A Dim-Small Target Tracking Approach Based on Robust Hyperspectral Features

Abstract

The dim-small target tracking is a central problem in several related applications. Subjected to the lack of objective information, traditional target tracking is difficult to meet the needs of dim-small target tracking. A dim-small target’s description based on robust spectral features is proposed to solve this problem. By creating a multi-dimensional feature space and extending the limited RGB information to the hyperspectral information, a new type of target’s description is presented to track the dim-small target with hyperspectral features.