期刊名称：International Journal of Signal Processing, Image Processing and Pattern Recognition
印刷版ISSN：2005-4254
出版年度：2015
卷号：8
期号：9
页码：267-274
DOI：10.14257/ijsip.2015.8.9.28
出版社：SERSC
摘要：A method of structure lane detection and departure based on inverse perspective mapping (IPM) and effective gradient points is proposed. Firstly, road image is transformed into plan view using inverse perspective transform to eliminate perspective effect. Secondly, "Sobel" operator and morphological image processing are used to detect edge and eliminate noise. According to the actual lane width, effective gradient points are extracted. Finally, lane is fitted using the least square method. Experimental results show that the method can detect both sides lane and analyze departure and lane detection rate is 97.5%. The method can provide reference for lane departure warning system
关键词：structure road; lane; gradient point; least squares method; departure
