Research On Moving Object Detection Based On EMP In Moving Camera

Moving target detection has always been one of the main subject in the field of computer vision and intelligent monitoringmost of the cameras in the actual scene are stationary.That i s,the detection of moving target in the scene of static camera.However,with the development of science and technology,the moving target detection under the scene of dynamic camera ha s been paid more attention by more researchers.In particular,the role of motion detection in t he field of machine vision and drone driving is particularly evident.But compared to the static camera,how to solve the separation of foreground and backgr ound when the camera and the target are both in motion has become one of the most difficult points for the detection of moving targets in the context of motion cameras.To solve the abov e problem,this article analyses the stability of feature points in different locations in matching.In particular,it is easy to appear error matching between the matching of the feature points o f the target and the background interface.This article based on feature points of these errors m atching,it proposes an EMP-based(EMP is fully called Error Match Point,that is,Error matc hing point)moving target detection algorithm,then get the moving target area based on EMP.Finally,combined with Grabcut algorithm,the detection of moving targets in the context of motion camera is realized.The main research content of this article has the following aspects:(1)The principle of EMP is analyzed,and giving its definition.The extraction method and process of EMP is presented.(2)In this paper,the combination of SIFT algorithm and ORB algorithm in space is proposed to enrich and improve EMP.In addition,the enrichment and improvement of EMP on different time scales are proposed.(3)In the five frame processing method,RANSAC algorithm was introduced to screen,and the extracted EMP was more stable.Through the vector matching method,the EMP in the background is removed.(4)The Grabcut algorithm is guided by the target information determined by the EMP,thereby realizing the detection of the moving object in the motion camera scenario.According to the erroneous matching feature points generated between consecutive frame matching in a motion camera scenario.Finally,Moving objects detection was realized.According to the different experimental results.The algorithm of this paper based on a variety of camera motion modes.All of them can be well extracted to EMP.and ultimately can achieve better test results.