Hough Transform Track Initiation Algorithm Based on Two-stage Clustering with Motion Constraints

Li Jing,Pan Quan
DOI: https://doi.org/10.16526/j.cnki.11-4762/tp.2011.11.041
2011-01-01
Abstract:Aiming to solve the problem of high false alarm and seriously missing target tracking from the Track Initiation Algorithm based on Hough Transform and Subtractive Clustering,a new Hough Transform track initiation method based on two-stage clustering with motion constraints is proposed.In this algorithm,first,the Object Method is used to filtrate parts of clutter,then the Hough Transform is adopted and filtrate the data by a low threshold values,following,the subtractive clustering is used to obtain the number and parameters of the target tracking,at last,the Nearest Neighbor method is used to decide which cluster center each sample point belong to.The simulation shows that under the density clutter environment,the initial probability of the proposed algorithm is 1.2 time than that of multi-Hough track initiation and the false track initiation probability of the present method is one third of the multi-Hough track initiation algorithm,which show that the present algorithm can improve the performance of track Initiation effectively,especially detect the targets in density clutter and low SNR environment.
What problem does this paper attempt to address?