Urban intelligent traffic monitoring system based on machine vision

Abstract

The invention relates to the technical field of intelligent traffic, in particular to an urban intelligent traffic monitoring system based on machine vision, which comprises a road data acquisition module, an edge calculation module, a target detection module, an action analysis module, an illegal behavior detection module, a traffic accident detection module, a responsibility division module, a database and a management platform, the target detection module is connected with the edge calculation module, the action analysis module is connected with the target detection module, the illegal behavior detection module is connected with the action analysis module, the database is connected with the illegal behavior detection module and the management platform, and the flow statistics module is also connected with the target detection module. And the traffic accident detection module is connected with the action analysis module, so that the technical problems that an urban traffic monitoring system in the prior art is single in function, only image data is provided when a traffic accident occurs, responsibility judgment is performed by a worker, the processing efficiency is low, and the system is easily influenced by subjective consciousness are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent transportation, and particularly to an urban intelligent transportation monitoring system based on machine vision. Background Art

[0002] With the accelerating advancement of urbanization, the urban traffic system has become increasingly complex, traffic flow has increased sharply, and traffic management and safety monitoring are facing unprecedented challenges. To address these challenges, the traffic monitoring system, as an important tool for urban traffic management, is undergoing a critical period of transformation from traditional to intelligent.

[0003] Traditional traffic monitoring systems mainly rely on manual monitoring, fixed camera shooting, and simple image processing technologies to monitor traffic flow, detect traffic violations, and record traffic accidents, etc.

[0004] However, the existing urban traffic monitoring systems have single functions. When a traffic accident occurs, only image data is provided for staff to determine liability, with low processing efficiency and being easily affected by subjective consciousness. Summary of the Invention

[0005] The purpose of the present invention is to provide an urban intelligent transportation monitoring system based on machine vision, aiming to solve the technical problems in the existing urban traffic monitoring systems, such as single functions, only providing image data for staff to determine liability when a traffic accident occurs, with low processing efficiency and being easily affected by subjective consciousness.

[0006] To achieve the above purpose, an urban intelligent transportation monitoring system based on machine vision adopted by the present invention includes a road data acquisition module, an edge computing module, an object detection module, an action analysis module, an illegal behavior detection module, a traffic accident detection module, a liability division module, a database, and a management platform. The edge computing module is connected to the road data acquisition module, the object detection module is connected to the edge computing module, the action analysis module is connected to the object detection module, the illegal behavior detection module is connected to the action analysis module, the database is connected to both the illegal behavior detection module and the management platform, the traffic flow statistics module is also connected to the object detection module, the traffic accident detection module is connected to the action analysis module, and the liability division module is connected to the traffic accident detection module;

[0007] The road data detection module is used to collect road image data in real time and transmit the data to the edge computing module;

[0008] The edge computing module integrates a GPU acceleration module and is used to perform real-time preprocessing on the collected road image data;

[0009] The target detection module adopts the improved YOLOv7 algorithm (with the attention mechanism CBAM) to achieve multi-category recognition of vehicles / pedestrians / non-motor vehicles;

[0010] The motion analysis module performs motion analysis on the detected target to determine whether it has abnormal behavior;

[0011] The illegal behavior detection module detects traffic violations through the preset rules in the database and the target action analysis results, and issues an early warning and records it through the management platform;

[0012] The traffic accident detection module is used to detect whether a traffic accident occurs on the road, and when a traffic accident occurs, the responsibility division module is used to divide the responsibility.

[0013] The urban intelligent traffic monitoring system based on machine vision further includes a traffic statistics module and a traffic suggestion module, and the traffic statistics module is connected to the target detection module;

[0014] The traffic suggestion module is connected to the traffic statistics module;

[0015] The traffic suggestion module uses a spatiotemporal Transformer model to predict traffic changes in the next 15 minutes and output traffic efficiency optimization suggestions.

