基于Otsu阈值与形态学处理的车道线检测  

Lane Detection Based on Otsu Threshold and Morphology

作　　者：王祥 罗素云[1] Wang Xiang;Luo Suyun(School of Automotive Engineering,Shanghai University of Engineering Science,Shanghai 201620,China)

机构地区：[1]上海工程技术大学汽车工程学院

出　　处：《农业装备与车辆工程》2019年第10期69-72,共4页Agricultural Equipment & Vehicle Engineering

摘　　要：为了提高车道线检测的识别能力和可靠性,提出了一种在对图像进行Hough变换前进行预处理的方法。该方法采用直方图均衡化、拉普拉斯滤波器图像增强,进而引用Otsu阈值分割处理转化为二值图像;再对该二值图像形态学处理消除物体边界点和微小结构像素,减少后期Hough变换的投票像素点;采用Canny算子对图像进行边缘检测以及Hough变换检测出车道线。In order to improve the recognition ability and reliability of lane detection,this paper put forward a lane detection method about image processing,which refers to histogram equalization,image enhancement with Laplace filtering,Otsu threshold segmentation converting to binary image,and then modifying binary image with morphology processing to eliminate the object boundary points and small pixel structure,to reduce the next voting space for Hough transform,then using Canny operator for edge detection and Hough transform to detect the lane line.

关 键 词：Otsu阈值分割 形态学处理 边缘检测 HOUGH变换 车道线检测 

分 类 号：U461[机械工程—车辆工程]