基于组合特征和级联分类器的防震锤检测算法  

Detection of vibration damper based on combination features and cascade classifier

作　　者：王赛娇[1,2] 李黎 徐晓宇[2] WANG Sai-jiao;LI Li;XU Xiao-yu(School of Higher Vocational,Taizhou Radio&TV University,Taizhou 318000,China;School of Computer Science and Technology,Hangzhou Dianzi University,Hangzhou 310018,China)

机构地区：[1]台州广播电视大学高职学院,浙江台州318000 [2]杭州电子科技大学计算机学院,浙江杭州310018

出　　处：《计算机工程与设计》2020年第5期1336-1344,共9页Computer Engineering and Design

摘　　要：提出一种高压输电线路上的防震锤检测识别算法,算法基于分块的Haar特征、基于区域的LBP特征以及HOG特征一起作为组合特征来检测防震锤。其主要分为5个步骤:预处理待检测图像;改进归一化互相关匹配算法并进行模板匹配,得到防震锤疑似区域样本集;提取防震锤疑似区域的组合特征;对防震锤疑似区域使用级联分类器进行多级分类;统计分类结果。实验结果表明,该算法具有较高的精确率、召回率和准确率。A vibration hammer detection algorithm for high-voltage transmission lines was proposed.The vibration hammer was detected based on the block Haar feature,the region-based LBP feature,and the HOG feature together.The detection procedures included five steps.The images to be detected were preprocessed.Template matching was implemented using an improved normalized cross-correlation matching algorithm to obtain a suspected regional sample set of the vibration damper.Combination features were extracted.Multi-level classification was implemented using cascaded classifiers.Statistical classification results were obtained.The practice shows that the algorithm has high precision,recall and accuracy.

关 键 词：组合特征 级联分类器 矩形特征 局部二值模式 防震锤检测算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]