[0016] Among them, the urban intelligent traffic monitoring system based on machine vision also includes a management module, a login module and an identity authentication module. The management module is implanted in the management platform, and the login module is connected to the management module through the identity authentication module.

[0017] The identity verification module includes a password verification unit and a fingerprint verification unit, and both the password verification unit and the fingerprint verification unit are connected to the login module.

[0018] Wherein, the urban intelligent traffic monitoring system based on machine vision further includes an authority granting module, and the authority granting module is connected to the identity authentication module;

[0019] The authority granting module grants different authorities according to rules and the identities of the persons logging into the management module.

[0020] Wherein, the urban intelligent traffic monitoring system based on machine vision further includes a data optimization module, and the data optimization module is connected to the edge computing module;

[0021] The data optimization module uses a motion compensation algorithm to eliminate camera jitter interference.

[0022] Wherein, the urban intelligent traffic monitoring system based on machine vision further includes a level classification module and a push module, the level classification module is connected to the traffic accident detection module, and the push module is connected to the level classification module;

[0023] The level classification module is used to classify traffic accidents detected by the traffic accident detection module, and push information to different managers according to different levels.

[0024] Among them, the urban intelligent traffic monitoring system based on machine vision also includes an equipment monitoring module and a warning module, and the equipment monitoring module and the warning module are respectively connected to the road data acquisition module and the illegal behavior detection module.

[0025] The present invention is a machine vision-based urban intelligent traffic monitoring system. When used, the present invention collects road image data in real time through the road data detection module and transmits it to the edge computing module in real time. The edge computing module integrates a GPU acceleration module to perform real-time preprocessing on the collected road image data, greatly improving the speed and efficiency of data processing;

[0026] At the same time, the target detection module adopts the improved YOLOv7 algorithm (adding the attention mechanism CBAM) to achieve accurate identification of multiple categories such as vehicles, pedestrians, and non-motor vehicles. The motion analysis module performs motion analysis on the detected target to determine whether it has abnormal behavior, such as running a red light, driving in the wrong direction, etc. The violation detection module automatically detects traffic violations through the preset rules and target motion analysis results in the database, and issues an early warning and records it through the management platform. The traffic accident detection module can detect in real time whether a traffic accident occurs on the road, and when a traffic accident occurs, objectively and accurately divide the responsibility through the responsibility division module. In this way, the technical problems of the existing urban traffic monitoring system being single in function, only providing image data for staff to make responsibility judgments when a traffic accident occurs, and being inefficient and easily affected by subjective consciousness are solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0028] Figure 1 It is a principle block diagram of the first embodiment of the present invention.

[0029] Figure 2 It is the principle block diagram of the second embodiment of the present invention.

[0030] Figure 3 It is the principle block diagram of the third embodiment of the present invention.

[0031] 101 - Road data acquisition module, 102 - Edge computing module, 103 - Target detection module, 104 - Action analysis module, 105 - Illegal behavior detection module, 106 - Traffic accident detection module, 107 - Liability division module, 108 - Database, 109 - Management platform, 110 - Traffic flow statistics module, 111 - Traffic advice module, 112 - Management module, 113 - Login module, 114 - Identity authentication module, 115 - Password verification unit, 116 - Fingerprint verification unit, 201 - Permission granting module, 202 - Data optimization module, 203 - Level division module, 204 - Push module, 205 - Device monitoring module, 206 - Warning module, 301 - Energy optimization module, 302 - Storage module, 303 - Compression module. Specific embodiments

[0032] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present invention and should not be construed as a limitation to the present invention.

[0033] The first embodiment of the present application is as follows:

[0034] Please refer to Figure 1 , Figure 1 It is the principle block diagram of the first embodiment of the present invention.

