Multiple Moving Objects Segmentation Algorithm Based on Regional Seed
Bingyu Liu · Journal of Chinese Computer Systems · 2013
A multi-object segmentation algorithm with regional seed is proposed for the problem of multiple moving targets segmentation in intelligent monitoring system.Firstly,it uses the background subtraction algorithm for foreground image containing multiple moving targets,and then uses quad-tree decomposition method to contain the sparse matrix corresponding to the foreground image,through the data distribution density of the sparse matrix,calculates the seed points corresponding to the regions which are containing the moving target,starting from these seed points,finally,execute the active contour model in parallel to complete the multiple moving objects segmentation.Experimental results show that the proposed algorithm can effectively segment multiple moving targets and the segmentation results are close to the judgment of the human visual,parallel contour extraction makes our algorithm has good realtime.