Statistical Properties of Hyperspectral Images

摘要

The aim of the paper is to analysis hyperspectral images using tensor principal component analysis for multi-way data sets. The method extracts statistically dominants information from hyperspectral images. Because of high resolution sampling in colour channel, images observed by hyperspectral camera system hyperspectral are expressed by the 3-mode tensors. Tensor principal component analysis allows to extract dominant light channel information from hyperspectral images.