[0035] The present invention provides an urban intelligent transportation monitoring system based on machine vision, including a road data acquisition module 101, an edge computing module 102, a target detection module 103, an action analysis module 104, an illegal behavior detection module 105, a traffic accident detection module 106, a liability division module 107, a database 108, a management platform 109, a traffic flow statistics module 110, a traffic advice module 111, a management module 112, a login module 113, and an identity authentication module 114. The identity authentication module 114 includes a password verification unit 115 and a fingerprint verification unit 116. The foregoing solution solves the technical problems in the prior art that the urban traffic monitoring system has a single function, and only provides image data for staff to determine liability when a traffic accident occurs, resulting in low processing efficiency and being easily affected by subjective consciousness.

[0036] For this specific embodiment, the road data detection module is used to collect road image data in real time and transmit the data to the edge computing module 102;

[0037] The edge computing module 102 integrates a GPU acceleration module for real-time preprocessing of the collected road image data;

[0038] The target detection module 103 adopts an improved YOLOv7 algorithm (adding the attention mechanism CBAM) to achieve multi-class recognition of vehicles / pedestrians / non-motor vehicles;

[0039] The action analysis module 104 performs action analysis on the detected targets to determine whether there are abnormal behaviors;

[0040] The illegal behavior detection module 105 detects traffic violations through the preset rules in the database 108 and the target action analysis results, issues warnings through the management platform 109 and records them;

[0041] The traffic accident detection module 106 is used to detect whether a traffic accident has occurred on the road, and when a traffic accident occurs, the responsibility is divided by the responsibility division module 107.

[0042] Among them, the edge computing module 102 is connected to the road data collection module 101, the target detection module 103 is connected to the edge computing module 102, the action analysis module 104 is connected to the target detection module 103, the illegal behavior detection module 105 is connected to the action analysis module 104, the database 108 is connected to both the illegal behavior detection module 105 and the management platform 109, the traffic flow statistics module 110 is also connected to the target detection module 103, the traffic accident detection module 106 is connected to the action analysis module 104, and the responsibility division module 107 is connected to the traffic accident detection module 106. In specific use, the present invention collects road image data in real time through the road data detection module and immediately transmits it to the edge computing module 102. The edge computing module 102 integrates a GPU acceleration module to perform real-time preprocessing on the collected road image data, greatly improving the speed and efficiency of data processing;

[0043] At the same time, the target detection module 103 adopts the improved YOLOv7 algorithm (adding the attention mechanism CBAM) to achieve accurate identification of multiple categories such as vehicles, pedestrians, and non-motor vehicles. The motion analysis module 104 performs motion analysis on the detected target to determine whether it has abnormal behavior, such as running a red light, driving in the wrong direction, etc. The illegal behavior detection module 105 automatically detects traffic violations through the preset rules and target motion analysis results in the database 108, and issues an early warning and records it through the management platform 109. The traffic accident detection module 106 can detect in real time whether a traffic accident occurs on the road, and when a traffic accident occurs, objectively and accurately divide the responsibility through the responsibility division module 107. In this way, the technical problems of the urban traffic monitoring system in the prior art being single in function, only providing image data for staff to make responsibility judgments when a traffic accident occurs, low processing efficiency and susceptibility to subjective consciousness are solved.

[0044] Secondly, the traffic statistics module 110 is connected to the target detection module 103;

[0045] The traffic suggestion module 111 is connected to the traffic statistics module 110;

[0046] The traffic suggestion module 111 uses a spatiotemporal Transformer model to predict traffic changes in the next 15 minutes and output traffic efficiency optimization suggestions;

[0047] The traffic statistics module 110 and the target detection module 103 work together to achieve real-time and accurate monitoring of traffic flow. It provides timely and reliable traffic data to the management department, helps to better understand the road congestion situation, and provides data support for subsequent traffic diversion and planning. The traffic suggestion module 111 adopts a spatiotemporal Transformer model, which can fully consider time and space factors and accurately predict traffic changes within the next 15 minutes. By predicting future traffic changes, the traffic suggestion module 111 can output traffic efficiency optimization suggestions, such as suggesting that drivers choose other routes, adjust departure times, etc., to effectively avoid congested sections and improve traffic efficiency.

