Hyperspectral data fusion technology for weak small sea surface target

Aiming at solving problem of weak small target missing during dimension reduction and fusion of hyperspectral data, hyperspectral data fusion method for weak small target is studied in this paper. Class pixel matrix is produced by similarity measures, dimension reduction projection matrix is calculated by principal component transforming class pixel matrix, and dimension reduction data is obtained by dimension reduction projection matrix projecting the original spectral data. False-color result image is produced by RGB color space transforming from reduction data. Fusion performance of improved method is verified by hyperspectral data of weak small target and underwater fish group. Experiments show that similarity classification fusion method can not only get one false color image from high dimension hyperspectral data, but also avoid missing weak small targets from background, which enhances abilities of detection and identification, and has a strong application prospects.