CN-122024470-A - Vehicle traffic violation identification method, early warning method, device, equipment and medium

CN122024470ACN 122024470 ACN122024470 ACN 122024470ACN-122024470-A

Abstract

The application provides a vehicle traffic violation identification method, an early warning device, electronic equipment and a storage medium, wherein the identification method is applied to MEC and comprises the steps of obtaining vehicle dynamic data and OBU registration information which are sent by an OBU of a target vehicle through an RSU cluster, obtaining multi-source perception data of the target vehicle which are collected by multi-source perception equipment and enter a core detection area based on the vehicle dynamic data, fusing the vehicle dynamic data and the OBU registration information, the multi-source perception data and the pre-stored vehicle registration information of the target vehicle to obtain multi-source fusion data, carrying out illegal behavior identification on the target vehicle based on the multi-source fusion data to obtain an illegal identification result, and sending the illegal identification result to a vehicle-road cloud platform so that the vehicle-road cloud platform can update digital twin information of the target vehicle according to the illegal identification result, and generating early warning information and illegal evidence package according to the illegal identification result.

Inventors

Song Shisuo
YU ZHONGTENG
JIA JIA
XIONG PENG

Assignees

中信科智联科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260109

Claims (11)

1. A method for identifying traffic violations of vehicles, characterized in that it is applied to an edge calculation unit MEC, comprising: Acquiring vehicle dynamic data and OBU registration information sent by an on-board unit (OBU) of a target vehicle through a Road Side Unit (RSU) cluster; acquiring multisource perception data of the target vehicle which is acquired by multisource perception equipment and enters a core detection area based on vehicle dynamic data; Fusing the vehicle dynamic data, the OBU registration information, the multi-source perception data and the pre-stored vehicle registration information of the target vehicle to obtain multi-source fusion data; Based on the multi-source fusion data, carrying out illegal behavior recognition on the target vehicle to obtain an illegal recognition result; and sending the illegal identification result to a vehicle-road cloud platform so that the vehicle-road cloud platform can update digital twin information of the target vehicle according to the illegal identification result, and generating early warning information and illegal evidence packets according to the illegal identification result.
2. The method for identifying traffic violations of vehicles according to claim 1, wherein the acquiring multi-source perception data of the target vehicle entering a core detection zone, which is acquired by the multi-source perception device, based on vehicle dynamic data, comprises: Predicting a predicted time for the target vehicle to reach a core detection zone based on the vehicle dynamics data; when the predicted time is less than or equal to the total preparation time for starting the preset multi-source sensing equipment, a wake-up instruction is issued to the multi-source sensing equipment so as to wake up the multi-source sensing equipment; and acquiring multi-source sensing data of the target vehicle which is acquired after the multi-source sensing equipment is awakened and enters a core detection area.
3. The method for identifying traffic violations of vehicles according to claim 1, wherein the fusing the vehicle dynamic data, the OBU registration information, the multisource awareness data, and the pre-stored vehicle registration information of the target vehicle to obtain multisource fusion data includes: performing space-time alignment, cross verification and consistency comparison on the vehicle dynamic data, the OBU registration information, the multi-source perception data and the pre-stored vehicle registration information of the target vehicle to obtain a comparison result; And according to a comparison result, fusing the vehicle dynamic data, the OBU registration information, the multi-source perception data and the pre-stored vehicle registration information of the target vehicle by utilizing a multi-source data fusion algorithm to obtain multi-source fusion data, wherein the multi-source data fusion algorithm comprises data preprocessing, feature layer fusion and decision layer fusion.
4. The method for identifying traffic violations of vehicles according to claim 3, wherein the fusing of the vehicle registration information, the multi-source perception data and the vehicle dynamic data using a multi-source data fusion algorithm to obtain multi-source fused data comprises: preprocessing the vehicle registration information, the multi-source awareness data, and the vehicle dynamics data, and The preprocessed vehicle registration information, the multi-source perception data and the vehicle dynamic data are weighted and integrated according to preset weights, and unified feature vectors are obtained; And carrying out multidimensional evidence combination on the unified feature vector to obtain structured multisource fusion data.
