Traffic illegal behavior analysis method and device, equipment and storage medium

Abstract

The invention relates to a traffic illegal behavior analysis method, device and equipment and a storage medium, and the method comprises the following steps: carrying out the real-time monitoring of a road through a camera, and obtaining the real-time monitoring video data; performing behavior analysis on the vehicle in the real-time monitoring video data to obtain a behavior analysis result; if the behavior analysis result shows that traffic illegal behaviors exist in the vehicles on the road, feature extraction is carried out on the vehicles with the traffic illegal behaviors in the real-time monitoring video data, and illegal vehicle extraction features are obtained; through a preset positioning system, based on illegal vehicle extraction features, determining a position corresponding to the illegal vehicle; encrypting the position corresponding to the illegal vehicle and the behavior analysis result corresponding to the illegal vehicle, generating an evidence chain data block, and sending the evidence chain data block to the law enforcement terminal; the problem that the urban traffic jam and potential safety hazards are aggravated due to the fact that comprehensive and effective supervision is difficult to achieve through a traditional means and many traffic illegal behaviors cannot be found and processed in time is solved.

Description

Technical Field

[0001] The present invention relates to the field of image recognition technology, and in particular to a method, device, equipment and storage medium for analyzing traffic violations. Background Art

[0002] With the acceleration of urbanization, road traffic volume has increased dramatically, and traffic violations have also increased. This not only seriously threatens people's lives and property safety, but also affects the city's traffic efficiency and environmental quality. Traditional traffic management methods mainly rely on manual patrols and fixed-point monitoring equipment. This method has great limitations in coverage, response speed and accuracy. Especially in the face of complex and changeable road conditions, traditional means are difficult to achieve comprehensive and effective supervision, resulting in many traffic violations not being discovered and handled in a timely manner, which in turn exacerbates the problems of urban traffic congestion and safety hazards.

[0003] In recent years, with the rapid development of information technology, especially the application of artificial intelligence, big data and other technologies, new ideas have been provided to solve the above problems. The use of advanced video analysis technology and intelligent algorithms can realize real-time monitoring of road conditions and automatic identification of illegal behaviors, greatly improving the level of intelligent traffic management. However, in the actual application process, how to ensure the effectiveness and reliability of these new technologies, as well as how to protect personal privacy and data security have become new challenges that need to be solved urgently. In addition, the adaptability of existing technical solutions in different regions and environments also needs to be further verified to ensure that they can be widely used in various complex traffic scenarios.

[0004] In summary, developing an efficient, accurate and safe method for analyzing traffic violations is of great significance to improving the level of urban traffic management. This method not only needs to have strong data analysis capabilities to quickly and accurately identify various traffic violations, but also needs a complete evidence collection mechanism to ensure that each violation can be effectively recorded and form a complete chain of evidence. At the same time, considering the public's high concern for privacy protection, this method must also adopt strict data encryption measures to ensure the security of information transmission, so as to win the trust and support of the public while improving law enforcement efficiency. Summary of the invention

[0005] The main purpose of the present invention is to provide a method, device, equipment and storage medium for analyzing traffic violations, which solves the problem that traditional means are difficult to achieve comprehensive and effective supervision, resulting in many traffic violations not being discovered and handled in a timely manner, thereby exacerbating urban traffic congestion and safety hazards.

[0006] To achieve the above object, the present invention provides a method for analyzing traffic violations, comprising the following steps: Real-time monitoring of roads is performed through cameras to obtain real-time monitoring video data; Performing behavior analysis on the vehicles in the real-time monitoring video data to obtain behavior analysis results; If the behavior analysis result shows that there are traffic violations among vehicles on the road, feature extraction is performed on the vehicles with traffic violations in the real-time monitoring video data to obtain illegal vehicle extraction features; Determine the location of the illegal vehicle based on the extracted features of the illegal vehicle through a preset positioning system; The location corresponding to the illegal vehicle and the behavior analysis results corresponding to the illegal vehicle are encrypted to generate an evidence chain data block and send it to the law enforcement terminal.

[0007] Furthermore, the real-time monitoring of the road by the camera to obtain real-time monitoring video data includes: The road is captured in real time by a preset high-resolution camera to obtain an original video stream; Preprocessing the original video stream to obtain road preprocessing data; Performing target extraction on the road preprocessing data to obtain extracted target data; wherein the extracted target data includes pedestrians and vehicles; Performing three-dimensional synthesis on the extracted target data by using a preset three-dimensional reconstruction technology to obtain a three-dimensional model of the traffic scene; Pedestrians and vehicles in the three-dimensional model are monitored in real time to obtain real-time monitoring video data.

[0008] Furthermore, the performing behavior analysis on the vehicle in the real-time monitoring video data to obtain the behavior analysis result includes: Performing motion vector analysis on vehicles and pedestrians in the real-time monitoring video data to obtain vehicle motion trajectory data and pedestrian trajectory data; Extracting spatiotemporal features from the vehicle motion trajectory data to obtain a vehicle behavior feature sequence; Performing a fusion analysis on the vehicle motion trajectory data and the vehicle behavior feature sequence to obtain a fusion behavior feature; wherein the fusion behavior feature includes the potential illegal tendency of the vehicle and the habitual characteristics of the driver; Based on a preset graph convolutional network algorithm, a contextual relationship analysis is performed on traffic violations between vehicles based on the fused behavior features to obtain behavior association results between vehicles; wherein the behavior association results between vehicles include relative positions, relative speeds, and interactive behaviors between vehicles; Performing pedestrian-vehicle interaction analysis on the vehicle motion trajectory data based on the pedestrian trajectory data to obtain a pedestrian-vehicle interaction analysis result; The behavior association result between the vehicles and the pedestrian-vehicle interaction analysis result are used as the behavior analysis result.

[0009] Furthermore, the extracting features of vehicles with traffic violations in the real-time monitoring video data to obtain illegal vehicle extraction features includes: Performing multi-level feature extraction on vehicles that have committed traffic violations in the real-time monitoring video data to obtain a multi-scale feature image of the vehicle; Performing region of interest screening on the multi-scale feature image of the vehicle to obtain a screening region of interest; wherein the screening region of interest includes a license plate position region and a vehicle contour region; Extracting feature descriptors from the vehicle contour area to obtain vehicle feature descriptors; Based on the vehicle feature descriptor, the vehicle with traffic violations is segmented to obtain the illegal vehicle contour features; Performing character recognition processing on the license plate position area to obtain the license plate number; The illegal vehicle outline features and the license plate number are used as illegal vehicle extraction features.

[0010] Furthermore, the determining the location of the illegal vehicle based on the extracted features of the illegal vehicle by a preset positioning system includes: Performing feature enhancement on the extracted features of the illegal vehicle to obtain enhanced vehicle features; Index encoding the enhanced vehicle feature to obtain a vehicle feature hash code; By using a preset positioning system, a relational reasoning is performed on the illegal vehicle in the three-dimensional model of the traffic scene based on the vehicle feature hash code to obtain a relative position of the vehicle; wherein the relative position of the vehicle is the relative position of the illegal vehicle in the three-dimensional model; Based on the particle filter, the relative positions of the vehicles are tracked and optimized in time series to obtain the positions corresponding to the illegal vehicles.

