融合改进帧差和边缘提取算法的运动目标检测  

Moving Target Detection Based on Improved Frame Difference and Edge Extraction Algorithm

作　　者：杜文汉 李东兴[1] 王倩楠 武帅 DU Wen-han;LI Dong-xing;WANG Qian-nan;WU Shuai(School of Mechanical Engineering, Shandong University of Technology, Zibo 255000, China)

机构地区：[1]山东理工大学机械工程学院,淄博255000

出　　处：《科学技术与工程》2022年第5期1944-1949,共6页Science Technology and Engineering

基　　金：国家自然科学基金(62076152,51705296)。

摘　　要：针对帧差算法检测运动目标存在的目标边缘缺失和空洞的问题,提出一种将改进帧差和基于改进Canny算子的运动目标边缘提取方法相融合的运动目标检测算法。首先,使用帧差法快速检测变化区域,并通过面积阈值判断是否进行当前帧的检测以提高算法实时性;其次,在运动目标边缘提取中,提出一种基于改进Canny算子的运动目标边缘提取方法提取当前帧运动目标边缘,以解决检测目标边缘缺失的问题;最后,将改进的三帧差分法提取运动目标像素点填充目标边缘图像,以解决检测目标内部空洞的问题。对比实验结果表明,该融合算法对视频中的运动目标能以较高的准确度和完整度实现运动目标的高效检测。Aiming at the lack of target contours and holes of moving target detection by frame difference algorithm,a moving target detection algorithm combined improved frame difference method with moving target edge extraction method based on improved Canny operator was proposed.Firstly,the frame difference method was used to quickly detect the changing region,and the area threshold was used to judge whether to detect the current frame,so as to improve the real-time performance of the algorithm.Secondly,in the process of moving target edge extraction,a moving object edge extraction method based on improved Canny operator was proposed to extract the current frame moving object edge to solve the problem of missing the target edge.Finally,the improved three-frame difference method was used to extract moving target pixels and fill the target edge image to solve the problem of detecting target holes.The comparison experimental results show that the algorithm can achieve high-efficiency detection of moving targets in the video with high accuracy and integrity.

关 键 词：运动目标检测 边缘提取 帧间差分 自适应阈值 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]