Improved surendra background updating algorithm for moving vehicle detection

Abstract

Moving object detection plays an important role in intelligent transportation system. Combined with dynamic segmentation threshold and the counter, this paper proposes an improved Surendra algorithm based on background subtraction method. Image sequence is firstly preprocessed by median filtering, followed by background model updating on basis of dynamic background subtraction. Counter is incorporated to offset the effects of stagnation of moving vehicles and camera vibration. Mathematical morphology operations including connectivity components analysis and iterative dilating are applied to remove noises and fill holes. Experimental results show that the proposed method is efficient and can perform in real time.