CN-122024474-A - Intelligent screening and controlling method for high-risk vehicles

CN122024474ACN 122024474 ACN122024474 ACN 122024474ACN-122024474-A

Abstract

The application provides an intelligent screening and controlling method of a high-risk vehicle, which relates to the technical field of intelligent traffic management and comprises the steps of acquiring multi-source data of the vehicle in real time, carrying out data cleaning and fusion processing on the multi-source data to construct a vehicle dynamic image, carrying out multidimensional feature extraction and risk assessment by adopting a deep learning model based on the vehicle dynamic image to generate a vehicle risk score and a risk grade, carrying out intelligent screening and generating targeted control advice based on the risk grade and user interaction operation, and optimizing the deep learning model and screening strategy by adopting an online learning mechanism based on user feedback data.

Inventors

SUN XINXIN
XUE YUPENG
SUN RUIYANG
WANG PO

Assignees

德州市公安局交通管理支队

Dates

Publication Date: 20260512
Application Date: 20260128

Claims (10)

1. An intelligent screening and controlling method for high-risk vehicles, which is characterized by comprising the following steps: Acquiring multi-source data of a vehicle in real time, wherein the multi-source data comprises vehicle basic information, a driving track, violation records and accident records; performing data cleaning and fusion processing on the multi-source data to construct a vehicle dynamic image; Based on the vehicle dynamic image, adopting a deep learning model to carry out multidimensional feature extraction and risk assessment, and generating a vehicle risk score and a risk level; based on the risk level and the user interaction operation, intelligent screening is carried out, and targeted management and control suggestions are generated; and optimizing the deep learning model and the screening strategy through an online learning mechanism based on the user feedback data.
2. The method of claim 1, wherein the acquiring the multi-source data of the vehicle in real time comprises: And acquiring basic information, driving tracks, violation records and accident records of the vehicle in real time through a data interface with the public security big data platform.
3. The method of claim 1, wherein the performing data cleansing and fusion processing on the multi-source data to construct a vehicle dynamic representation comprises: Removing abnormal data and noise by adopting a data cleaning algorithm, and complementing the missing data; Based on the spatiotemporal association rules, the multi-source data is fused into a unified vehicle dynamic representation that includes static attributes and dynamic behavior features.
4. A method according to claim 3, wherein said employing a deep learning model for multi-dimensional feature extraction and risk assessment comprises: Extracting time sequence characteristics including a moving mode and abnormal stay points by adopting a transducer model or a time sequence convolution network based on an attention mechanism; calculating the violation frequency and the accident association degree by adopting a statistical analysis method as statistical characteristics; Analyzing a special mark in the vehicle image by adopting a convolutional neural network as a visual characteristic; adopting a reinforcement learning model to identify high-risk driving behaviors as behavior characteristics; based on the extracted multidimensional features, vehicle risk scores are generated through a risk assessment model, and risk grades are dynamically classified based on an adaptive threshold.
5. The method of claim 4, wherein the intelligently filtering and generating targeted management and control suggestions based on the risk level and user interaction comprises: Providing a visual interaction interface, and supporting a user to combine screening conditions in a dragging or choosing mode; dynamically recommending high-frequency or high-weight screening conditions based on importance analysis of the multi-dimensional features; based on the screening result and the risk characteristic, a rule engine is combined with a machine learning model to generate a management and control suggestion, wherein the management and control suggestion comprises a key monitoring period, a point position of a necessary checkpoint and an early warning frequency.
6. The method of claim 5, wherein optimizing the deep learning model and screening strategy via an online learning mechanism based on user feedback data comprises: Collecting feedback data of a user on screening results and management and control suggestions, wherein the feedback data comprise false alarm records and missing report records; and updating parameters of the deep learning model by adopting an incremental learning technology, and adjusting weights and management and control rules of screening conditions.
7. The method according to claim 1, wherein the method further comprises: And formulating a differentiated vehicle management and control strategy based on the risk grade and the management and control suggestion.
8. High risk vehicle intelligent screening and management and control device, characterized in that, the device includes: The data acquisition module is used for acquiring multi-source data of the vehicle in real time; the data processing module is used for carrying out data cleaning and fusion processing on the multi-source data to construct a vehicle dynamic image; the risk assessment module is used for carrying out multidimensional feature extraction and risk assessment by adopting a deep learning model based on the vehicle dynamic image to generate a vehicle risk score and a risk grade; the screening and controlling module is used for carrying out intelligent screening based on the risk level and the user interaction operation and generating targeted controlling suggestions; and the optimization module is used for optimizing the deep learning model and the screening strategy through an online learning mechanism based on user feedback data.
9. A computer device, comprising: A memory and a processor in communication with each other, the memory having stored therein computer instructions which, upon execution, cause the processor to perform the steps of the method of any of claims 1-7.
10. A computer readable storage medium, on which a program is stored, which program, when being executed by a processor, carries out the steps of the method according to any one of claims 1-7.

