A Fast Hyperspectral Object Tracking Method Based On Channel Selection Strategy
Yifan Zhang, Xu Li, Fei–Yue Wang, Baoguo Wei, Lixin Li
Abstract
Hyperspectral object tracking aims to take advantage of the rich spatial and spectral information in hyperspectral videos to effectively improve the robustness and accuracy of object tracking. Compared with color videos, hyperspectral videos have huge amount of data bringing a challenge to the efficiency of object tracking. We propose a fast hyperspectral object tracking method based on channel selection strategy. The strategy considers the spatial and spectral changes of local regions in the frame image and selects only three channels fed to tracker to speed up. The experimental results show that our method reaches 11.5 FPS on the dataset of Hyperspectral Object Tracking Challenge, which is faster than the state-of-the-art methods.