A Road Detection Algorithm Based on EF-boosting
To exploiting monocular vision for road detection, the EF-boosting algorithm for road detection is proposed in this paper. When the road detection is taken for a problem of region classification road or off-road, the boosting algorithm can be used as a classifier. Traditional classification-based road detection algorithm is on supposes that the features are independence to each other. But in fact it is not always the truth. In this paper, to improve the performance of classification, the relationship of features is explored. First, an over-complete feature combination set is constructed. Then features are selected and weighted according to the EF (Effect of each Feature for classification), which is calculated by the train and test error rate of boosting classifiers composed by each single feature in the combined vector on road image. After that, the EF-boosting classifier is constructed based on the EF. The experimental result shows the algorithm is effective especially for heavy shadow.
Road Detection Computer Vision Boosting Feature Combination
Sha Yun Yang Yong Zhang Guo-ying
Department of Computer Science, Beijing Institute of Petro Chemical Technology, Beijing 102617 Department of Computer Science, Chang Chun Institute of Technology, Changchun 130025
国际会议
北京
英文
2007-08-05（万方平台首次上网日期，不代表论文的发表时间）