A human-computer collaborative traffic violation intelligent detection and identification method and system

By caching video streams and receiving driver voice triggers on the in-vehicle terminal, and combining cloud analysis to generate credible evidence files, the high power consumption and incomplete evidence chain problems of existing traffic violation detection systems are solved, achieving low power consumption and high efficiency in violation detection and evidence collection.

CN121686794BActive Publication Date: 2026-07-03GUANGZHOU RUNQIANLONG INFORMATION TECHNOLOGY CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
GUANGZHOU RUNQIANLONG INFORMATION TECHNOLOGY CO LTD
Filing Date
2025-12-18
Publication Date
2026-07-03

AI Technical Summary

Technical Problem

Existing traffic violation detection systems rely on highly complex algorithms, resulting in high computational power and energy consumption. They are unable to accurately identify violations in complex environments and lack human-machine collaboration, leading to incomplete evidence chains.

Method used

The vehicle terminal uses low-power caching of video streams and records spatiotemporal information. It receives voice triggers from the driver, reviews the video, and uploads it to the cloud for analysis. It then uses a multi-task deep neural network to determine traffic violations and generates credible evidence files locally.

Benefits of technology

It reduces system power consumption, enables efficient and complete evidence collection of violations, ensures the timeliness and legal validity of evidence, and realizes intelligent detection through human-machine collaboration.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121686794B_ABST
    Figure CN121686794B_ABST
Patent Text Reader

Abstract

The application discloses a kind of man-machine collaborative traffic violation intelligent detection identification method, system.The method includes: through vehicle terminal continuous caching with space-time information video stream;Through voice recognition, driver instruction is received and event reference point T0 is marked;Based on T0, extract the historical video segment and metadata of preset time length and upload to cloud end by backtracking;Cloud end carries out analysis to video segment by multi-task deep neural network model, executes vehicle tracking, lane line and sign identification based on vision, violation behavior determination and license plate recognition, generates structured report;Report is returned to vehicle terminal, after driver confirmation or correction, in local, generate and store with legal force violation event archives with digital signature and trusted time stamp in combination.This application realizes the efficient cooperation of active trigger of driver and intelligent analysis of machine, evidence solidification, improves the convenience, accuracy and evidence credibility of traffic violation evidence.
Need to check novelty before this filing date? Find Prior Art