Method For Moving Object Detection Based On The Construction Of Feature Frames

Along with the rapid development of science and technology, the extraction of targetarea of the movement and the analysis of the target motion’s characteristics and states invideo image sequence, has become one of the most important branches in the field of imageprocessing and computer. And it has found a solid foundation for better exercise targettracking and estimate motion parameters. In our life, moving targets detection technology arewidely used, such as traffic flow, the control of important place of the monitoring, automaticdriving and military aircraft on the guidance system.Moving object detection algorithm is varied, in this paper, using the same experimentalvideo data respectively experimental it with the various classical algorithm and the algorithmin recent years. The classic algorithm includes background minus division and framedifferential method. The main method of the outlying algorithm is density estimation. Thispaper carries on the analysis of the experiment, sums up the advantages and disadvantages ofthe algorithm, and finally chose the density estimation method of the parameters of the kerneldensity estimation method for the focus.The parameters in sample analysis density estimation modeling got a lot of attention,especially kernel density estimation. But because the kernel density estimation method needto calculate the large amount, applied to movement in the inspection of the target is difficultto reach the real-time effects. This paper proposes a characteristics of building up the frame ofthe kernel density estimation method. Because the kernel density estimation don’t need thatthe density distribution of background model function, all samples and meet with the principleof independent distribution, so can through the features of building up the frame backgroundmodeling method, and application of this method for background updates. The experimentalresults show that the method can adapt the changing of the environment and has a fastcalculation speed and good real-time etc. This method can be used under complex backgroundof monitoring and control system.