井下人数统计的视觉分析方法设计  

Design of Visual Analysis Method for Statistics of Underground Mining Personnel

作　　者：袁海娣[1] 姚蓓蓓[2] YUAN Haidi;YAO Beibei(Fondation Expriment Teaching Center, Anhui Sanlian Collage, Hefei, Anhui 230601, China;Department of Chemical and Materials Engineering, Hefei University, Hefei, Anhui 230601, China)

机构地区：[1]安徽三联学院基础实验教学中心,安徽合肥230601 [2]合肥学院化学与材料工程系,安徽合肥230601

出　　处：《宜宾学院学报》2018年第12期24-28,共5页Journal of Yibin University

基　　金：安徽省省级实验实训中心项目(2014SXZX022);校级平台重点项目"基于改进布谷鸟算法的油墨预置技术研究"(PTZD2017009)

摘　　要：运用井下人员视频计数的计算机智能视觉监控方法,首先生成背景图,然后用背景减影法提取运动目标区域,再利用安全帽的颜色特征缩小提取、分割运动目标,计算出运动目标,并完成形态学处理.最后采用区域质心匹配法实现目标跟踪、匹配与计数.实验证明该算法简单可行,能够有效克服矿井复杂环境影响,并能很好地满足实时性的要求.Using the computer intelligent visual monitoring method of the downhole personnel video counting,the background map was generated first,then the moving target area was extracted with the background subtraction method,the moving target was extracted and divided by the color feature of the helmet,the moving target was calculated,and the morphological processing was completed.Finally,the region centroid matching method was used to achieve target tracking,matching and counting.Experiments show that the algorithm is simple and feasible,and can effectively overcome the complex environmental impact of mines,and can meet the real-time requirements well.

关 键 词：智能视觉 背景减影法 区域质心匹配 Kalman滤波 最大类间方差法(OTSU) 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]