[0011] Furthermore, the index encoding of the enhanced vehicle feature to obtain the vehicle feature hash code includes: Performing hash mapping on the enhanced vehicle features using a preset local sensitive hashing algorithm to obtain an initial hash code; Quantizing the initial hash code to obtain a quantized feature; Using a preset Bloom filter to index encode the quantized features to obtain a preliminary feature hash code; Performing multi-level hash processing on the preliminary feature hash code to obtain a multi-level feature hash code; The multi-level feature hash codes are indexed and stored using a hash table to obtain a vehicle feature hash code.

[0012] Furthermore, the position corresponding to the illegal vehicle and the behavior analysis result corresponding to the illegal vehicle are encrypted to generate an evidence chain data block and send it to the law enforcement terminal, including: Performing data fusion on the position corresponding to the illegal vehicle and the behavior analysis results corresponding to the illegal vehicle to obtain illegal vehicle fusion data; Performing hash coding on the illegal vehicle fusion data to obtain a hash value of the illegal vehicle data; Encrypting the illegal vehicle data hash value by a preset asymmetric encryption algorithm to obtain an encrypted hash value; Using blockchain technology, the encrypted hash value is block generated to obtain an evidence chain data block; Digitally signing the evidence chain data block to obtain an evidence chain data block with a digital signature; The evidence chain data block with the digital signature is sent to the law enforcement terminal through a preset secure communication protocol.

[0013] The present invention also provides a traffic violation analysis device, comprising: A monitoring module is used to monitor the road in real time through a camera to obtain real-time monitoring video data; An analysis module, used to perform behavior analysis on the vehicles in the real-time monitoring video data to obtain behavior analysis results; An extraction module, for extracting features of vehicles with traffic violations in the real-time monitoring video data to obtain extracted features of illegal vehicles if the behavior analysis result shows that there are traffic violations in vehicles on the road; A determination module, configured to determine the location corresponding to the illegal vehicle based on the extracted features of the illegal vehicle through a preset positioning system; The encryption module is used to encrypt the location corresponding to the illegal vehicle and the behavior analysis result corresponding to the illegal vehicle, generate an evidence chain data block and send it to the law enforcement terminal.

[0014] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0015] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.

[0016] The present invention provides a method for analyzing traffic violations, comprising the following steps: monitoring a road in real time through a camera to obtain real-time monitoring video data; analyzing the behavior of vehicles in the real-time monitoring video data to obtain behavior analysis results; if the behavior analysis results show that there are traffic violations in the vehicles on the road, extracting features of the vehicles with traffic violations in the real-time monitoring video data to obtain illegal vehicle extraction features; determining the position corresponding to the illegal vehicle based on the illegal vehicle extraction features through a preset positioning system; encrypting the position corresponding to the illegal vehicle and the behavior analysis results corresponding to the illegal vehicle, generating an evidence chain data block and sending it to the law enforcement terminal; through the above technical solution, the problem that traditional means are difficult to achieve comprehensive and effective supervision, resulting in many traffic violations not being discovered and handled in time, thereby aggravating urban traffic congestion and safety hazards, is solved, and once a traffic violation is detected, the system will extract features of the illegal vehicle and determine its specific location in combination with the preset positioning system. Subsequently, the location information of the illegal vehicle and its behavior analysis results are encrypted to generate an evidence chain data block. This process ensures that each illegal behavior is traceable, forms a complete chain of evidence, and enhances the legality and persuasiveness of the law enforcement basis. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the steps of a method for analyzing traffic violations in one embodiment of the present invention; Figure 2 It is a structural block diagram of a traffic violation analysis device in one embodiment of the present invention; Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0018] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0020] like Figure 1 As shown, Figure 1 It is a schematic diagram of the steps of a method for analyzing traffic violations in one embodiment of the present invention; In one embodiment of the present invention, a method for analyzing traffic violations is provided, comprising the following steps: Step S1, real-time monitoring of the road is performed through a camera to obtain real-time monitoring video data.

[0021] Specifically, the real-time monitoring of roads by cameras to obtain real-time monitoring video data is the basis of the entire traffic violation analysis method, which involves the selection of monitoring equipment, the determination of installation locations, and the collection and transmission of video data. In order to ensure the effectiveness of monitoring, high-resolution cameras with night vision functions are usually selected and installed in key sections with large traffic flow and prone to or frequent accidents, such as intersections, around schools, or at highway entrances. These cameras can not only cover a wide range, but also clearly capture important information such as the driving status and license plate number of the vehicle. In actual operation, the camera will continuously record the video stream, convert the captured image into a digital signal, and then transmit it to the back-end server through the network. For example, on a city main road, when a car is speeding, the high-definition camera installed on the road section will capture this behavior immediately, and then upload the video clip containing the vehicle's driving trajectory, speed and other information to the cloud platform for subsequent behavior analysis. In this way, by real-time monitoring of the road, not only can potential traffic violations be discovered in a timely manner, but also reliable data support is provided for subsequent analysis and processing.

[0022] Step S2, performing behavior analysis on the vehicles in the real-time monitoring video data to obtain behavior analysis results.

[0023] Specifically, the behavior analysis of the vehicles in the real-time monitoring video data is performed to obtain the behavior analysis results. This process mainly relies on advanced image processing and pattern recognition technology. Specifically, when the video data captured by the camera is transmitted to the background system, the system will first use the target detection algorithm to accurately identify each moving vehicle from the video stream. Then, through the trajectory tracking technology, the system can continuously track the movement trajectory of each vehicle, and then analyze the vehicle's speed, direction, and whether it complies with traffic rules and other behavioral characteristics. For example, on the main road of the city, if a vehicle continues to move forward when the red light is on, the system will immediately identify this violation and mark it as a red light running event. Similarly, when a vehicle is detected speeding in a speed limit area, the system will also make a corresponding judgment. In order to improve the accuracy and efficiency of the analysis, the system will also combine historical data and machine learning models to continuously optimize the behavior analysis algorithm, so that the system can more intelligently identify various complex traffic violations. In this way, not only can the vehicle behavior on the road be monitored in real time, but also accurate behavior analysis results can be quickly generated, providing strong support for subsequent violation handling.

[0024] Step S3: If the behavior analysis result shows that there are traffic violations in the vehicles on the road, feature extraction is performed on the vehicles with traffic violations in the real-time monitoring video data to obtain illegal vehicle extraction features.

[0025] Specifically, if the behavior analysis result shows that there are traffic violations in the vehicles on the road, the feature extraction of the vehicles with traffic violations in the real-time monitoring video data is performed to obtain the extracted features of the illegal vehicles. This process is to further capture and record the specific information of the illegal vehicles in detail after confirming that the vehicle has traffic violations. For example, on the main roads of the city, when the system finds that a car runs a red light through behavior analysis, it will start the feature extraction program next. The program will select the video clip containing the illegal vehicle from the real-time monitoring video data, and use image processing technology, such as edge detection, color segmentation and other methods, to accurately extract the appearance features of the vehicle, including key information such as model, color, and license plate number. At the same time, the system will also record the time and place of the violation and the specific type of violation, such as speeding, driving in the wrong direction, occupying the emergency lane, etc. Through this comprehensive feature extraction, not only can detailed evidence support be provided for the subsequent violation processing, but also misjudgment can be effectively avoided, and the fairness and accuracy of law enforcement can be ensured. In addition, these extracted feature information will also be used as an important data source for statistical analysis of violations and optimization of traffic management strategies, thereby further improving the intelligent level of urban traffic management.

