Moving Object Detection Based On Low-rank And Sparse Decomposition Of Matrix

Moving object detection is a technique which can accurately extract one or more moving objects from the video,and it’s is an important branch in the field of image processing and computer vision.The accuracy of moving object detection directly impacts on the results of subsequent moving target tracking,moving target recognition and moving target behavior analysis.At present,the moving target detection is widely used in traffic monitoring,weather forecast,satellite image analysis and other fields.Therefore,the research on moving object detection technology has important theoretical significance and application value.Usually,there are three moving object detection methods:frame difference method,background subtraction method and optical flow method.In recent years,the theory of low-rank and sparse decomposition for matrix attracts the high attention of relevant scholars and is applied to moving object detection.This paper focuses on the research of video moving object detection based on low-rank and sparse decomposition of matrix to improve the accuracy of moving object detection and algorithm efficiency.The main innovations include two aspects:firstly,propose an improved video moving object detection method based on low-rank and sparse decomposition of matrix,secondly,considering the difference between adjacent frames,combine edge detection with frame difference to compensate the edge of the moving object.The main contents of this paper include the following aspects:1.This paper elaborates several moving object detection methods:frame difference method,optical flow method,background subtraction method and low-rank and sparse decomposition of matrix theory,and digitally realizes a part of algorithms,and also gives the experiments and analysis.2.Based on the theory study of low-rank and sparse decomposition of matrix and its algorithm,this paper proposes an improved moving object detection method based on low-rank and sparse decomposition of matrix.In this method,the video frames are extracted by equal interval,and the sparse component in the low-rank and sparse decomposition model are further constrained to improve the accuracy of moving detection and algorithm efficiency.Experimental results show that this method can extract the moving object accurately,and is an effective method.3.On the basis of moving object detection method via low-rank and sparse decomposition of matrix,considering the edge is the most important feature of an image,this paper proposes a moving object detection method fused edge and different information between adjacent frame,which combines the advantages of frame difference.It compensates the edge information of moving object,and further improves the accuracy of object detection.We also give the corresponding experimental results and analysis in the paper.All in all,around moving object detection,this paper puts forward and digitally realizes an improved moving object detection method based on low-rank sparse and decomposition of matrix and further introduces the information of edge detection and difference between adjacent frame to compensate the edge information of moving object.This method can not only extract the moving object accurately,but also has high efficiency and occupies less memory space.