Hyperspectral video for illumination-invariant tracking

Abstract

Recent advances in electronics and sensor design have enabled the development of a hyperspectral video camera, which can capture hyperspectral datacubes at near video rates. In this work, we show how high-speed hyperspectral imaging can be used to address several challenging problems in video surveillance. In particular, we combine traditional methods for hyperspectral image analysis and computer vision to achieve illumination-invariant motion detection and object tracking. Experiments using real hyperspectral video images are provided.