Hyperspectral Nonlinear Unmixing by Using Plug-and-Play Prior for Abundance Maps

Lianru Gao,Michael K Ng,Zhicheng Wang,Bing Zhang,Andrea Marinoni,Lina Zhuang

doi:10.3390/rs12244117

Abstract

Spectral unmixing (SU) aims at decomposing the mixed pixel into basic components, called endmembers with corresponding abundance fractions. Linear mixing model (LMM) and nonlinear mixing models (NLMMs) are two main classes to solve the SU. This paper proposes a new nonlinear unmixing method base on general bilinear model, which is one of the NLMMs. Since retrieving the endmembers’ abundances represents an ill-posed inverse problem, prior knowledge of abundances has been investigated by conceiving regularizations techniques (e.g., sparsity, total variation, group sparsity, and low rankness), so to enhance the ability to restrict the solution space and thus to achieve reliable estimates. All the regularizations mentioned above can be interpreted as denoising of abundance maps. In this paper, instead of investing effort in designing more powerful regularizations of abundances, we use plug-and-play prior technique, that is to use directly a state-of-the-art denoiser, which is conceived to exploit the spatial correlation of abundance maps and nonlinear interaction maps. The numerical results in simulated data and real hyperspectral dataset show that the proposed method can improve the estimation of abundances dramatically compared with state-of-the-art nonlinear unmixing methods.

Highlights

Hyperspectral remote sensing imaging is a combination of imaging technology and spectral technology
The details of the simulated data can be obtained with the previous steps, we generated a series of noisy images with signal-to-noise ratio (SNR) = {15, 20, 30} dB to evaluate performance of the proposed method and compare with other methods
Note that a direct compasion with FCLS unmixing results is unfair and FCLS is served as a benchmark, which shows the impact of using a linear unmixing method on nonlinear mixed images

Summary

Introduction

Hyperspectral remote sensing imaging is a combination of imaging technology and spectral technology. Benefiting from the rich spectral information, hyperspectral images (HSIs) can be used to identity materials precisely. For NLU, various models have been proposed to describe the mixing of pixels, taking into account the more complex reflections in the scene. They are the generalized bilinear model (GBM) [15], the polynomial post nonlinear model (PPNM) [16], the multilinear mixing model (MLM) [17], the p-linear model [18], the multiharmonic postnonlinear mixing model (MHPNMM) [19], the nonlinear non-negative matrix factorization (NNMF) [20] and so on. The NLU is a more challenging problem than LU, and we mainly focus on the NLU in the paper

Objectives

Methods

Results

Conclusion