Multi-view Lithology Remote Sensing Scene Classification Based on Transfer Learning

In vegetated areas, the lithology image features are complex and the distribution of lithology species in different regions is different, which makes it difficult to accurately classify lithology across regions. Aiming at the problem that it is difficult to identify new lithology in the cross-region prediction of conventional models, this chapter uses the idea of transfer learning to study the model migration ability on the basis of the lithology scene classification model based on multi-view data fusion, and proposes a transfer learning method based on multi-view data fusion, which can achieve the identification of new lithology types across regions and improve the model generalization ability.