[0026] Step S4, determining the position of the illegal vehicle based on the extracted features of the illegal vehicle through a preset positioning system.

[0027] Specifically, the location corresponding to the illegal vehicle is determined based on the extracted features of the illegal vehicle through a preset positioning system, which is an important part of the traffic violation analysis method. In actual operation, when the system confirms that a certain vehicle has committed a traffic violation through video analysis, it will further use the preset positioning system to accurately locate the location of the illegal vehicle. The preset positioning system referred to here usually refers to a comprehensive positioning technology that combines multiple information sources such as the GPS global positioning system, electronic map data, and vehicle driving paths. For example, on a city's main road, when a car is identified as an illegal vehicle by the system due to speeding, the system will use the positioning system to calculate the exact geographic location of the vehicle based on the previously extracted vehicle features, such as license plate number, model, etc., combined with the location and time of the vehicle appearing in the video. This process requires not only accurate geographic information system support, but also efficient algorithms to process a large amount of real-time data to ensure the accuracy and timeliness of positioning. In this way, law enforcement agencies can obtain the specific location information of illegal vehicles, which is crucial for subsequent law enforcement actions. It can not only help law enforcement personnel quickly arrive at the scene, but also provide accurate geographical evidence for the punishment of illegal acts. In addition, this feature-based positioning method also helps reduce the possibility of misjudgment and ensures that every enforcement is based on sufficient and accurate information, thereby improving the efficiency and credibility of the entire traffic management system.

[0028] Step S5, encrypting the location corresponding to the illegal vehicle and the behavior analysis result corresponding to the illegal vehicle, generating an evidence chain data block and sending it to the law enforcement terminal.

[0029] Specifically, the location corresponding to the illegal vehicle and the behavior analysis results corresponding to the illegal vehicle are encrypted, and the evidence chain data block is generated and sent to the law enforcement terminal. This is an important step to ensure the transparency and security of the traffic violation processing process. In practical applications, when the system confirms that a car has committed a traffic violation through behavior analysis and feature extraction, and determines its specific location, it is necessary to encrypt these sensitive information to protect the security and integrity of the data. For example, on the main road of the city, if a car is identified by the system for running a red light, the system will first extract the characteristic information of the vehicle, such as license plate number, model, color, etc., and record the specific time and place of the violation. This information will then be integrated into a data packet and encrypted using the Advanced Encryption Standard (AES) or other secure encryption algorithms to ensure that the data will not be illegally intercepted or tampered with during transmission. The encrypted data packet becomes the evidence chain data block, which not only contains the detailed information of the illegal vehicle, but also includes metadata such as timestamps and geographic locations. These metadata are crucial to prove the authenticity and validity of the evidence. After the encryption is completed, the evidence chain data block will be sent to the law enforcement terminal, such as the traffic police command center or mobile law enforcement equipment. After receiving the data, the law enforcement terminal can restore the original information through decryption, ensuring that law enforcement officers can obtain accurate evidence of violations so that they can take corresponding legal measures. In addition, this encryption processing and the establishment of a chain of evidence not only improves data security, but also enhances the transparency and credibility of the entire traffic violation handling process, helping to enhance the public's trust and support for traffic management.

[0030] In a specific embodiment, the real-time monitoring of the road by a camera to obtain real-time monitoring video data includes: The road is captured in real time by a preset high-resolution camera to obtain an original video stream; Preprocessing the original video stream to obtain road preprocessing data; Performing target extraction on the road preprocessing data to obtain extracted target data; wherein the extracted target data includes pedestrians and vehicles; Performing three-dimensional synthesis on the extracted target data by using a preset three-dimensional reconstruction technology to obtain a three-dimensional model of the traffic scene; Pedestrians and vehicles in the three-dimensional model are monitored in real time to obtain real-time monitoring video data.

[0031] Specifically, first, a preset high-resolution camera is used to collect real-time video of the road to obtain an original video stream. At this stage, it is crucial to choose the right camera model and installation location. The camera needs to have high resolution, wide dynamic range, low illumination performance, and night vision to ensure that the situation on the road can be clearly captured under different lighting conditions. For example, on the main roads in the city, cameras may be installed on traffic signal poles or street lamp poles to cover multiple lanes to ensure a wide field of view without blind spots. These cameras will record videos continuously, converting real-time traffic images into digital signals to form an original video stream. The original video stream contains a lot of information, but it also brings the problem of large data volume, so the next step of preprocessing is required. Next, the original video stream is preprocessed to obtain road preprocessing data. The main purpose of preprocessing is to remove noise, stabilize the picture, adjust brightness and contrast, etc., to improve the accuracy of subsequent analysis. Preprocessing techniques include but are not limited to denoising algorithms, inter-frame difference methods, background modeling, etc. For example, the denoising algorithm can eliminate random noise in the video and make the picture clearer; the inter-frame difference rule can help detect moving objects in the video, thereby reducing background interference; and background modeling can distinguish between static background and dynamic foreground, which is convenient for subsequent target extraction. Through these preprocessing steps, the quality of the video can be significantly improved, laying a good foundation for subsequent target extraction. Next, the road preprocessing data is subjected to target extraction to obtain extracted target data. Target extraction is a very critical part of the whole process, and its goal is to accurately identify pedestrians and vehicles from the preprocessed video. Modern target extraction technology is usually based on deep learning methods, such as convolutional neural networks (CNNs), which can automatically learn and identify the characteristics of targets. For example, the system trains a deep learning model to enable it to identify different types of vehicles (such as cars, trucks, motorcycles, etc.) and pedestrians from the video. Target extraction requires not only identifying the existence of the target, but also accurately calibrating the position and size of the target. This process generates a series of bounding boxes, each of which contains an identified target. Through target extraction, the system can simplify complex traffic scenes into a series of discrete targets, preparing for subsequent three-dimensional reconstruction. Subsequently, the extracted target data is synthesized in three dimensions through a preset three-dimensional reconstruction technology to obtain a three-dimensional model of the traffic scene. Three-dimensional reconstruction technology can convert a two-dimensional video image into a three-dimensional spatial model, so that each element of the traffic scene can be accurately represented in the virtual space. Commonly used three-dimensional reconstruction methods include reconstruction based on stereo vision, structured light reconstruction, and multi-view geometric reconstruction. For example, in reconstruction based on stereo vision, the system uses multiple cameras to shoot the same scene from different angles, and infers the three-dimensional coordinates of the object by calculating the parallax between different perspectives. In multi-view geometric reconstruction, the system obtains images of the target from multiple perspectives and reconstructs the three-dimensional shape of the target through feature matching and geometric constraints.Through three-dimensional reconstruction, the system can create a realistic traffic scene model, which includes not only the location of vehicles and pedestrians, but also their dynamic information such as movement trajectory and speed. Finally, the pedestrians and vehicles in the three-dimensional model are monitored in real time to obtain real-time monitoring video data. Based on the three-dimensional model, the system can continuously track and analyze each target and monitor its behavior in real time. For example, the system can determine whether there is speeding by calculating the speed and acceleration of the vehicle; and determine whether there is jaywalking by detecting whether pedestrians cross the road outside the crosswalk. Real-time monitoring can not only provide an intuitive display of the current traffic conditions, but also provide data support for the automatic detection of traffic violations. For example, on the main roads of the city, when the system finds that a car continues to move forward when the red light is on, it will immediately mark the behavior as running a red light and record the specific time, location, license plate number and other information. This information will be further processed to eventually form a complete chain of evidence, providing strong support for subsequent law enforcement. In short, through the above series of steps, the system can extract high-quality traffic scene information from the original video stream, which not only provides a solid data foundation for the analysis of traffic violations, but also lays a technical foundation for the intelligent upgrade of urban traffic management. This process not only reflects the widespread application of modern information technology in the field of transportation, but also makes positive contributions to improving urban traffic safety and efficiency.

