Moving object detection in the low illumination night scene

Images captured at night for visual surveillance have low SNR, low contrast, poor distinction between the object and the background, that pose more challenges to moving object detection. In this paper, a moving object detection algorithm for night surveillance based on dual-scale Approximate Median Filter background models is proposed. Firstly, to reduce the image noise, a low resolution image is reconstructed by block-wise down-sampling the original image, and used to coarsely detect the ROI of moving objects; then the moving object contour is refined by using coarse detected result ROI and the original image. Experiments show that the proposed algorithm can detect moving objects robustly at night under adverse illumination conditions, with high spatial resolution, low computational complexity and high resistant to noise.