[0048] At the same time, the management module 112 is implanted in the management platform 109, and the login module 113 is connected to the management module 112 through the identity authentication module 114;

[0049] The management module 112 is embedded in the management platform 109, making the management functions of the platform more centralized and efficient. Managers can easily access and manage the data and information of the traffic monitoring system through the platform, which greatly improves management efficiency and convenience;

[0050] Through the verification mechanism of the identity verification module 114, the system can identify and filter out illegal or abnormal access requests in real time, thereby reducing the possibility of the system being attacked or interfered with. This helps to improve the reliability and stability of the system and ensure the normal operation of the traffic monitoring system.

[0051] In addition, the password verification unit 115 and the fingerprint verification unit 116 are both connected to the login module 113, and the administrator can log in to the management module 112 through the password verification unit 115 and the fingerprint verification unit 116 to prevent illegal personnel from using it.

[0052] When using the urban intelligent traffic monitoring system based on machine vision of this embodiment, the present invention collects road image data in real time through the road data detection module and transmits it to the edge computing module 102 in real time. The edge computing module 102 integrates a GPU acceleration module to perform real-time preprocessing on the collected road image data, which greatly improves the speed and efficiency of data processing;

[0053] At the same time, the target detection module 103 adopts the improved YOLOv7 algorithm (adding the attention mechanism CBAM) to achieve accurate identification of multiple categories such as vehicles, pedestrians, and non-motor vehicles. The motion analysis module 104 performs motion analysis on the detected target to determine whether it has abnormal behavior, such as running a red light, driving in the wrong direction, etc. The illegal behavior detection module 105 automatically detects traffic violations through the preset rules and target motion analysis results in the database 108, and issues an early warning and records it through the management platform 109. The traffic accident detection module 106 can detect in real time whether a traffic accident occurs on the road, and when a traffic accident occurs, objectively and accurately divide the responsibility through the responsibility division module 107. In this way, the technical problems of the urban traffic monitoring system in the prior art being single in function, only providing image data for staff to make responsibility judgments when a traffic accident occurs, low processing efficiency and susceptibility to subjective consciousness are solved.

[0054] The second embodiment of the present application is:

[0055] Based on the first embodiment, please refer to Figure 2 , Figure 2 It is a principle block diagram of the second embodiment of the present invention.

[0056] The present invention provides a city intelligent traffic monitoring system based on machine vision, which also includes an authority granting module 201, a data optimization module 202, a level classification module 203, a push module 204, an equipment monitoring module 205 and a warning module 206.

[0057] For this specific embodiment, the permission granting module 201 is connected to the authentication module 114;

[0058] The permission granting module 201 grants different permissions according to rules and the identity of the person logging in to the management module 112. By introducing the permission granting module 201, the system can obtain the identity information and actual needs of the user logging in to the management module 112 based on the authentication module 114, and dynamically allocate permissions for them, realizing the flexibility and fine-grained control of permission management.

[0059] Among them, the data optimization module 202 is connected to the edge computing module 102;

[0060] The data optimization module 202 uses a motion compensation algorithm to eliminate camera jitter interference. The motion compensation algorithm adopted by the data optimization module 202 can effectively eliminate the jitter of the camera caused by external factors (such as vehicle driving, wind, etc.), making the captured images more stable and clear, which is helpful for subsequent image analysis and processing.

[0061] Secondly, the level classification module 203 is connected to the traffic accident detection module 106, and the push module 204 is connected to the level classification module 203;

[0062] The level classification module 203 is used to classify the traffic accidents detected by the traffic accident detection module 106, and push the information to different management personnel according to different levels;

[0063] The level classification module 203 inputs the collected accident images into the trained YOLOv8 model, and the model will output the severity rating of the accident;

[0064] The push module 204 can adopt a rule-based push algorithm to push the information to different management personnel according to the level of the accident;

[0065] The specific steps are as follows:

[0066] Rule setting: Set different management personnel corresponding to different levels of accidents. For example, minor accidents can be pushed to grass-roots traffic police, and serious accidents can be pushed to senior management personnel of the traffic management department.

