Hyperspectral Image Dimensionality Reduction Algorithm Based on Spatial–Spectral Adaptive Multiple Manifolds

Hyperspectral images contain rich spatial–spectral information and have high dimensions, which can lead to challenges related feature extraction for classification tasks, resulting in suboptimal performance. We propose a hyperspectral image dimensionality reduction algorithm based on adaptive multiple manifolds address the problem of small differences between features dissimilar samples subspace caused by uniform projection transformation traditional methods. Firstly, spatial boundary mismatch problems re-characterizing pixel using pixels fixed area around it as its near neighbors algorithms, an weight representation method super-pixel segmentation is proposed, enhances similarity similar dissimilarity samples. Secondly, that single manifold cannot completely characterize neighbor different categories, multi-manifold proposed. The entire data low-dimensional obtained adaptively fusing intra- inter-manifold maps constructed each category spectral dimensions. Experimental results two public datasets show proposed achieves better when performing task.