An improved object detection algorithm based on SIFT features matching in dynamic background

Moving object detection under the conditions of camera moving have considerable difficulty and challenge. The problems are usually dealt with by the algorithm based on SIFT features matching. But this algorithm is time consuming and with some defects in the integrity of the detection results, which cannot detect the real moving objects from the dynamic background efficiently. Therefore, we present an improved algorithm which has better performances in speed and accuracy. Firstly, we introduce an efficient new method that constructs circular area around the feature point and uses the key direction to correct the gradient directions for SIFT descriptors generating. Then we use a local feature points matching instead of the global matching algorithm. Finally we use the ViBe method which has a good performance to detect foreground objects. Experimental results show that the proposed algorithm significantly improves the efficiency and accuracy of the moving object detection. © 2015 by Binary Information Press.