基于视频的施工升降机乘客头部识别方法研究

摘要

In view of the current number of elevator is given priority to with artificial,manmade interference,a approach based on video processing to counting the number of the people in the elevator is presented.This paper mainly studied the head identification.Using the template matching algorithm finds the position that is similar to the template.For the two kinds of false alarm,we use moving region detection and Hu moment invariants to remove them.The experimental results show that using this method can effectively remove the false alarm,can meet the demand of practical application.%针对目前人数统计检测主要以人工为主，存在人为因素干扰的问题，提出一种基于视频处理的升降机内人数统计方法，本文主要对系统中头部识别方面进行了研究，利用模板匹配算法寻找视频帧中与模板相近的人头位置，针对出现的两种虚警误判，分别采用运动区域检测和Hu不变矩形状匹配进行去除。实验结果表明，利用该方法能够有效的去除虚警，符合实际应用需求。