Abstract: Hyperspectral data with high dimensionality always needs more storage space and increases the computational consumption, manifold learning is extensively used in dimensionality reduction. A ...