铁道科学与工程学报2017，Vol.14Issue(2)：370-378,9.
基于LBP-HOG特征的接触网旋转双耳区域识别
Detection of catenary rotation binaural based on LBP-HOG feature
摘要
Abstract
In this paper, the principle of target location based on the combination of LBP feature, HOG feature and support vector machine is studied. And the positioning of the rotation binaural is investigated. The LBP feature and support vector machine method or HOG feature and support vector machine method are used to locate the rotation binaural. On this basis, the combination of LBP features and HOG features is proposed to form the LBP-HOG feature which is used to locate the position of the rotation binaural. Test results show that LBP-HOG feature effect is better than that of the single use of LBP features or HOG features under the same conditions.关键词
接触网检测/LBP/HOG/支持向量机/旋转双耳Key words
catenary detection/LBP/HOG/SVM/rotation binaural分类
交通工程引用本文复制引用
徐丹,于龙,陈唐龙,王佳祺..基于LBP-HOG特征的接触网旋转双耳区域识别[J].铁道科学与工程学报,2017,14(2):370-378,9.基金项目
四川省科技计划项目(2015GZ0148) （2015GZ0148）
