微型机与应用2017，Vol.36Issue(7)：39-42,4.DOI:10.19358/j.issn.1674-7720.2017.07.012
雨天环境基于HOG-SIFT特征稀疏表示的行人检测
A pedestrian detection algorithm based on HOG-SIFT feature sparse representation in rainy environment
陶春 1陈淑荣1
作者信息
- 1. 上海海事大学 信息工程学院,上海 201612
- 折叠
摘要
Abstract
In rainy environment, the rain noise and the weakened gray value of the image will make the pedestrian contour feature lost, which leads to the misdetection and error detection of surveillance video.Given this,the paper puts forward a kind of a pedestrian detection algorithm based on HOG-SIFT feature sparse representation.The histogram equalization is applied to reduce rain noise and the HOG-SIFT fusion feature is used to characterize pedestrians in video image to cut the loss of contour features.And it takes advantage of the sparse representation to reduce the dimension of the fusion feature and the amount of computation, retain valid pedestrian features and reduce misdetection and error detection combined with AdaBoost classifier.Experimental results show that the proposed method can effectively improve the accuracy of pedestrian detection in rainy environment.关键词
HOG特征/SIFT特征/稀疏表示/行人检测Key words
HOG feature/SIFT feature/sparse representation/pedestrian detection分类
计算机与自动化引用本文复制引用
陶春,陈淑荣..雨天环境基于HOG-SIFT特征稀疏表示的行人检测[J].微型机与应用,2017,36(7):39-42,4.
