Feature selection-oriented interactive visual analysis approach for hyperspectral images

Hyperspectral images (HSIs) have become increasingly prominent as they can maintain the subtle spectral differences of the imaged objects. Designing approaches and tools for analyzing HSIs presents a unique set of challenges due to their high-dimensional characteristics. Given the problems existing in the current visual analysis methods of HSIs, such as insufficient guidance and difficulty in achieving an accurate selection of specific spectral pixels, a universal interactive visual analysis approach is proposed in this article, which enables observers to visually interpret the rich information contained in HSIs with guidance, and pertinence through a graphical interface. The selection of the region of interest can realize the interactive screening from the spatial dimension of HSIs. Three information indicators are used to guide observers to select bands effectively. The clustering calculation and its scatter plot play an important guiding role in the selection and interpretation of feature classes for observers. Aiming at the precise selection of specific spectral pixels, a parallel coordinate method with reordering calculation of spectral bands is proposed to make it easier to distinguish spectral data curves and improve the clarity of target class expression. Finally, the usability and effectiveness of the proposed approach are analyzed through experiments.