An assisted driving method and system based on machine vision

By using multi-view vision sensors and spatiotemporal fusion technology, a three-dimensional scene representation model and a hierarchical environmental cognition map are constructed, generating a dynamic risk distribution heat map. A multi-objective optimization algorithm is used to generate a driving trajectory that balances safety, comfort, and efficiency, solving the problem of inaccurate environmental understanding in intelligent driving systems and achieving precise, flexible, and safe assisted driving.

CN122009210BActive Publication Date: 2026-07-24SHENZHEN LEADER AUTOMOTIVE INTELLIGENT TECH DEV CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN Β· China
Patent Type
Patents(China)
Current Assignee / Owner
SHENZHEN LEADER AUTOMOTIVE INTELLIGENT TECH DEV CO LTD
Filing Date
2026-02-03
Publication Date
2026-07-24

AI Technical Summary

Technical Problem

Existing intelligent driving systems lack accurate understanding of the surrounding environment, resulting in poor assisted driving performance and even safety risks.

Method used

Multi-view image data is collected by an onboard multi-view vision sensor array to construct a raw visual dataset containing spatial coordinate information. Spatiotemporal fusion processing is then performed to establish a unified three-dimensional spatial coordinate system. Semantic features are extracted to construct a hierarchical environmental cognition map, and a dynamic risk distribution heat map is generated. A multi-objective optimization algorithm is used to generate candidate driving trajectories that take into account safety, comfort, and efficiency, and refined control commands are output.

Benefits of technology

It achieves more precise, flexible and safer assisted driving effects, improves the driving experience and driving safety, and avoids the problems of being overly conservative or overly aggressive in traditional methods.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122009210B_ABST
    Figure CN122009210B_ABST
Patent Text Reader

Abstract

The application provides an assisted driving method and system based on machine vision, comprising: constructing an original vision data set; establishing a three-dimensional space coordinate system, converting multi-view image data into a three-dimensional scene representation model, and extracting semantic features therefrom to construct a hierarchical environment cognition graph; identifying key interactive objects in the current driving scene according to the interactive layer information in the hierarchical environment cognition graph, and calculating the risk influence weight of each key interactive object on the driving of the vehicle to generate a dynamic risk distribution heat map; based on the dynamic risk distribution heat map and the current driving state of the vehicle, a multi-objective optimization algorithm is used to generate a candidate driving trajectory set, and each candidate trajectory is verified; selecting the target trajectory with the optimal comprehensive score from the verified candidate trajectories, and outputting a control instruction to the vehicle control system. The application realizes more accurate, flexible and safe assisted driving effect, and improves the driving experience and driving safety.
Need to check novelty before this filing date? Find Prior Art