Spectral-Spatial Method for Hyperspectral Image classification in Noisy Environment

Abstract

Target detection is an important issue in the HyperSpectral Image (HSI) processing field. However, current spectral-identification-based target detection algorithms are sensitive to the noise and most denoising algorithms cannot preserve small targets, therefore it is necessary to design a robust detection algorithm that can preserve small targets. Because the signal-dependent photonic noise has become as dominant as the signal-independent noise generated by the electronic circuitry in HSI data collected by new-generation hyperspectral sensors, the reduction of the additive signal-dependent photonic noise becomes the focus of the current research in this field. To reduce the opto-elctronic noise from HSIs, a new method is developed in this paper. Firstly, a pre-whitening procedure is proposed to whiten noise in HSIs. Secondly, a three-dimensional wavelet packet transform (3-WPT) in tensor form is presented to find different component tensors of HSI. Then, to jointly filter a component tensor in each mode, multiway Wiener filter (MWF) is introduced. Moreover, to determine the best transform level and basis of 3-WPT a risk function is proposed. The effectiveness of our method in classification is experimentally demonstrated on a real-world HSI acquired by airborne sensor.