基于双目视觉的智能车避障方法研究

打开文本图片集

中图分类号：TP391.4 文献标识码：A 文章编号：2096-4706（2025）22-0189-05

Research on Obstacle Avoidance Method of Intelligent Vehicles Based on BinocularVision

LI Xie， XIAO Kaili， LI Ye， ZHONG Zhuoliang （Guangzhou University of Software，Guangzhou 51o990，China）

Abstract： In the complex and changeableroad environment，the obstacle avoidanceabilityof intellgent vehicles is veryimportant to ensure driving safetyand improve driving experience.This paper proposesanintellgentvehicleobstacle avoidance system basedon binocular vision.The systemuses ahigh-precision binocularcamera tocollct front road images in realtime，andintegratestheSGBMstereomatchingalgorithmandSLICsuperpixelsegmentationtechnologytoachieveaccurate identificationofobstaclesandacurateestimationofdistance，which provides reliabledatasupport forautonomous obstacle avoidance decision-makingofintellgentvehicles.Experimentsshowthatthesystemhashighrecognitionaccuracyinthecloserangeobstacleecognitionsene，efectivelypromotes thedevelopmentofintellgentveiclebstacleavodancetechologyand provides valuable innovativeideasand practicalreferences forthedesignofobstaclerecognition modulesin future intellgent vehicle obstacle avoidance systems.

Keywords：binocularvision;obstacle avoidance system;obstacle detection

0 引言

随着科技的飞速发展，智能车避障技术已成为交通领域的研究热点，越来越多的车企将环境感知技术作为重点研究方向[1。（剩余5423字）