Motion detection using background constraints

摘要

We introduce a technique for detecting moving objects from a moving camera using the background constraint. The camera is mounted on a platform that can both rotate and translate. Motion is detected by computing a mapping that corresponds pixels in successive images whenever possible. Assuming the background is planar provides knowledge that is used to recover projected local velocity (u, v). Background compensation can be used to eliminate the effects of rotation and translation on the mapping function. To increase the robustness of this technique, false motion caused by inaccuracies in sensor readings are eliminated. This is achieved by using a morphological filter, which consists of two successive operations, erosion and dilation, performed on the motion detection image. Experimental results with real images are presented, which show the robustness of the proposed algorithm in detecting independently moving objects from a mobile platform.