[0032] In a specific embodiment, the performing behavior analysis on the vehicle in the real-time monitoring video data to obtain the behavior analysis result includes: Performing motion vector analysis on vehicles and pedestrians in the real-time monitoring video data to obtain vehicle motion trajectory data and pedestrian trajectory data; Extracting spatiotemporal features from the vehicle motion trajectory data to obtain a vehicle behavior feature sequence; Performing a fusion analysis on the vehicle motion trajectory data and the vehicle behavior feature sequence to obtain a fusion behavior feature; wherein the fusion behavior feature includes the potential illegal tendency of the vehicle and the habitual characteristics of the driver; Based on a preset graph convolutional network algorithm, a contextual relationship analysis is performed on traffic violations between vehicles based on the fused behavior features to obtain behavior association results between vehicles; wherein the behavior association results between vehicles include relative positions, relative speeds, and interactive behaviors between vehicles; Performing pedestrian-vehicle interaction analysis on the vehicle motion trajectory data based on the pedestrian trajectory data to obtain a pedestrian-vehicle interaction analysis result; The behavior association result between the vehicles and the pedestrian-vehicle interaction analysis result are used as the behavior analysis result.

[0033] Specifically, first, motion vector analysis is performed on the vehicles and pedestrians in the real-time monitoring video data to obtain vehicle motion trajectory data and pedestrian trajectory data. This process mainly uses computer vision technology, especially optical flow method or background subtraction method, to analyze the motion of each pixel in the video, so as to extract the motion trajectory of vehicles and pedestrians. For example, on the main road of the city, when the camera captures a video containing multiple cars and pedestrians, the system will first perform motion vector analysis on the video, calculate the position change of each target at different time points, and form a series of motion trajectory data. These data include not only the driving path of the vehicle, but also the walking route of the pedestrian, which provides basic information for subsequent behavior analysis. Next, the spatiotemporal feature extraction of the vehicle motion trajectory data is performed to obtain a vehicle behavior feature sequence. The spatiotemporal feature extraction is to mine more meaningful information from the vehicle's motion trajectory, such as speed, acceleration, steering frequency, etc. These features can reflect the driving behavior of the vehicle and the driver's habits. For example, by analyzing the speed changes of the vehicle in different time periods, it can be determined whether there is frequent acceleration or sudden braking, which are potential traffic violations. In addition, the steering frequency can also be used to determine whether the driver changes lanes frequently, which is also an important indicator for evaluating driving habits. Through spatiotemporal feature extraction, the system can generate a series of feature sequences reflecting vehicle behavior, providing data support for subsequent fusion analysis. Then, the vehicle motion trajectory data and the vehicle behavior feature sequence are fused and analyzed to obtain fusion behavior features. The purpose of fusion analysis is to combine the vehicle's motion trajectory data and behavior feature sequence to form a more comprehensive description of vehicle behavior. For example, the system can generate a fusion behavior feature by combining the vehicle's speed, acceleration, steering frequency and other features with the vehicle's actual driving path. These fusion behavior features can not only reflect the vehicle's driving behavior, but also reveal the driver's habitual characteristics, such as whether he likes speeding, whether he changes lanes frequently, etc. In addition, the system can also predict the vehicle's potential violation tendency by analyzing the vehicle's behavior characteristics in a specific time period, such as whether speeding is likely to occur during peak traffic hours. Through fusion analysis, the system can more accurately identify potential traffic violations and provide support for subsequent contextual relationship analysis. = Next, based on the preset graph convolutional network algorithm, the contextual relationship analysis of traffic violations between vehicles is performed based on the fusion behavior features to obtain the behavior association results between vehicles. Graph convolutional network (GCN) is a deep learning model that can process graph structured data and is particularly suitable for analyzing the interactions between vehicles. For example, on a city's main road, when multiple cars appear at the same intersection in the same time period, the system uses the graph convolutional network algorithm to construct a traffic scene graph by treating each car as a node and the relative positions, relative speeds, and interactions between vehicles as edges.By performing convolution operations on the graph, the system can analyze the mutual influence between vehicles, such as whether a sudden deceleration of a vehicle will cause the vehicle behind to brake urgently, or whether two vehicles will get dangerously close when changing lanes. These behavioral association results can not only reveal the interaction patterns between vehicles, but also help identify potential traffic conflict points and provide a basis for preventing traffic accidents. =Finally, pedestrian-vehicle interaction analysis is performed on the vehicle motion trajectory data based on the pedestrian trajectory data to obtain pedestrian-vehicle interaction analysis results. Pedestrian-vehicle interaction analysis is to evaluate the safe distance and potential risks between pedestrians and vehicles. For example, on the main roads in the city, when pedestrians cross the road, the system will combine the pedestrian trajectory data and the vehicle motion trajectory data to analyze the relative position and speed between pedestrians and vehicles. By calculating the minimum safe distance between the two, the system can determine whether there is a risk of pedestrians being hit. In addition, the system can also predict the time required for pedestrians to cross the road by analyzing the walking speed of pedestrians and the driving speed of vehicles, thereby evaluating whether vehicles need to slow down to avoid. The pedestrian-vehicle interaction analysis results can not only provide a reference for the identification of traffic violations, but also help optimize traffic signal control and improve the safety of pedestrians crossing the road. = The behavior association results between the vehicles and the pedestrian-vehicle interaction analysis results are used as behavior analysis results. This process summarizes all the analysis results to form a comprehensive traffic behavior report. For example, on the main roads of the city, when the system completes the analysis of a certain video, it will generate a detailed report. The report includes not only basic information such as the driving path, speed, acceleration, etc. of each vehicle, but also the relative position, relative speed and interaction behavior between vehicles, as well as the safe distance and potential risks between pedestrians and vehicles. This report can not only provide decision-making support for traffic management departments, helping them to discover and deal with traffic violations in a timely manner, but also provide a scientific basis for urban traffic planning and safety management. Through this comprehensive behavior analysis, the system can more effectively ensure the safety and smoothness of urban traffic and improve the level of intelligent urban traffic management.

