Small Moving Target Detection Model Based on Improved Mixed Disturbance Gauss Model

Abstract: In order to obtain ideal moving target detection results, a kind of moving target detection algorithm of improved mixture Gaussian model is proposed in this paper on the basis of analyzing the deficiency oftraditional Mixture Gaussian model. Firstly, in this algorithm, the update rate is increased to better adapt to the change of environment and the idea of pixel connected region is introduced to remove the noise, then the invariance property of chroma is adopted to avoid arising of “shadow” phenomenon; finally, simulation comparison experiment is used to test its performance. The experiment results show that this algorithm can generate reliable background image and accurately and completely detect the moving target to obtainmore superior detection results of comparison algorithm, which can meet the practical application requirement of target tracking.
Keywords: gaussian mixture model; target tracking; learning rate; shadow inhibition
Author: Ting Bai, Jinwen Tian, Xiao Sun
Journal Code: jptkomputergg160025

Artikel Terkait :