[0067] Information push: According to the accident level output by the level classification module 203, the push module 204 pushes the accident information to the corresponding management personnel through the API interface. The push methods can include text messages, emails, APP pushes, etc.

[0068] The construction of the YOLOv8 model is as follows:

[0069] Collect image data of traffic accidents and perform annotation. The annotation content includes the severity of the accident (such as "medium" and "severe").

[0070] Model training: Use the YOLOv8 model to train the annotated data. The model can output accurate bounding boxes and classification results, intuitively showing the rating information of the accident.

[0071] Again, the device monitoring module 205 and the warning module 206 are respectively connected to the road data collection module 101 and the illegal behavior detection module 105;

[0072] The device monitoring module 205 can self-diagnose device abnormalities such as camera occlusion and offset. The technologies adopted are: anomaly detection based on Autoencoder: comparing real-time video with historical scene features; pan-tilt automatic calibration: matching landmark points through the SLAM algorithm (calibration accuracy ±0.3°);

[0073] When the illegal behavior detection module 105 determines that a pedestrian runs a red light, the warning module 206 can project a red warning image in front of the pedestrian's trajectory to warn the pedestrian and prevent the pedestrian from running a red light.

[0074] When using a machine vision-based urban intelligent transportation monitoring system according to this embodiment, in specific use, by introducing the permission granting module 201, the system can obtain the identity information and actual needs of the user logging in to the management module 112 according to the identity verification module 114, and dynamically allocate permissions for it, realizing the flexibility and fine-grained control of permission management.

[0075] The device monitoring module 205 can self-diagnose device abnormalities such as camera occlusion and offset. The technologies adopted are: anomaly detection based on Autoencoder: comparing real-time video with historical scene features; pan-tilt automatic calibration: matching landmark points through the SLAM algorithm (calibration accuracy ±0.3°);

[0076] When the illegal behavior detection module 105 determines that a pedestrian runs a red light, the warning module 206 can project a red warning image in front of the pedestrian's trajectory to warn the pedestrian and prevent the pedestrian from running a red light.

[0077] The third embodiment of this application is:

[0078] Based on the second embodiment, please refer to Figure 3 , Figure 3 which is the principle block diagram of the third embodiment of the present invention.

[0079] The present invention provides an urban intelligent transportation monitoring system based on machine vision, which further includes an energy optimization module 301, a storage module 302, and a compression module 303.

[0080] For this specific embodiment, the energy optimization module 301 is connected to the road data collection module 101. By setting the energy optimization module 301, the system can intelligently adjust the device power according to real-time traffic data and system load conditions, effectively reducing the system operation energy consumption, which conforms to the development trend of green and low-carbon.

[0081] Among them, the storage module 302 is connected to the management platform 109, and the compression module 303 is connected to the storage module 302. The storage module 302 is responsible for storing various data passing through the system, and at the same time provides fast data retrieval and access functions, supporting the real-time data analysis and decision-making support of the management platform 109. The compression module 303 performs efficient compression processing on the original data, significantly reducing the storage space occupation and lowering the data storage device and operation and maintenance costs.

[0082] When using the urban intelligent transportation monitoring system based on machine vision of this embodiment, during specific use, by setting the energy optimization module 301, the system can intelligently adjust the device power according to real-time traffic data and system load conditions, effectively reducing the system operation energy consumption, which conforms to the development trend of green and low-carbon. The storage module 302 is responsible for storing various data passing through the system, and at the same time provides fast data retrieval and access functions, supporting the real-time data analysis and decision-making support of the management platform 109. The compression module 303 performs efficient compression processing on the original data, significantly reducing the storage space occupation and lowering the data storage device and operation and maintenance costs.

