AUTHOR(S): 

Caroline Fossati, Salah Bourennane

TITLE

Multidimensional Signal Processing Methods: Target Detection Methods Based on Tensor Decompositions

KEYWORDS

Prewhitening, Hyperspectral image, small target detection, multiway Wiener Filter

ABSTRACT

Reducing noise is an important preprocessing step for further analyze the information in the hyperspectral image (HIS). Commonly, filtering methods for HSIs are based on the data vectorization or matricization while ignore the related information between different bands. So there are new approaches considering multidimensional data as whole entities based on tensor decomposition. However, it can not cope with the HSIs disturbed by nonwhite noise which is the most cases in the actual world and cannot preserve small targets. In this paper, we propose a new method for the reduction of non-white noise from images. The first step of this method is to change the noise in HSIs being a white one through a prewhitening procedure(PW). Then multidimensional wavelet packet transform with multiway Wiener filter (PW-MWPT-MWF) to improve the target detection efficiency of HIS with small targets in the noise environment. The performances of the our method (PW-MWPT-MWF) are exemplified using simulated and real-world HIS

Cite this paper

Caroline Fossati, Salah Bourennane. (2017) Multidimensional Signal Processing Methods: Target Detection Methods Based on Tensor Decompositions. International Journal of Signal Processing, 2, 67-71