基于运动目标检测的工业安全系统设计  

Design of Industrial Safety System Based on Moving Target Detection

作　　者：包晓敏 李平 BAO Xiao-min;LI Ping(School of Information Science and Technology,Zhejiang Sci-Tech University,Hangzhou 310018,China)

机构地区：[1]浙江理工大学信息学院,浙江杭州310018

出　　处：《软件导刊》2021年第4期158-162,共5页Software Guide

基　　金：浙江省重点研发计划项目(2018C02027)。

摘　　要：为改善工业生产车间安全防护手段低效、防护设备造价高等缺点,基于运动目标检测技术设计安全防护系统,通过摄像头采集的视频信息对图像进行预处理,利用OpenCV软件库,将相邻三帧图像进行两两差分运算得到差分图像,依据最大模糊熵准则自适应选取阈值,对差分图像进行二值化,对所得二值化差分图像进行逻辑“与”运算,检测运动目标。系统还可根据距离判断运动目标是否进入设定的危险区域,从而实现自动报警。实验结果表明,改进的最大模糊熵阈值三帧差算法准确率、系统自动报警功能正检率较高,分别达到98.8%、96.7%。该安全系统能够实现危险自动报警,保障工业安全生产,具有一定的应用价值。In order to improve the shortcomings of the current industrial production workshop,such as inefficient safety protection methods and high protection equipment cost,Based on the moving target detection technology,a set of security protection system is de⁃signed.The video information collected by the camera is used to preprocess the image.Using the OpenCV software library,the adjacent three frames of the image are subjected to pairwise difference operation to obtain the difference image,according to the maximum fuzzy entropy criterion The threshold value is adaptively selected,the difference image is binarized,and the resulting binary difference im⁃age is logically AND to detect the moving target.According to the distance,it is judged whether the moving target enters the set danger zone to realize automatic alarm.The experimental results show that the accuracy of the three-frame difference algorithm based on the improved maximum fuzzy entropy threshold and the positive detection rate of the system's automatic alarm function are relatively high,reaching 98.8%and 96.7%.The safety system based on the detection of moving targets can realize the automatic danger alarm,which guarantees the industrial safety production and has certain use value.

关 键 词：运动目标检测 模糊熵 差分图像 OPENCV 安全防护 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]