SDTN and TRN: Adaptive Spectral-Spatial Feature Extraction for Hyperspectral Image Classification

2025-07-13

AI Summary

To address challenges in hyperspectral image classification—including high-dimensional redundancy, strong spectral-spatial coupling, and scarcity of labeled samples—this paper proposes an adaptive spectral-spatial feature extraction framework. Methodologically, it introduces (1) adaptive tensor decomposition with dynamic rank regularization to jointly model spectral and spatial structures in a low-rank manner, and (2) a lightweight multi-scale tensor residual network (TRN) that fuses hierarchical spectral-spatial features while significantly compressing model parameters. Experimental results on the PaviaU dataset demonstrate that the proposed method achieves an average classification accuracy improvement of 2.3% over state-of-the-art approaches, with a 68% reduction in parameter count. The framework thus achieves an effective trade-off among high classification accuracy, low computational cost, and strong generalization capability—making it particularly suitable for real-time crop monitoring and soil analysis in precision agriculture.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Computer Vision: Remote Sensing / Geospatial AI
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Hyperspectral image classification plays a pivotal role in precision agriculture, providing accurate insights into crop health monitoring, disease detection, and soil analysis. However, traditional methods struggle with high-dimensional data, spectral-spatial redundancy, and the scarcity of labeled samples, often leading to suboptimal performance. To address these challenges, we propose the Self-Adaptive Tensor- Regularized Network (SDTN), which combines tensor decomposition with regularization mechanisms to dynamically adjust tensor ranks, ensuring optimal feature representation tailored to the complexity of the data. Building upon SDTN, we propose the Tensor-Regularized Network (TRN), which integrates the features extracted by SDTN into a lightweight network capable of capturing spectral-spatial features at multiple scales. This approach not only maintains high classification accuracy but also significantly reduces computational complexity, making the framework highly suitable for real-time deployment in resource-constrained environments. Experiments on PaviaU datasets demonstrate significant improvements in accuracy and reduced model parameters compared to state-of-the-art methods.

Problem

• Overcoming high-dimensional data challenges in hyperspectral image classification
• Reducing spectral-spatial redundancy and labeled sample scarcity issues
• Achieving real-time deployment in resource-constrained agricultural environments

Innovation

• Self-Adaptive Tensor-Regularized Network for dynamic feature extraction
• Tensor decomposition with regularization for optimal feature representation
• Lightweight multi-scale spectral-spatial feature integration network