5. The method for identifying traffic violations of vehicles according to claim 1, wherein the identifying the target vehicles for violations based on the multi-source fusion data to obtain the result of identifying violations comprises: And carrying out illegal behavior recognition on the multisource fusion data by using an AI illegal recognition model to obtain an illegal recognition result comprising an illegal type, illegal dynamic characteristics, illegal confidence and illegal evidence index, wherein the illegal behavior recognition comprises at least one of overrun, detour escape detection, shielding license plate, suspected license plate, S-bend, emergency brake, S-bend dynamic compensation and emergency brake dynamic compensation, and the AI illegal recognition model is obtained by training the multi-task recognition model based on deep learning in advance according to a prompt word template.
6. The vehicle traffic violation identification method of claim 1, further comprising: Monitoring the state of the RSU cluster in the core detection area; and when the monitoring result meets a preset condition, controlling the multi-source sensing equipment in the core detection area to enter a dormant state, wherein the preset condition comprises when the last RSU in the core detection area detects that the OBU signal of the last vehicle in the current traffic disappears, and when the upstream first RSU does not detect the OBU signal of the new target vehicle in a synchronous time window, and when the dormant hold time still does not detect the OBU signal of the new target vehicle in.
7. The vehicle traffic violation early warning method is characterized by being applied to a vehicle road cloud platform and comprising the following steps of: Receiving an illegal identification result of the target vehicle sent by an edge calculation unit MEC; Updating digital twin information of the target vehicle based on the violation identification result, and Generating early warning information and a illegal evidence packet based on the illegal recognition result; and sending the early warning information and the illegal evidence packet of the target vehicle to a traffic police distribution control platform so that the traffic police distribution control platform distributes and controls according to the early warning information and the illegal evidence packet.
8. A vehicle traffic violation identification device, characterized in that the device is applied to an edge calculation unit MEC, comprising: The system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring vehicle dynamic data and OBU registration information sent by an on-board unit (OBU) of a target vehicle through a Road Side Unit (RSU) cluster; the second acquisition module is used for acquiring multisource perception data of the target vehicle which is acquired by the multisource perception device and enters a core detection area based on vehicle dynamic data; The fusion module is used for fusing the vehicle dynamic data, the OBU registration information, the multi-source perception data and the pre-stored vehicle registration information of the target vehicle to obtain multi-source fusion data; the identification module is used for carrying out illegal behavior identification on the target vehicle based on the multi-source fusion data to obtain an illegal identification result; And the sending module is used for sending the illegal identification result to a vehicle-road cloud platform so that the vehicle-road cloud platform can update the digital twin information of the target vehicle according to the illegal identification result and generate early warning information and an illegal evidence packet according to the illegal identification result.
9. The utility model provides a vehicle traffic violation early warning device which characterized in that, the device is applied to the vehicle road cloud platform, includes: The receiving module is used for receiving the illegal recognition result of the target vehicle sent by the edge computing unit MEC; the updating module is used for updating digital twin information of the target vehicle based on the illegal identification result; The generation module is used for generating early warning information and a illegal evidence packet according to the illegal recognition result; And the sending module is used for sending the early warning information and the illegal evidence packet of the target vehicle to the traffic police distribution control platform so that the traffic police distribution control platform can distribute and control the target vehicle according to the early warning information and the illegal evidence packet.
10. An electronic device, comprising: comprising a processor, a memory, and a program or instructions stored on the memory and executable on the processor, which when executed by the processor, implement the vehicle traffic violation identification method of any of claims 1 to 6 or the vehicle traffic violation pre-warning steps of claim 7.
11. A readable storage medium, characterized in that it stores thereon a program or instructions that, when executed by a processor of an electronic device, implement the vehicle traffic violation identification method of any of claims 1 to 6 or the vehicle traffic violation pre-warning step of claim 7.