[0083] The above-disclosed is only a preferred embodiment of the present invention. Of course, it cannot be used to limit the scope of the rights of the present invention. Those of ordinary skill in the art can understand the entire or part of the processes of implementing the above embodiments, and the equivalent changes made according to the claims of the present invention still fall within the scope covered by the invention.

Claims

1. An urban intelligent transportation monitoring system based on machine vision, characterized in that it includes a road data acquisition module, an edge computing module, an object detection module, an action analysis module, an illegal behavior detection module, a traffic accident detection module, a liability division module, a database, and a management platform. The edge computing module is connected to the road data acquisition module, the object detection module is connected to the edge computing module, the action analysis module is connected to the object detection module, the illegal behavior detection module is connected to the action analysis module, the database is connected to both the illegal behavior detection module and the management platform, the traffic flow statistics module is also connected to the object detection module, the traffic accident detection module is connected to the action analysis module, and the liability division module is connected to the traffic accident detection module; The road data detection module is used to collect road image data in real time and transmit the data to the edge computing module; The edge computing module integrates a GPU acceleration module and is used to preprocess the collected road image data in real time; The object detection module uses an improved YOLOv7 algorithm (adding an attention mechanism CBAM) to achieve multi-class recognition of vehicles / pedestrians / non-motor vehicles; The action analysis module analyzes the actions of the detected objects to determine whether there are abnormal behaviors; The illegal behavior detection module detects traffic violations through the preset rules in the database and the object action analysis results, issues warnings through the management platform and records them; The traffic accident detection module is used to detect whether a traffic accident has occurred on the road, and when a traffic accident occurs, the liability is divided through the liability division module.

2. The urban intelligent transportation monitoring system based on machine vision according to claim 1, characterized in that the urban intelligent transportation monitoring system based on machine vision further includes a traffic flow statistics module and a traffic flow advice module, and the traffic flow statistics module is connected to the object detection module; the traffic flow advice module is connected to the traffic flow statistics module; the traffic flow advice module uses a spatio-temporal Transformer model to predict the traffic flow change in the next 15 minutes and outputs traffic efficiency optimization suggestions.

3. The urban intelligent transportation monitoring system based on machine vision according to claim 2, characterized in that the urban intelligent transportation monitoring system based on machine vision further includes a management module, a login module, and an identity verification module. The management module is implanted in the management platform, and the login module is connected to the management module through the identity verification module.

4. The urban intelligent transportation monitoring system based on machine vision according to claim 3, characterized in that the identity verification module includes a password verification unit and a fingerprint verification unit, and both the password verification unit and the fingerprint verification unit are connected to the login module.

5. The urban intelligent transportation monitoring system based on machine vision according to claim 4, characterized in that the urban intelligent transportation monitoring system based on machine vision further includes a permission granting module, and the permission granting module is connected to the identity verification module; The permission granting module grants different permissions according to rules and the identity of the personnel logging in to the management module.

6. The machine vision-based urban intelligent transportation monitoring system according to claim 5, characterized in that the machine vision-based urban intelligent transportation monitoring system further includes a data optimization module, and the data optimization module is connected to the traffic flow monitoring module; the data optimization module uses a motion compensation algorithm to eliminate the interference of camera jitter.

7. The machine vision-based urban intelligent transportation monitoring system according to claim 6, characterized in that the machine vision-based urban intelligent transportation monitoring system further includes a grading module and a push module, the grading module is connected to the traffic accident detection module, and the push module is connected to the grading module; the grading module is used to grade the traffic accidents detected by the traffic accident detection module and push the information to different management personnel according to different grades.

8. The machine vision-based urban intelligent transportation monitoring system according to claim 7, characterized in that the machine vision-based urban intelligent transportation monitoring system further includes a device monitoring module and a warning module, and the device monitoring module and the warning module are respectively connected to the road data collection module and the illegal behavior detection module.

Images