Description

Intelligent screening and controlling method for high-risk vehicles Technical Field The application belongs to the technical field of intelligent traffic management, and particularly relates to an intelligent screening and controlling method for high-risk vehicles. Background In the field of public security traffic management, how to accurately and efficiently identify high-risk vehicles (such as suspected violations, accidents or vehicles with suspicious behaviors) from massive vehicle information and effectively manage and control the vehicles is a core challenge for guaranteeing road traffic safety and public safety. At present, the conventional vehicle screening system mainly has the following limitation that firstly, the system usually depends on manual operation, and a user needs to manually combine various conditions (such as license plates, vehicle types, violation types and the like) to perform query screening. Because of heterogeneous data sources (including basic information, running tracks, violation records and the like), the manual screening process is extremely tedious, time-consuming and labor-consuming, and cannot meet the real-time response requirement on dynamic risk scenes. Secondly, the existing system lacks the capability of quantitatively evaluating the risk of the vehicle, and cannot automatically learn and identify the risk mode from multidimensional features such as driving behaviors, violation frequencies, special marks and the like, so that the management and control decision is seriously dependent on manual experience, and the accuracy and consistency are difficult to guarantee. Finally, the screening result is severely disjointed with the subsequent control measures, and the system cannot automatically generate accurate control suggestions (such as monitoring time periods, bayonet points and the like) according to specific vehicle risk characteristics, so that police resources are unreasonable to be distributed, and early warning and interception effects are limited. Disclosure of Invention The application provides an intelligent screening and controlling method for high-risk vehicles, which aims to solve one of the technical problems. The technical scheme adopted by the application is as follows: The embodiment of the application provides an intelligent screening and controlling method for a high-risk vehicle, which comprises the following steps: Acquiring multi-source data of a vehicle in real time, wherein the multi-source data comprises vehicle basic information, a driving track, violation records and accident records; performing data cleaning and fusion processing on the multi-source data to construct a vehicle dynamic image; Based on the vehicle dynamic image, adopting a deep learning model to carry out multidimensional feature extraction and risk assessment, and generating a vehicle risk score and a risk level; based on the risk level and the user interaction operation, intelligent screening is carried out, and targeted management and control suggestions are generated; and optimizing the deep learning model and the screening strategy through an online learning mechanism based on the user feedback data. According to one embodiment of the present application, the acquiring the multi-source data of the vehicle in real time includes: And acquiring basic information, driving tracks, violation records and accident records of the vehicle in real time through a data interface with the public security big data platform. According to one embodiment of the present application, the data cleaning and fusion processing is performed on the multi-source data to construct a vehicle dynamic image, including: Removing abnormal data and noise by adopting a data cleaning algorithm, and complementing the missing data; Based on the spatiotemporal association rules, the multi-source data is fused into a unified vehicle dynamic representation that includes static attributes and dynamic behavior features. According to one embodiment of the present application, the multi-dimensional feature extraction and risk assessment using the deep learning model includes: Extracting time sequence characteristics including a moving mode and abnormal stay points by adopting a transducer model or a time sequence convolution network based on an attention mechanism; calculating the violation frequency and the accident association degree by adopting a statistical analysis method as statistical characteristics; Analyzing a special mark in the vehicle image by adopting a convolutional neural network as a visual characteristic; adopting a reinforcement learning model to identify high-risk driving behaviors as behavior characteristics; based on the extracted multidimensional features, vehicle risk scores are generated through a risk assessment model, and risk grades are dynamically classified based on an adaptive threshold. According to an embodiment of the present application, the intelligent screening based on the ri