Application of Bayesian Object Tracking in Substation Job Safety Management
By reading the data from substation video surveillance system and locating the moving objects in the video footage,the object tracking technology can be applied to realize substation job safety management.This paper proposes the object tracking technology based on Bayesian principles.Thereby,describe the target appearance with multiple local image features such as gray,direction gradient and local binary pattern; dynamically fuse multiple features based on chunk incremental discriminant analysis; achieve the optimal separation of target and background.Moreover,apply iterative evolution operator resampling to maximize the number of effective particles in order to maintain the diversity of particles and achieve stable tracking.The target tracking experiments on Chengdu 220 kV substation show that this method is superior in tracking stability in the case of background illumination,target attitude change and partial occlusions,quite applicable to substation job safety management.
NETL
NSTL
万方数据
维普
