Semi-NMF-Based Reconstruction for Hyperspectral Compressed Sensing

Abstract

Hyperspectral compressed sensing (HCS) is a new imaging method that effectively reduces the power consumption of data acquisition. In this paper, we present a novel HCS algorithm by incorporating spatial-spectral hybrid compressed sensing, followed by a reconstruction based on spectral unmixing. At the sampling stage, the measurements are acquired by a spatial-spectral hybrid compressive sampling scheme to preserve the necessary information for the following spectral unmixing, where spatial compressive sampling mainly retains the endmember information, and spectral compressive sampling mainly retains the abundance information. Due to the limitations of the traditional linear mixed model (LMM), an improved mixed model is proposed for HCS reconstruction, which considers spectral variability, nonlinear mixing and other factors. At the reconstruction stage, based on the improved mixed model, semi-nonnegative matrix factorization (semi-NMF) is introduced to realize spectral unmixing on the measurements to achieve the final reconstruction by using an alternate iteration manner. The proposed algorithm is tested on real hyperspectral data and the selection of parameters is fully analyzed. Experimental results demonstrate that the proposed algorithm can significantly outperform state-of-the-art HCS algorithms in terms of reconstruction performance.

Low-Rank Tensor Optimization With Nonlocal Plug-and-Play Regularizers for Snapshot Compressive Imaging

“…The underlying principle of the SCI system [41]- [44] is to compress over 30 channels of data on a 2D detector with the temporal-variant mask for high-speed imaging [41] and the dispersion for spectral imaging [19], [43]. These SCI systems compress the 3D cube across the third dimension, which is different from the random sampling in CS systems [45], [46]. The degradation of the SCI process [27], [41], [43] is usually formulated as the vector form, i.e.,…”

Survey on compressed sensing reconstruction method for 3D data

“…In another hand, Wang et al 104 improves the linear mixture model by considering factors such as spectral variability and nonlinear mixing, and retain more prior knowledge by semi‐NMF. Vargas et al 105 proposed a linear fusion model and an optimization algorithm based on a fast coordinate descent strategy to solve the HSI fusion problem of different spectral and spatial ranges.…”

RGB-Based Compressed Medical Imaging Using Sparsity Averaging Reweighted Analysis for Wireless Capsule Endoscopy Images

“…This approach minimized the issue related to a linear mixed model (LMM) such as surrounding circumstances, configurations of the instruments, and tangible nonlinear mixing impacts [12]. Furthermore, LMM model was enhanced using seminonnegative matrix factorization for the reconstruction of the HSI which is performed by fusing spatial-spectral (SS) hybrid CS [15]. In the restoration process, fidelity restraints were proposed for an objective function in spectral and spatial estimations [16].…”