结合肤色和Hough变换的人手检测方法  

Hand Detection Method Based on Skin Color and Hough Transform

作　　者：杨永辉 鄢阳[1,2] 宋友 张海龙[1,2] 冯运 郭子娟[4] YANG Yonghui;YAN Yang;SONG You;ZHANG Hailong;FENG Yun;GUO Zijuan(NARI Group Corporation,State Grid Electric Power Research Institute,Nanjing,Jiangsu 211006,China;Wuhan NARI Limited Liability Company,State Grid Electric Power Research Institute,Wuhan,Hubei 430074,China;State Grid Sichuan Electric Power Research Institute,Chengdu,Sichuan 610072,China;China Electric Power Research Institute,Wuhan,Hubei 430074,China)

机构地区：[1]南瑞集团(国网电力科学研究院)有限公司,江苏南京211106 [2]国网电力科学研究院武汉南瑞有限责任公司,湖北武汉430074 [3]国网四川省电力公司电力科学研究院,四川成都610072 [4]中国电力科学研究院有限公司,湖北武汉430074

出　　处：《广东电力》2019年第9期56-61,共6页Guangdong Electric Power

基　　金：国家电网有限公司科技项目(52199916024X)

摘　　要：为了实现电气测试设备校准全景仿真中的人机交互,需要根据人手的自然特征来识别手势。针对手势识别中的手部检测问题,提出了一种结合肤色和Hough变换的人手检测与分割方法。首先,分析肤色的色调值的分布情况,设计了基于阈值的肤色检测方法以获取肤色二值掩码图像;然后,对二值掩码图像先进行边缘检测,再进行Hough变换,来检测与手部相关的轮廓,排除与脸部相关的轮廓;最后,用Hough变换所得到的手部轮廓对肤色二值掩码图像进行约束,即可得到人手区域的信息。实验结果表明,所提方法能够有效识别出图像中的人手信息,并且算法速度较快。To realize man-machine interaction in panorama simulation of electrical test equipment verification,it is necessary to identify gesture according to natural characteristics of the hand.Aiming at the problem of hand detection in gesture identification,this paper proposes a kind of hand detection and partition method based on skin color and Hough transform.Firstly,it analyzes distribution of hue values of skin and designs a threshold-based detection method for skin color to obtain binary mask images.Then,by means of edge detection and Hough transform for the binary mask images,it could detect relevant profile of the hand and exclude the profile of the face.Finally,by using hand profile obtained from Hough transform,it could restrict the binary mask images of skin color and acquire information of hand area.The test result indicates the proposed method can effectively identify hand information in images and the algorithm is fast in calculation.

关 键 词：人手检测 肤色 HOUGH变换 边缘检测 掩码图像 

分 类 号：TM938.7[电气工程—电力电子与电力传动] TP391.41[自动化与计算机技术—计算机应用技术]