Hyperspectral Unmixing with 3D Convolutional Sparse Coding and Projected Simplex Volume Maximization

2025-12-05

AI Summary

Hyperspectral unmixing (HSU) aims to identify endmembers and estimate their fractional abundances from mixed pixels. This paper proposes 3D-CSCNet, an algorithm-unrolled deep network based on 3D convolutional sparse coding autoencoding, which jointly models spectral–spatial features. Its core innovations are: (i) embedding an interpretable 3D sparse coding module into a deep architecture, and (ii) introducing a projection-based simplex volume maximization (PSVM) strategy for endmember matrix initialization, enabling joint optimization of endmembers and abundances. The resulting framework balances physical interpretability with strong representational capacity. Extensive experiments on three real-world and one synthetic hyperspectral dataset demonstrate that 3D-CSCNet consistently outperforms state-of-the-art methods across varying signal-to-noise ratios, validating its robustness and superiority.

Technology Category

• Computer Vision: 3D Computer Vision
• Search and Optimization: Mixed Discrete/Continuous Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Web Mining and Content Analysis: Web data integration and cleaning
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

Hyperspectral unmixing (HSU) aims to separate each pixel into its constituent endmembers and estimate their corresponding abundance fractions. This work presents an algorithm-unrolling-based network for the HSU task, named the 3D Convolutional Sparse Coding Network (3D-CSCNet), built upon a 3D CSC model. Unlike existing unrolling-based networks, our 3D-CSCNet is designed within the powerful autoencoder (AE) framework. Specifically, to solve the 3D CSC problem, we propose a 3D CSC block (3D-CSCB) derived through deep algorithm unrolling. Given a hyperspectral image (HSI), 3D-CSCNet employs the 3D-CSCB to estimate the abundance matrix. The use of 3D CSC enables joint learning of spectral and spatial relationships in the 3D HSI data cube. The estimated abundance matrix is then passed to the AE decoder to reconstruct the HSI, and the decoder weights are extracted as the endmember matrix. Additionally, we propose a projected simplex volume maximization (PSVM) algorithm for endmember estimation, and the resulting endmembers are used to initialize the decoder weights of 3D-CSCNet. Extensive experiments on three real datasets and one simulated dataset with three different signal-to-noise ratio (SNR) levels demonstrate that our 3D-CSCNet outperforms state-of-the-art methods.

Problem

• Develops a 3D convolutional sparse coding network for hyperspectral unmixing
• Estimates endmembers and abundances by learning spectral-spatial relationships jointly
• Proposes a projected simplex volume maximization algorithm to initialize endmembers

Innovation

• 3D convolutional sparse coding network for hyperspectral unmixing
• Autoencoder framework with algorithm-unrolling-based 3D-CSC block
• Projected simplex volume maximization for endmember initialization