Multi-source remote sensing images such as MSS, TM, ETM, OLI, and Sentinal-2.

Using multi-source and multi-temporal remote sensing images, using the zoning and hierarchical decision tree method, first of all, object-oriented multi-scale image segmentation is completed based on e-Cognition remote sensing interpretation software. By superimposing and analyzing single element maps such as altitude, precipitation, temperature, soil, vegetation, and landform, the secondary interpretation results of large-scale geological zoning plots are obtained. On the basis of large-scale segmentation units with geological characteristics, we continue to use The spectral characteristics of the image are used as the basis for classification and evaluation, and the interpretation results of the desert are obtained in turn.

This data is based on large-scale segmentation units with geological characteristics, and continues to use the image spectral characteristics as the basis for classification and evaluation, and then stratified to obtain the interpretation results of the desert. The interpretation accuracy is better than 85%.