研究目的
To provide a comprehensive review of state-of-the-art graph-based semi-supervised learning methods for hyperspectral image classification, categorizing techniques and discussing methodologies, challenges, and future research directions.
研究成果
The survey summarizes that existing graph-based methods face issues with multiple overlapping manifolds, computational complexity, and reliance on KNN. Future research should focus on developing sparse representation-based spectral-spatial graph learning with optimization algorithms and band selection to improve efficiency.
研究不足
The review highlights challenges such as overlapping manifolds in high-dimensional data, insufficient discriminant data from KNN methods, computational complexity, and the need for band reduction and optimization algorithms.
