I. Introduction
In recent years, owing to the rapid advancement of remote sensing technology, hyperspectral image (HSI) has emerged as a prominent tool for extracting valuable information from the Earth’s surface [1]. Compared to other remote sensing data, HSI has a remarkable ability to subtle spectral variations among pixels, enabling precise differentiation of image components beyond spatial limitations. However, the complex spectral nature of HSI often poses challenges in fully harnessing its inherent potential [2].