METHOD AND APPARATUS FOR MOVING OBJECT DETECTION

The present invention provides a method for detecting a moving object and apparatus for the same. The image data .DELTA..sub.ij (k)! outputted by subtracter 4 are alternatively contained in image buffers 21 and 22. In subtracter 24, the image data .DELTA..sub.ij (k-1)! of one timing beforehand is removed from the image data .DELTA..sub.ij (k)! of the present time point. In addition, the image data .DELTA..sub.ij (k-1)! of one timing beforehand and the image data .DELTA..sub.ij (k)! of the present time point are compared using comparator 25. In this case, if image data .DELTA..sub.ij (k)! and image data .DELTA..sub.ij (k-1)! differ, the image data .DELTA..sub.ij (k-1)! of one timing beforehand is supplied to subtracter 26. The initial moving object image is excluded by subtracter 26. Consequently, it is possible to precisely distinguish the shape, size, etc. of a moving object without traces (afterimage) of the moving object at the present time point remaining in the final image data .DELTA..sub.ij (k)! .sub.R. According to another method also proposed in the present invention, the image data .DELTA..sub.ij (k-1)! of the moving object are obtained by subtracting the image data D.sub.ij (k-1)! of one timing beforehand or the average image data M.sub.ij (k-1)! of one timing beforehand from the present image data D.sub.ij (k)!, and the image data .DELTA..sub.ij (k-2)! of the moving object are obtained by subtracting the image data D.sub.ij (k-2)! of two timings beforehand or the average image data M.sub.ij (k-2)! of two timings beforehand from the present image data D.sub.ij (k)!. The traces (afterimage) of a moving object is erased by obtaining theoretical product of the image data .DELTA..sub.ij (k-1)! and .DELTA..sub.ij (k-2)!.