Novel Data-Driven Spatial-Spectral Correlated Scheme for Dimensionality Reduction of Hyperspectral Images

Hyperspectral imaging technology has been popularly applied in remote sensing because it collects echoed signals from across the electromagnetic (EM) spectrum and thereby contributes fruitfully spatial-spectral information. However, processing or storage of high-data-volume hyperspectral images (HSIs), also viewed as snapshots varying with EM spectrum, burdens hardware resources, especially for high spectral resolution spatial cases. To address this challenge, a novel unsupervised dimensionality reduction method based on dynamic mode decomposition~(DMD) algorithm is proposed to analyze data. This decomposes HSIs terms modes corresponding patterns. Then, these patterns are combined reconstruct raw via low-rank model. Furthermore, we extend DMD data tensor form title CubeDMD actualize compression horizontal, vertical, dimensions. Our data-driven scheme benchmarked by real measured at Salinas scenes Pavia University. It demonstrated can be reconstructed accurately effectively The mean peak signal-to-noise ratio (PSNR) between original reach 31.47 dB, angle mapper (SAM) only 0.1037. work provides useful tool analysis representation.