基于轮廓方向特征的头部特征识别  

Method for Recognizing Human Head Based on Contour Directional Feature

作　　者：黄杰贤[1] 黄志平 杨冬涛[1] 欧阳玉平[3] 

机构地区：[1]广东嘉应学院电子信息工程学院,广东梅州514015 [2]广东振声科技股份有限公司,广东梅州514000 [3]华南农业大学工程学院,广东广州510642

出　　处：《电视技术》2015年第12期107-112,共6页Video Engineering

基　　金：广东省重大科技专项计划项目(2012A080104012)

摘　　要：针对危险生产作业场所的定员管理问题,采用机器视觉技术对进出作业场所的人数进行在线、实时统计。首先引入纹理梯度算子,定义方向熵函数以获取人头轮廓的方向特征并对轮廓点进行编码;接着,基于标准圆的轮廓及其方向二元特征实现对人头轮廓的准确匹配;最后建立弧长置信度函数完成对人员目标的定位。在目标跟踪过程中,建立预测函数辅助人头运动轨迹的获取,最终根据运动轨迹线判断目标是否进入或离开生产作业场所。实验表明,提出的方法准确率在90%以上,具有良好的实用性与学术价值,对于确保安全生产具有重大现实意义。A statistical method of headcount based on machine vision technology is proposed in the research, which is implemented in quota management in dangerous production place. The texture gradient operator is firstly introduced to define directional entropy fimction which is used to extract directional feature of head＇s contour and realize co,ltour coding. So,head＇s contour can be accurately matched based on standard circle＇s contour and directionality dual characteristic. And then, arc confidence function is finally constructed to complete the location of people. During the process of moving objective tracking, moving production function is set assisting to obtain head＇s moving trajectory. So whether the people enter or leave production place can be judged according to trajectory. Finally, the experiment proves the practical effectiveness and academic value of the method, the accuracy can reach 90% above. It is also very significant for safety in production.

关 键 词：机器视觉 危险生产作业场所 定员 方向熵 弧长置信度 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]