<title>New color road segmentation method</title>
Lists
Only you see this
Abstract
Road segmentation is the critical step in a vision system for outdoor road following. This paper presents a new method for color road segmentation. The basic idea of the method is as follows: each pixel of a color image consists of red, green, and blue intensity values; it can be considered as a feature vector in 3-D RGB space, and the road segmentation can be translated into a statistic classification problem. First, a color normalization transform is used to erase the shadows in a color image; then, the optimal discriminant plane technique is used for feature decorrelation; finally, road segmentation is completed by a minimum distance classifier designed on the optimal discriminant plane. We have done a lot of experiments, and the results show that the present method is effective.
BibTeX
@inproceedings{liu1992lt,
author = {Ke Liu and Yanxing Liu and Jingyu Yang and Yong-Qing Cheng and Nian-Chun Gu},
title = {{\<title\>New color road segmentation method\</title\>}},
booktitle = {Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE},
year = {1992},
doi = {10.1117/12.135189},
}Loading related content...