基于CMOS图像采集的车道偏移识别系统设计与实现  

Design and implementation of lane deviation recognition system based on CMOS image acquisition

作　　者：李成勇 王莎 陈成瑞 LI Chengyong;WANG Sha;CHEN Chengrui(School of Electronic Information,Chongqing Engineering College,Chongqing 400056,China)

机构地区：[1]重庆工程学院电子信息学院,重庆400056

出　　处：《中国测试》2022年第6期106-110,共5页China Measurement & Test

基　　金：国家自然科学基金资助项目(61401052);重庆市教委科学技术基金资助项目(KJQN202001907)。

摘　　要：在目标体运动过程中对其运行轨迹信息采集是光电图像处理的一大难点,尤其是行驶速度较快的车辆。该系统基于CMOS图像采集传感器,以智能小车为运动目标模型,设计用于车道偏移识别的一套汽车辅助驾驶系统,选用CMOS型摄像头作为路面信息的采集提取,通过内部编写的图像二值化处理程序、中值滤波算法和图像锐化处理程序,实现车道线图像特征的提取,得到的车道线图像轮廓清晰。通过程序识别车道信息从而转动舵机角度,实现车道识别并偏移的目的,与电磁导航对车道偏移识别相比,识别率超过95%,远大于一般识别系统,且系统性能良好。It is one of the most difficult problems in photoelectric image processing to collect the tracking information of the moving target,especially for the vehicles with high speed.This system based on CMOS image acquisition sensor,intelligent car for moving target model,designed for lane migration to identify a set of the auxiliary driving system,type selection of CMOS camera as the road information acquisition extraction,image binarization processing program written by internal,median filtering algorithm and image sharpening processing procedure,realize the lane line image features extraction,the lane line image obtained has clear contour.Recognize the lane information through the program to turn the steering gear angle to achieve the purpose of lane identification and deviation,compared with electromagnetic navigation on the lane deviation recognition,the recognition rate is more than 95%,much higher than the general identification system,and the system performance is good.

关 键 词：车道偏移 面阵摄像头 二值化 中值滤波 

分 类 号：TN710.9[电子电信—电路与系统]