Temporal difference method based on positive and negative energy distribution in moving objects detection

摘要

The traditional temporal difference method under dynamic background will not extracted moving objects. In this paper, we propose a method based on positive and negative energy distribution and combine the LK optical flow method to achieve effective moving objects detection. Using series frames obtain the positive and negative energy distribution. The Gaussian kernel density function is used to estimate the density of the energy distribution histogram of the second frame. When the kernel density is smaller than the threshold value, it is considered to be the background movement with smaller energy. While the kernel density is larger than the threshold value, it is considered to be the moving objects with larger energy. Experimental results demonstrate that the method effectively detect the moving target in the dynamic background.