Research on Remote Sensing Image Scene Classification Method Based on Transfer Learning and Multi-scale Fusion
With the improvement of computer computing power and the development of Deep Learning technology,Deep Learning methods that do not require human intervention have become the mainstream method for remote sensing image classification.Therefore,this paper proposes a neural network based on Deep Learning,embedding Attention Mechanism and blending multi-scale features for scene classification of remote sensing images.This model uses Transfer Learning to reduce the negative impact from insufficient training samples.It embeds Attention Mechanisms and blends multi-scale features in the network to improve the ability to classify small-sized terrain targets,and verifying the effectiveness of the model.Through experimental analysis,it is concluded that the proposed model is feasible and effective for remote sensing image scene classification.