Abstract:We proposed an algorithm combining the Gaussian background mixture modelling and three-frame difference to tackle the problem of traditional object detection for its low processing capacity in an environment of sudden light change.Based on Gaussian background mixture modelling,this method determines whether there is the change in light using statistical method,and uses three-frame difference method to detect the targets in an environment of sudden light change,meanwhile it divides the detecting region into two areas,the static and the dynamic.For static area it detects every discontinuous frame so as to reduce computational load.Experimental results show that the proposed algorithm can well detect the moving objects in the environment of sudden light change.

RESEARCH ON MOVING TARGET DETECTION IN AN ENVIRONMENT OF SUDDEN LIGHT CHANGING