Research on land-use automatical classification based on HJ -1 satellite data

Based on the calibrated HJ-1 satellite images, Dark pixel method was used to do the atmospheric correction, then applying the pre-processing of the CCD data of HJ-1 satellite to the quantitative inversion model. Acquired four kinds of surface features index: Normalized Difference Vegetation Index, Land Surface Temperature, Vegetation Coverage Index, Normalized Difference Water Index, based on these indicators, extracted the six feature types and achieved the automatic classification of land-use.