[0034] In a specific embodiment, the feature extraction of vehicles with traffic violations in the real-time monitoring video data to obtain illegal vehicle extraction features includes: Performing multi-level feature extraction on vehicles that have committed traffic violations in the real-time monitoring video data to obtain a multi-scale feature image of the vehicle; Performing region of interest screening on the multi-scale feature image of the vehicle to obtain a screening region of interest; wherein the screening region of interest includes a license plate position region and a vehicle contour region; Extracting feature descriptors from the vehicle contour area to obtain vehicle feature descriptors; Based on the vehicle feature descriptor, the vehicle with traffic violations is segmented to obtain the illegal vehicle contour features; Performing character recognition processing on the license plate position area to obtain the license plate number; The illegal vehicle outline features and the license plate number are used as illegal vehicle extraction features.

[0035] Specifically, first, multi-level feature extraction is performed on the vehicles with traffic violations in the real-time monitoring video data to obtain a multi-scale feature image of the vehicle. This process mainly uses computer vision technology, especially multi-scale analysis methods, to extract vehicle features from different scales. For example, on the main roads of the city, when the system confirms that a car is speeding through behavioral analysis, it will extract the image of the vehicle from the real-time monitoring video data. In order to ensure the comprehensiveness and accuracy of feature extraction, the system will use a multi-scale analysis method to extract vehicle features from different resolutions and scales. These features may include the overall outline of the vehicle, local details (such as lights, windows, etc.) and texture information. Through multi-level feature extraction, the system can generate a series of feature images of different scales, which provide rich information for subsequent feature screening and description. Next, the multi-scale feature image of the vehicle is screened for regions of interest to obtain a screening region of interest. The purpose of region of interest screening is to extract the part that best reflects the characteristics of the vehicle from the multi-scale feature image, mainly including the license plate position area and the vehicle contour area. For example, the system will use target detection algorithms (such as YOLO, Faster R-CNN, etc.) to identify the license plate position and contour of the vehicle. Through these algorithms, the system can quickly locate the position of the license plate and the contour area of ​​the vehicle in the multi-scale feature image. This process not only improves the efficiency of feature extraction, but also ensures that the extracted features are the most representative and recognizable. For example, on the main road of the city, when the system identifies a speeding vehicle, it will filter out the license plate position area and the vehicle contour area from the multi-scale feature image to provide a basis for subsequent feature description and recognition. Then, the feature descriptor is extracted from the vehicle contour area to obtain the vehicle feature descriptor. Feature descriptor extraction is to extract features that can uniquely identify the vehicle from the vehicle contour area. For example, the system can use algorithms such as SIFT (Scale-Invariant Feature Transform) or SURF (Speeded-Up Robust Features) to extract key points and descriptors from the vehicle contour area. These descriptors include not only the shape features of the vehicle contour, but may also include information such as color and texture. Through feature descriptor extraction, the system can generate a set of feature vectors that can uniquely identify the vehicle, which provide a basis for subsequent vehicle recognition and contour segmentation. For example, on a city's main road, after the system extracts the outline area of ​​a speeding vehicle, it uses the SIFT algorithm to extract key points and descriptors from the area to generate a vehicle feature descriptor. Next, based on the vehicle feature descriptor, the vehicle with traffic violations is segmented to obtain the illegal vehicle's outline features. The purpose of outline segmentation is to separate the vehicle from the background and obtain a clear vehicle outline.For example, the system can use the GrabCut algorithm or a deep learning segmentation model (such as U-Net) to achieve this purpose. Through these algorithms, the system can extract the vehicle's contour information from the vehicle feature descriptor and generate an accurate vehicle contour image. This process can not only remove background interference, but also ensure that the extracted vehicle contour is accurate and complete. For example, on the main road of the city, after the system generates the feature descriptor of the speeding vehicle, the GrabCut algorithm is used to segment the vehicle contour to obtain a clear vehicle contour feature image. Subsequently, the license plate position area is subjected to character recognition processing to obtain the license plate number. The character recognition processing is to extract the license plate number from the license plate position area, and this process usually uses optical character recognition (OCR) technology. For example, the system can use the Tesseract OCR engine or a deep learning character recognition model (such as CRNN) to achieve this purpose. Through these technologies, the system can accurately identify the license plate number from the license plate position area. This process requires not only a high-precision character recognition algorithm, but also pre-processing of the license plate image, such as denoising, binarization, etc., to improve the accuracy of recognition. For example, on the main roads of the city, after the system has screened out the license plate location area from the multi-scale feature image of the speeding vehicle, the Tesseract OCR engine will be used to perform character recognition on the license plate to obtain the license plate number. Finally, the illegal vehicle contour features and the license plate number are used as illegal vehicle extraction features. This process is to summarize all the features extracted in the previous steps to form a complete feature description of the illegal vehicle. For example, the system will combine the contour feature image and the license plate number of the speeding vehicle to generate a feature description containing the vehicle contour and license plate information. These feature descriptions can not only uniquely identify the illegal vehicle, but also provide detailed evidence support for subsequent law enforcement. For example, on the main roads of the city, after the system completes the feature extraction of the speeding vehicle, it will generate a feature description containing the vehicle contour feature image and the license plate number. This information will be encrypted and sent to the law enforcement terminal to provide accurate evidence of the violation for law enforcement personnel. Through this series of feature extraction steps, the system can accurately extract the detailed features of vehicles with traffic violations from real-time monitoring video data, which not only improves the accuracy and efficiency of illegal behavior identification, but also provides reliable evidence support for subsequent law enforcement. This process not only reflects the widespread application of modern computer vision and deep learning technologies in traffic management, but also makes positive contributions to improving the level of urban traffic management and ensuring traffic safety.

[0036] In a specific embodiment, the determining the position corresponding to the illegal vehicle based on the extracted features of the illegal vehicle by a preset positioning system includes: Performing feature enhancement on the extracted features of the illegal vehicle to obtain enhanced vehicle features; Index encoding the enhanced vehicle feature to obtain a vehicle feature hash code; By using a preset positioning system, a relational reasoning is performed on the illegal vehicle in the three-dimensional model of the traffic scene based on the vehicle feature hash code to obtain a relative position of the vehicle; wherein the relative position of the vehicle is the relative position of the illegal vehicle in the three-dimensional model; Based on the particle filter, the relative positions of the vehicles are tracked and optimized in time series to obtain the positions corresponding to the illegal vehicles.

