera to collect data of the traffic signs corresponding to the detection point area. Step 2: Analyze the traffic sign spatial data collected by the laser scanner to determine whether the traffic sign spatial state is abnormal; Step 3: Analyze the traffic sign image data collected by the CCD camera to determine whether the display status of the traffic sign is abnormal.

4. The traffic sign detection system based on image recognition according to claim 3, characterized in that: The process of analyzing the spatial data of traffic signs collected by laser scanners includes: The spatial data deviation coefficient Q of traffic signs is obtained by analyzing and calculating formula (1)-(2) s ; Among them, E s is the displacement parameter of the traffic sign, ΔR s is the tilt angle of the traffic sign, R std is the allowable tilt angle of the traffic sign, ω1 and ω2 are the first weight coefficients, N is the number of analysis reference points preset on the traffic sign, i∈[1,N], L i is the distance between the i-th analysis reference point and the preset comparison reference point, L i,std is the standard distance between the i-th analysis reference point and the preset comparison reference point; The spatial data deviation coefficient Q of the traffic sign s Critical coefficient Q of deviation from preset spatial data risk Make a comparison; If Q s ≥Q risk , it is determined that the spatial state of the traffic sign is abnormal, and the traffic sign is marked as a first abnormal state; If Q s risk , proceed to step 3.​ 5. The traffic sign detection system based on image recognition according to claim 4, characterized in that: The process of analyzing traffic sign image data collected by the CDD camera includes: The bad evaluation parameter Z of the traffic sign is obtained by analyzing and calculating formula (3): s ; Where M is the number of factors affecting traffic sign quality, j∈[1,M], W j is the bad evaluation value of the jth quality influencing factor, is the weight coefficient corresponding to the jth quality influencing factor; The bad evaluation parameter Z of the traffic sign s and the preset critical parameter Z risk Make a comparison; If Z s ≥Z risk , it is determined that the traffic sign display state is abnormal, and the traffic sign is marked as a second abnormal state.

6. The traffic sign detection system based on image recognition according to claim 5, characterized in that: The process of comprehensive analysis of all traffic signs within a specified detection area includes: The evaluation parameter A of all traffic sign status in the specified detection area is obtained by analyzing and calculating formulas (4)-(6): s ; Among them, C k is the average value of abnormal parameters of traffic signs in normal state, m is the number of traffic signs marked as the second abnormal state, n is the number of traffic signs marked as the first abnormal state, B is the number of all traffic signs in the specified detection area, D s is the average usage time of all traffic signs in the specified detection area, D std is the standard usage time of the traffic sign, τ1, τ2, τ3 are the second weight coefficients, k∈[1,Bnm], Q k is the spatial data deviation coefficient of the kth normal traffic sign, Z k is the bad evaluation parameter of the kth normal traffic sign, σ1 and σ2 are the third weight coefficients, l∈[1,B], D l is the usage time of the lth traffic sign; All traffic sign status evaluation parameters A s And the preset status evaluation warning parameter A risk Make a comparison; If A s ≥A risk , generating a first warning signal; If A s risk , generating a second warning signal.​ 7. The traffic sign detection system based on image recognition according to claim 6, characterized in that: The system further comprises: The detection control module reads and analyzes the historical maintenance record data of traffic signs in the specified detection area and generates a recommended time for executing the traffic sign detection task.

8. The traffic sign detection system based on image recognition according to claim 7, characterized in that: The process of generating the recommended duration for executing the traffic sign detection task includes: The recommended execution time T of the traffic sign detection task is obtained by analyzing and calculating formula (7): s ; Among them, G std is the preset number of traffic sign changes within the specified detection area, G s is the average number of traffic signs changed in a single time in the specified detection area, X std X is the preset number of times to detect traffic signs in the specified detection area. s T is the cumulative number of traffic sign detections in the specified detection area. std The standard duration for performing traffic sign detection tasks is preset.

9. A traffic sign detection method based on image recognition, characterized in that: The method is used in a traffic sign detection system based on image recognition according to any one of claims 1 to 8, and the method comprises: S1, control the motion platform to move within the specified detection area, and collect the real-time position information of the motion platform through the data acquisition terminal; S2. Analyze the position information of the motion platform through the preprocessing module, and control the data collection end to collect traffic sign related data; S3, preprocessing and analyzing the collected traffic sign related data through the preprocessing module to determine whether the status of each traffic sign is abnormal; S4. Perform a comprehensive analysis on all traffic signs in the designated detection area through a comprehensive analysis module to generate corresponding warning signals.

Images