Research On Motion Extraction Of Moving Background On Smart Mobile Device

Motion detect is to extract object by dividing motion region and background region using detection algorithm.However,change of light,complex background and speed of moving object make challenge for object detection,the accuracy of which directly influence the result of moving taregt recognition,tracking and other abnormal behavior analysis.Traditional cameras has certain limitaion to more and more complex social activities,Thanks to the generation of smart mobile,which has made more and more influence on our life,but also accelerating motion target detection intelligent and integrated.However,which will causing relative rest between foreground and banckground region,making new challenge for moving target detection,so this article research is focused on moving target detection of small movement based on intelligent mobile device.Having studied a series of classical algorithm,this paper focus on the small moving of intelligent mobile terminal.Moving target detection under the moving background is established on the basis of the static background,so in this paper,the research of moving object detection based on the intelligent mobile terminal vdeo in this paper include 3 aspects:(1)algorithm of static background algorithm based on real-time and accuracy;(2)moving object extraction algorithm based on moving background;(3)the application of moving object detection algorithm on intelligent mobile terminal.Considering the real-time and accuracy of the algorithm in Static background,this paper propose a fusion of ViBe algorithm and temporal difference method.Firstly,initialize ViBe background images from a single frame and build a ViBe model with real observed pixel values.At the same time,frame difference algorithm has been preliminary used to extract potential moving object area,detecting moving object in the potential area again by ViBe algorithm to make up for the hole effect caused by frame difference algorithm in order to extract moving object more accuratly.this algorithm will be employed in the following dynamic background.Aimed the problem that mobile terminals are prone to tiny movement,this paper overcome this problem by compensate the background.Firstly,feature points is extracted by Harris operator and descriptied by SIFT operator.In order to calculate the motion parameter of camera accurately,it is important to match the Feature points by RANSAC algorithm,calculating the motion parameter by affine parameters matrix so as to compensate the background image,finally making the moving background static,extracting moving object by the fusion algorithm that has been proposed before.After coding in eclipse,this algorithm has been apply in smart phone based on Android platform.Experimental results show that daynamic background can be effectively resisted and a precise extraction be achieved.