Research on background segmentation method for sports video

Background difference method is important method of moving object in video image sequence, in order to solve the problem such as illumination change, in the background of each sentenced to establish single Gaussian model for background pixel, is put forward through a hypothesis test method to estimate the IIR filter background updating coefficient alpha, updates to the background. And the traditional video image background subtraction model on the basis of the analysis of the background noise and movement target grey value distribution characteristics, in view of the traditional single threshold segmentation of background difference method of some deficiencies, this paper puts forward a more reasonable model of background subtraction: through the bright and dark sides of background difference image separate processing, at the same time use Otsu (the betweencluster variance) algorithm to find out about the two best threshold to achieve better video image segmentation. The experimental results show that the method proposed in the research has lower complexity, high computational effectiveness and low computational amount. It can get better extraction effects on video image. © 2016 UK Simulation Society. All rights reserved.