Additionally, since augmented samples may still be insufficient, using overly complex networks can lead to overfitting. Therefore, we introduce a hyperspectral image classification network called ...