A Multi-Layer Background Subtraction Based on Gaussian Pyramid for Moving Objects Detection

In this paper, a real-time multi-layer background subtraction based on Gaussian pyramid is proposed for moving object detection. The proposed method models background on two levels: region analysis in the high-resolution level with averaging background model and pixel analysis in the low-resolution level with hierarchical non-parametric kernel density estimation method. The new method has lower time and space complexities and is more effective than Elgammal’s method. Meanwhile, time factor is introduced to refine foreground, and a novel background updating strategy is proposed to adapt to the changes in the scene. Experiment results on both the public video database and our own video database show that the proposed approach has good accuracy and speed, especially against drastic camera shaking. KeywordsBackground Subtraction; Kernel Density Estimation; Motion Detection; Gaussian Pyramid