The density estimation of group targets based on state information

In view of the problems of unstable image quality, large demand for training data and high computational cost, which are existed in the method of group targets density estimation based on image information, a new method based on state information and combined with statistics is proposed. Specifically, the state information derived from target tracking with GLMB filter is used as sample data, and the kernel density estimation (KDE) with adaptive bandwidth is introduced to calculate the probability density function (PDF) and visualize the density distribution of group targets. The results of simulation show that the proposed algorithm can effectively visualize the density distribution of group targets, and the mean square error can be reduced by nearly 60% after the introduction of adaptive bandwidth.