[0037] Specifically, first, the extracted features of the illegal vehicle are enhanced to obtain enhanced vehicle features. This process is mainly to improve the robustness and discrimination of the features, so that they are more advantageous in the subsequent positioning process. Feature enhancement techniques usually include normalization, standardization, feature selection and other methods. For example, on the main roads of the city, after the system extracts the contour features and license plate number of a speeding vehicle through the previous steps, these features will be further enhanced. Specifically, the system may normalize the vehicle contour features to ensure that the contour features of different vehicles are compared at the same scale; the license plate number is standardized to ensure the accuracy of character recognition. Through these enhancement processes, the system can generate a set of more stable and reliable enhanced vehicle features, providing a basis for subsequent index coding and positioning. Next, the enhanced vehicle features are indexed and encoded to obtain a vehicle feature hash code. The purpose of index coding is to convert the enhanced vehicle features into an efficient and compact representation for fast retrieval and matching. Common index coding methods include local sensitive hashing (LSH), hash coding, etc. For example, the system can use the local sensitive hashing (LSH) algorithm to map the enhanced vehicle features into a high-dimensional hash space and generate vehicle feature hash codes. These hash codes can not only retain the main information of the original features, but also significantly reduce the storage and computing overhead. For example, on the main roads of the city, when the system generates the enhanced features of the speeding vehicle, it will use the LSH algorithm to index and encode it to generate a compact vehicle feature hash code, which will be used for subsequent positioning and matching. Then, through the preset positioning system, the illegal vehicle in the three-dimensional model of the traffic scene is reasoned based on the vehicle feature hash code to obtain the relative position of the vehicle. This process mainly uses the preset positioning system, combined with the vehicle feature hash code and the three-dimensional model of the traffic scene, to determine the relative position of the illegal vehicle in the three-dimensional model. The preset positioning system usually includes global positioning system (GPS), electronic map data, multi-sensor fusion and other technologies. For example, the system will match the generated vehicle feature hash code with the vehicle features in the three-dimensional model, and determine the specific position of the illegal vehicle in the three-dimensional model through relational reasoning. The process of relational reasoning not only considers the similarity of vehicle features, but also combines the geometric position and motion trajectory of the vehicle in the three-dimensional model. For example, on the main roads of the city, when the system generates the feature hash code of the speeding vehicle, it will match it with the vehicle features in the three-dimensional model, and determine the relative position of the vehicle in the three-dimensional model through relational reasoning, including which lane it is in, the distance from the intersection, etc. Finally, the relative position of the vehicle is tracked and optimized in time based on the particle filter to obtain the corresponding position of the illegal vehicle. The particle filter is a recursive Bayesian estimation method that is particularly suitable for processing nonlinear, non-Gaussian dynamic systems.In traffic scenarios, particle filters can be used to track and optimize the motion trajectory of vehicles in a timely manner, thereby improving the accuracy and stability of positioning. For example, after determining the relative position of the illegal vehicle in the three-dimensional model, the system uses a particle filter to track these position information in a timely manner. Specifically, the particle filter generates a set of particles, each of which represents a possible vehicle position, and gradually approaches the actual vehicle position by continuously updating the weight and position of the particles. This process can not only handle the uncertainty and noise of the vehicle in motion, but also ensure the continuity and consistency of the positioning results. For example, on the main roads of the city, after the system determines the relative position of the speeding vehicle through relational reasoning, it uses a particle filter to track and optimize these positions in a timely manner, and finally obtains the precise position of the illegal vehicle in the actual traffic scene. Through this series of positioning steps, the system can accurately determine the specific position of the illegal vehicle in the traffic scene from the extracted features of the illegal vehicle, which not only improves the accuracy and stability of positioning, but also provides reliable geographic information support for subsequent law enforcement. This process not only reflects the wide application of modern positioning technology and particle filtering algorithms in traffic management, but also makes positive contributions to improving the level of urban traffic management and ensuring traffic safety. For example, on a city's main roads, once the system has located a speeding vehicle, it will send the precise location information of the illegal vehicle to the law enforcement terminal, helping law enforcement officers to quickly arrive at the scene and deal with traffic violations in a timely manner.

[0038] In a specific embodiment, the index encoding of the enhanced vehicle feature to obtain the vehicle feature hash code includes: Performing hash mapping on the enhanced vehicle features using a preset local sensitive hashing algorithm to obtain an initial hash code; Quantizing the initial hash code to obtain a quantized feature; Using a preset Bloom filter to index encode the quantized features to obtain a preliminary feature hash code; Performing multi-level hash processing on the preliminary feature hash code to obtain a multi-level feature hash code; The multi-level feature hash codes are indexed and stored using a hash table to obtain a vehicle feature hash code.

[0039] Specifically, first, the enhanced vehicle features are hash mapped using a preset local sensitive hashing algorithm to obtain an initial hash code. Local sensitive hashing (LSH) is an efficient approximate nearest neighbor search algorithm, which is particularly suitable for fast retrieval of high-dimensional data. In traffic violation analysis, enhanced vehicle features are usually high-dimensional, such as vehicle profile features and license plate numbers. Through the local sensitive hashing algorithm, these high-dimensional features can be mapped to a low-dimensional hash space to generate an initial hash code. For example, on the main roads of the city, after the system obtains the profile features and license plate numbers of speeding vehicles through feature enhancement, the local sensitive hashing algorithm is used to map these features to a low-dimensional hash space to generate an initial hash code. This process not only significantly reduces the dimension of the data, but also retains the main information of the features, providing a basis for subsequent processing. Next, the initial hash code is quantized to obtain a quantized feature. The purpose of quantization is to convert continuous hash codes into discrete values ​​for further index encoding. Common quantization methods include binarization, quantization buckets, etc. For example, the system can binarize each value in the initial hash code and convert continuous floating point numbers to 0 or 1. Through quantization, the system can generate a set of discrete quantitative features, which are not only convenient for storage and retrieval, but also can improve the efficiency of subsequent index coding. For example, on the main roads of the city, after the system generates the initial hash code of the speeding vehicle, it will binarize it to generate a set of discrete quantitative features. Then, the quantized features are indexed and encoded using a preset Bloom filter to obtain a preliminary feature hash code. The Bloom filter is an efficient probabilistic data structure, which is mainly used to quickly determine whether an element exists in a set. In traffic violation analysis, the Bloom filter can be used to index and encode the quantitative features to generate a preliminary feature hash code. For example, the system can input the quantitative features into the Bloom filter to generate a preliminary feature hash code. This process can not only quickly determine whether the vehicle features exist in the database, but also significantly reduce the storage and retrieval overhead. For example, on the main roads of the city, after the system generates the quantitative features of the speeding vehicle, it will use the Bloom filter to index and encode it to generate a preliminary feature hash code. Next, the preliminary feature hash code is subjected to multi-level hash processing to obtain a multi-level feature hash code. The purpose of multi-level hash processing is to further improve the distinguishability and robustness of the feature through multiple hash operations. For example, the system can use multiple different hash functions to perform multiple hash processing on the preliminary feature hash code to generate a multi-level feature hash code. These multi-level feature hash codes can not only better retain the main information of the feature, but also improve the retrieval efficiency of the feature in large-scale data sets. For example, on the main roads in the city, after the system generates the preliminary feature hash code of the speeding vehicle, it will use multiple different hash functions to perform multi-level hash processing on it to generate a multi-level feature hash code.Finally, the multi-level feature hash code is indexed and stored using a hash table to obtain a vehicle feature hash code. A hash table is an efficient data structure for quickly finding and storing data. In traffic violation analysis, a hash table can be used to store multi-level feature hash codes to generate a final vehicle feature hash code. For example, the system can store the multi-level feature hash code in a hash table to generate a unique vehicle feature hash code. This process can not only quickly retrieve vehicle features, but also ensure the uniqueness and stability of the features. For example, on a city's main road, when the system generates a multi-level feature hash code for a speeding vehicle, it will store it in a hash table to generate a final vehicle feature hash code. These feature hash codes will be used for subsequent positioning and matching to help the system accurately determine the specific location of the illegal vehicle in the traffic scene. Through this series of index encoding steps, the system can efficiently convert enhanced vehicle features into a compact and easy-to-retrieve form, which not only improves the robustness and discrimination of the features, but also provides reliable technical support for subsequent positioning and matching. This process not only reflects the wide application of modern hashing technology and data structures in traffic management, but also makes positive contributions to improving the level of urban traffic management and ensuring traffic safety. For example, on a city's main roads, after the system completes the feature extraction and index coding of speeding vehicles, it will use the generated vehicle feature hash code for subsequent positioning and matching, helping law enforcement officers to quickly and accurately find illegal vehicles and deal with traffic violations in a timely manner.

