Hybrid Moving Object Detection Algorithm with Controlled Temporal Scale Frame Difference

Abstract

Moving object detection is the first stage in any video surveillance system. It is followed by two other stages; tracking and classification. The accuracy of the later two stages strongly depends on the first one. This paper proposes a hybrid moving object detection algorithm, which combines controlled temporal scale frame difference and selective running average as a method of background subtraction. In order to overcome the main drawbacks of traditional temporal difference methods that depend strongly on the speed of the moving object, we propose a new algorithm that controls the temporal scale between successive frames to solve the problem of emptiness phenomenon, whenan object moves slowly. Experimental results show that the proposed algorithm outperforms other commonly used object detection algorithms.