Description

Vehicle traffic violation identification method, early warning method, device, equipment and medium Technical Field The present application relates to the field of vehicle monitoring technologies, and in particular, to a vehicle traffic violation identification method, an early warning device, an electronic device, and a readable storage medium. Background With the rapid increase of the urban automobile conservation amount and the limitation of urban road space, vehicle illegal behaviors, such as suspected license plates/shielding license plates, overrun, detour escape detection, S-bend, sudden braking and other vehicle illegal behaviors, are increasingly frequently generated. In the related technology, aiming at the detection of the illegal behaviors of the vehicles, a fixed detection area is generally arranged at a specific position of a road, when the vehicles drive into the detection area, monitoring equipment such as a high-definition camera, a dynamic weighing sensor, a laser contour detector and the like are triggered by a ground induction coil to start working, related data such as appearance images, actual weight and vehicle body contour of the vehicles are collected, matching association of the collected data of multiple equipment is completed based on uniform time stamps and space coordinates, and then the associated matching result is uploaded to a rear-end highway overrun comprehensive management platform. And the comprehensive management platform firstly carries out preliminary analysis on the matching result, screens out suspected illegal overrun vehicles, then carries out auditing on related suspected illegal data of the suspected illegal overrun vehicles in a manual mode, and finally judges whether illegal evidences corresponding to the vehicles are generated according to the manual auditing result. Therefore, in the vehicle illegal action recognition process, the related technology has the defects of single dimension of sensing data, delayed response of illegal recognition and the like, can only rely on a single or few devices such as a camera, a wagon balance and the like to acquire monitoring data, cannot effectively detect the identity legitimacy of the vehicle, and particularly can accurately recognize complex illegal actions such as suspected fake license plates, shielding license plates, overrun, detour escape detection, S-bend, sudden braking and the like, thereby reducing the overall recognition efficiency of the complex illegal actions of the vehicle. Disclosure of Invention The application provides a vehicle traffic violation identification method, an early warning device, electronic equipment and a readable storage medium, which at least solve the problem that the overall identification efficiency of complex vehicle illegal behaviors is low because the acquired monitoring data is single and the validity of the vehicle identity cannot be accurately detected in the related technology. The technical scheme of the application is as follows: According to a first aspect of an embodiment of the present application, there is provided a vehicle traffic violation identification method, which is applied to an edge calculation unit MEC, including: Acquiring vehicle dynamic data and OBU registration information sent by an on-board unit (OBU) of a target vehicle through a Road Side Unit (RSU) cluster; acquiring multisource perception data of the target vehicle which is acquired by multisource perception equipment and enters a core detection area based on vehicle dynamic data; Fusing the vehicle dynamic data, the OBU registration information, the multi-source perception data and the pre-stored vehicle registration information of the target vehicle to obtain multi-source fusion data; Based on the multi-source fusion data, carrying out illegal behavior recognition on the target vehicle to obtain an illegal recognition result; And sending the illegal identification result to a vehicle-road cloud platform so that the vehicle-road cloud platform can update the digital twin information of the target vehicle according to the illegal identification result, and generating early warning information and an illegal evidence packet based on the illegal identification result. Optionally, the acquiring the multi-source sensing data of the target vehicle driving into the core detection area based on the vehicle dynamic data includes: Predicting the predicted time of the target vehicle reaching a core detection area based on the vehicle dynamic data, and calculating the total preparation time of the multi-source sensing equipment when the multi-source sensing equipment is awakened; when the predicted time is less than or equal to the total preparation time for starting the preset multi-source sensing equipment, a wake-up instruction is issued to the multi-source sensing equipment so as to wake up the multi-source sensing equipment; and acquiring multi-source sensing data of the target vehicle which is acquired