The Research Of Moving Target Detection Method Based On Three-frame Difterence

Moving object detection studies how to extract moving object region in image sequence. It is one of active branches of image processing and computer vision. This technology has been applied widely in many ways such as traffic flow control, human tracing, automatic drive, and military application.The frame difference method is contrasted two different time images which have the same background, and its result can reflect the movement of a motion object in this background. According to this thought, it introduces three-frame difference method which is a common motion target detection method. First, it makes a difference between the current frame image difference and the previous frame in image video sequences. Second, it makes a difference between the next frame image and the current frame. Third, using these two differences to make an and operation. Then, based on the gray value of the moving object detection change region to set a threshold, so that we can getting a better location of moving object in the video images. From the descriptions above and we could detect moving targets.According to Three-frame difference method, this paper researches two methods on moving object detection:1. A moving object detection method based on three-frame difference and threshold segmentation is proposed. First the one dimensional cross entropy threshold segmentation and two-dimensional cross entropy threshold segmentation method are discussed in greater detail, second deduced the fast two-dimensional cross entropy threshold segmentation algorithm, third a moving object detection method based on three-frame difference and threshold segmentation is proposed, then using three-frame difference method calculate three consecutive frame images in the video. Finally using the fast2D cross-entropy threshold segmentation algorithm to the results which is obtained from the previous step. This method not only satisfied the requirements of real-time but also got accurately motion target detection results.2. A moving object detection method based on three-frame difference and data fusion is proposed. First using the operator Kirsch to detect the edge of the video image. Then, combining the detection image and the detection image that obtained from the first method to get a fusion image. Last, the fusion image was processed through morphology to get moving target. This method not only satisfied the requirements of real-time but also did get motion target detection results accurately and completely.