Entrance detection of a moving object using intensity average variation of subtraction images

The position of a moving object can be acquired from subtraction of contiguous two images. The centroid of the subtraction image is the center point of the moving object since the intensity values of subtraction image for moving regions are presented with higher density than those of other static regions. So the centroid can be used as a position index of moving object. For the algorithms focusing on position variation(e.g., calculating the velocity of a moving object) position information has to indicate the same point of the object. When the object overlaps with a border line of a camera image plane, the centroid is not the center point of whole object area but the center point of certain part of the object. And it is not identical with the centroid of the object locating in the center region of the plane. In this paper, we observe the average variation of the subtraction image intensity values and find automatically the duration that moving object completely enters in the plane. Effectiveness of the suggested scheme is confirmed by experiments conducted in outdoor circumstances.