[0040] In a specific embodiment, the encryption processing of the position corresponding to the illegal vehicle and the behavior analysis result corresponding to the illegal vehicle, generating an evidence chain data block and sending it to the law enforcement terminal includes: Performing data fusion on the position corresponding to the illegal vehicle and the behavior analysis results corresponding to the illegal vehicle to obtain illegal vehicle fusion data; Performing hash coding on the illegal vehicle fusion data to obtain a hash value of the illegal vehicle data; Encrypting the illegal vehicle data hash value by a preset asymmetric encryption algorithm to obtain an encrypted hash value; Using blockchain technology, the encrypted hash value is block generated to obtain an evidence chain data block; Digitally signing the evidence chain data block to obtain an evidence chain data block with a digital signature; The evidence chain data block with the digital signature is sent to the law enforcement terminal through a preset secure communication protocol.

[0041] Specifically, first, the location corresponding to the illegal vehicle and the behavior analysis results corresponding to the illegal vehicle are fused to obtain the illegal vehicle fusion data. This process mainly processes the location information and behavior analysis results of the illegal vehicle comprehensively to generate a data set containing all relevant information. For example, on the main road of the city, when the system determines the location of a speeding vehicle through the previous steps and analyzes its behavior in detail, it will fuse this information. Specifically, the system will integrate the precise location of the illegal vehicle, the time and place of the speeding behavior, the outline characteristics of the vehicle and the license plate number into a data structure to generate the illegal vehicle fusion data. This process not only ensures the integrity and consistency of the data, but also provides a unified data format for subsequent processing. Next, the illegal vehicle fusion data is hash-encoded to obtain the illegal vehicle data hash value. The purpose of hash coding is to convert the fusion data into a fixed-length hash value to ensure the uniqueness and immutability of the data. Hash coding usually uses a secure hash algorithm such as SHA-256. For example, the system will input the illegal vehicle fusion data into the SHA-256 algorithm to generate a 256-bit hash value. This hash value can not only uniquely identify the fusion data of the illegal vehicle, but also be used to verify the integrity and consistency of the data during the data transmission process. For example, on the main roads of the city, when the system generates the fusion data of the speeding vehicle, it will use the SHA-256 algorithm to hash it and generate the hash value of the illegal vehicle data. Then, the illegal vehicle data hash value is encrypted by a preset asymmetric encryption algorithm to obtain an encrypted hash value. Asymmetric encryption algorithms (such as RSA) can ensure the security of data during transmission and prevent data from being illegally intercepted and tampered with. For example, the system will use the preset RSA public key to encrypt the illegal vehicle data hash value to generate an encrypted hash value. This process not only protects the privacy of the data, but also ensures that only the law enforcement terminal holding the corresponding private key can decrypt and restore the original hash value. For example, on the main roads of the city, when the system generates the data hash value of the speeding vehicle, it will use the RSA public key to encrypt it and generate an encrypted hash value. Next, the encrypted hash value is generated into blocks using blockchain technology to obtain the evidence chain data block. Blockchain technology ensures the immutability and transparency of data through distributed ledgers and consensus mechanisms. For example, the system will use the encrypted hash value as a transaction data, generate a new block, and add it to the blockchain. This block contains not only the encrypted hash value, but also the hash value of the previous block, forming an unalterable chain structure. For example, on a city main road, when the system generates the encrypted hash value of a speeding vehicle, it will use it as a transaction data, generate a new block, and add it to the blockchain to form an evidence chain data block. Then, the evidence chain data block is digitally signed to obtain an evidence chain data block with a digital signature.The purpose of digital signature is to ensure the source and integrity of data and prevent data from being forged or tampered with. For example, the system will use the preset private key to digitally sign the evidence chain data block and generate an evidence chain data block with a digital signature. This process not only verifies the source of the data, but also ensures the integrity of the data during transmission. For example, on the main roads of the city, when the system generates an evidence chain data block for speeding vehicles, it will use the preset private key to digitally sign it and generate an evidence chain data block with a digital signature. Finally, the evidence chain data block with a digital signature is sent to the law enforcement terminal through the preset secure communication protocol. Secure communication protocols (such as TLS / SSL) can ensure the security and confidentiality of data during transmission. For example, the system will use the TLS / SSL protocol to send the evidence chain data block with a digital signature to the law enforcement terminal. This process not only protects the security of data during transmission, but also ensures the integrity and consistency of the data. For example, on the main roads of the city, when the system generates an evidence chain data block with a digital signature for speeding vehicles, it will send it to the law enforcement terminal through the TLS / SSL protocol. After receiving the data, the law enforcement terminal can verify the digital signature and decrypt the hash value to ensure the authenticity and integrity of the data. Through this series of data processing and transmission steps, the system can ensure the security, integrity and non-tamperability of the relevant data of illegal vehicles during the transmission and storage process, which not only improves the reliability and credibility of the data, but also provides solid evidence support for subsequent law enforcement. This process not only reflects the widespread application of modern cryptography and blockchain technology in traffic management, but also makes positive contributions to improving the level of urban traffic management and ensuring traffic safety. For example, on the main roads in the city, after the system completes the data processing and transmission of speeding vehicles, the law enforcement terminal can quickly and accurately obtain the detailed information of the illegal vehicles, take corresponding legal measures in time, and ensure that traffic violations are effectively handled.

[0042] The above describes a method for analyzing a traffic violation in an embodiment of the present invention. The following describes a device for analyzing a traffic violation in an embodiment of the present invention. Figure 2 In one embodiment of the present invention, a device for analyzing traffic violations includes: A monitoring module 21 is used to monitor the road in real time through a camera to obtain real-time monitoring video data; An analysis module 22, used for performing behavior analysis on the vehicles in the real-time monitoring video data to obtain behavior analysis results; The extraction module 23 is used to extract features of vehicles with traffic violations in the real-time monitoring video data to obtain illegal vehicle extraction features if the behavior analysis result shows that there are traffic violations in the vehicles on the road; A determination module 24, configured to determine the location of the illegal vehicle based on the extracted features of the illegal vehicle through a preset positioning system; The encryption module 25 is used to encrypt the position corresponding to the illegal vehicle and the behavior analysis result corresponding to the illegal vehicle, generate an evidence chain data block and send it to the law enforcement terminal.

