Vegetation Detection and Feature Extraction in Remote Sensing Images Based on Neural Networks

In this paper, we propose a method based on Transformer architecture to combine generative adjunctive network (GAN) with new model for vegetation detection and feature extraction in remote sensing images. First, the powerful self-attention mechanism of Transformer can effectively capture spatial and spectral features in remote sensing images, thus enhancing the ability to identify vegetation in complex backgrounds. The generative adversarial network improves the diversity of the training set by generating photorealistic composite images, thus improving the robustness and generalization ability of the model. The new model proposed in this paper involves multi-level feature fusion and context information integration in the feature extraction process, which effectively enhances the detection and classification accuracy of different types of vegetation. Experiments on several standard remote sensing data sets show that the proposed model is significantly superior to traditional methods, and has achieved excellent performance in both accuracy and efficiency of vegetation extraction.