Moving Object Detection Algorithm Based on Mixed Gaussian Model and Five-frame Difference
LI Zhanming
TAN Xiangyang
Abstract：The moving target according to the traditional Gaussian mixture model to extract contains lots of holes,uncovered background area easily detected as foreground and Susceptible to light mutation,a improved algorithm combination of the five-frame difference moving target detection method and Gaussian mixture model algorithm is proposed. Compared to the traditional three frames difference algorithm,the improved five frame difference algorithm can extract the foreground region with fewer holes and the image is divided into background region,uncovered region and moving target region through the Gaussian mixture model.Given the large update rate to the background to solve the problem that the uncovered area be detected as moving target area.In the foreground target segmentation stage,only the distribution whose weight is greater than the average weight is used to distinguish the foreground segmentation in order to further reduce the internal cavity.The experimental results show that it is able to extract the complete mov-ing target and the background region contains less noise.
Keywords：Gaussian mixture modelmoving object detectionfive-frame difference methoduncovered background regionforeground segmentation
Publication Date：2018-01-01
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：5( 284-288 )