[0043] In this embodiment, for the specific implementation of each unit in the above system embodiment, please refer to the above method embodiment, which will not be repeated here.

[0044] Reference Figure 3 The present invention also provides a computer device in an embodiment, wherein the internal structure of the computer device can be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a display screen, an input device, a network interface and a database connected through a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.

[0045] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0046] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0047] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided by the present invention and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM.

[0048] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0049] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for analyzing traffic violations, characterized in that: The following steps are involved: Real-time monitoring of roads is performed through cameras to obtain real-time monitoring video data; Performing behavior analysis on the vehicles in the real-time monitoring video data to obtain behavior analysis results; If the behavior analysis result shows that there are traffic violations among vehicles on the road, feature extraction is performed on the vehicles with traffic violations in the real-time monitoring video data to obtain illegal vehicle extraction features; Determine the location of the illegal vehicle based on the extracted features of the illegal vehicle through a preset positioning system; The location corresponding to the illegal vehicle and the behavior analysis results corresponding to the illegal vehicle are encrypted to generate an evidence chain data block and send it to the law enforcement terminal.

2. The method for analyzing traffic violations according to claim 1, characterized in that: The real-time monitoring of the road by the camera to obtain real-time monitoring video data includes: The road is captured in real time by a preset high-resolution camera to obtain an original video stream; Preprocessing the original video stream to obtain road preprocessing data; Performing target extraction on the road preprocessing data to obtain extracted target data; wherein the extracted target data includes pedestrians and vehicles; Performing three-dimensional synthesis on the extracted target data by using a preset three-dimensional reconstruction technology to obtain a three-dimensional model of the traffic scene; Pedestrians and vehicles in the three-dimensional model are monitored in real time to obtain real-time monitoring video data.

3. The method for analyzing traffic violations according to claim 1, characterized in that: The performing behavior analysis on the vehicle in the real-time monitoring video data to obtain the behavior analysis result includes: Performing motion vector analysis on vehicles and pedestrians in the real-time monitoring video data to obtain vehicle motion trajectory data and pedestrian trajectory data; Extracting spatiotemporal features from the vehicle motion trajectory data to obtain a vehicle behavior feature sequence; Performing a fusion analysis on the vehicle motion trajectory data and the vehicle behavior feature sequence to obtain a fusion behavior feature; wherein the fusion behavior feature includes the potential illegal tendency of the vehicle and the habitual characteristics of the driver; Based on a preset graph convolutional network algorithm, a contextual relationship analysis is performed on traffic violations between vehicles based on the fused behavior features to obtain behavior association results between vehicles; wherein the behavior association results between vehicles include relative positions, relative speeds, and interactive behaviors between vehicles; Performing pedestrian-vehicle interaction analysis on the vehicle motion trajectory data based on the pedestrian trajectory data to obtain a pedestrian-vehicle interaction analysis result; The behavior association result between the vehicles and the pedestrian-vehicle interaction analysis result are used as the behavior analysis result.

4. The method for analyzing traffic violations according to claim 1, characterized in that: The extracting features of vehicles with traffic violations in the real-time monitoring video data to obtain illegal vehicle extraction features includes: Performing multi-level feature extraction on vehicles that have committed traffic violations in the real-time monitoring video data to obtain a multi-scale feature image of the vehicle; Performing region of interest screening on the multi-scale feature image of the vehicle to obtain a screening region of interest; wherein the screening region of interest includes a license plate position region and a vehicle contour region; Extracting feature descriptors from the vehicle contour area to obtain vehicle feature descriptors; Based on the vehicle feature descriptor, the vehicle with traffic violations is segmented to obtain the illegal vehicle contour features; Performing character recognition processing on the license plate position area to obtain the license plate number; The illegal vehicle outline features and the license plate number are used as illegal vehicle extraction features.

5. The method for analyzing traffic violations according to claim 2, characterized in that: The method of determining the location of the illegal vehicle based on the extracted features of the illegal vehicle by using a preset positioning system includes: Performing feature enhancement on the extracted features of the illegal vehicle to obtain enhanced vehicle features; Index encoding the enhanced vehicle feature to obtain a vehicle feature hash code; By using a preset positioning system, a relational reasoning is performed on the illegal vehicle in the three-dimensional model of the traffic scene based on the vehicle feature hash code to obtain a relative position of the vehicle; wherein the relative position of the vehicle is the relative position of the illegal vehicle in the three-dimensional model; Based on the particle filter, the relative positions of the vehicles are tracked and optimized in time series to obtain the positions corresponding to the illegal vehicles.

6. The method for analyzing traffic violations according to claim 5, characterized in that: The step of indexing the enhanced vehicle feature to obtain a vehicle feature hash code includes: Performing hash mapping on the enhanced vehicle features using a preset local sensitive hashing algorithm to obtain an initial hash code; Quantizing the initial hash code to obtain a quantized feature; Using a preset Bloom filter to index encode the quantized features to obtain a preliminary feature hash code; Performing multi-level hash processing on the preliminary feature hash code to obtain a multi-level feature hash code; The multi-level feature hash codes are indexed and stored using a hash table to obtain a vehicle feature hash code.

7. The method for analyzing traffic violations according to claim 1, characterized in that: The encryption processing of the position corresponding to the illegal vehicle and the behavior analysis result corresponding to the illegal vehicle, generating an evidence chain data block and sending it to the law enforcement terminal includes: Performing data fusion on the position corresponding to the illegal vehicle and the behavior analysis results corresponding to the illegal vehicle to obtain illegal vehicle fusion data; Performing hash coding on the illegal vehicle fusion data to obtain a hash value of the illegal vehicle data; Encrypting the illegal vehicle data hash value by a preset asymmetric encryption algorithm to obtain an encrypted hash value; Using blockchain technology, the encrypted hash value is block generated to obtain an evidence chain data block; Digitally signing the evidence chain data block to obtain an evidence chain data block with a digital signature; The evidence chain data block with the digital signature is sent to the law enforcement terminal through a preset secure communication protocol.

8. A traffic violation analysis device, characterized in that: include: A monitoring module is used to monitor the road in real time through a camera to obtain real-time monitoring video data; An analysis module, used to perform behavior analysis on the vehicles in the real-time monitoring video data to obtain behavior analysis results; An extraction module, for extracting features of vehicles with traffic violations in the real-time monitoring video data to obtain extracted features of illegal vehicles if the behavior analysis result shows that there are traffic violations in vehicles on the road; A determination module, configured to determine the location corresponding to the illegal vehicle based on the extracted features of the illegal vehicle through a preset positioning system; The encryption module is used to encrypt the location corresponding to the illegal vehicle and the behavior analysis result corresponding to the illegal vehicle, generate an evidence chain data block and send it to the law enforcement terminal.

9. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images