Hyperspectral reconstruction image classification based on UPDM

For the restriction of the air, sensor and kinds of natural conditions, satellites can´t get enough remote sensing data to meet the needs of scientific research, especially for the western mountainous area. This study reconstructed the TM images of Guizhou mountainous areas to achieve their hyperspectral images based on UPDM analysis, and classified the hyperspectral images for vegetation information extraction. The results of the study show that this method can expand the depth and width of multi-spectral remote sensing data when it applied in Landsat5.