Hyperspectral Band Clustering for Visualisation.

Hyperspectral imaging captures images in narrow wavelength bands, providing rich and detailed information for various applications. However, accessing and interpreting this information is challenging due to the large size of hyperspectral images and the need for ground truth data. This article presents BC4V, a novel framework for visualising hyperspectral images through band clustering. BC4V uses salience detection to produce a composite image with three channels that convey rich visual information. It preserves physical meaning and facilitates interpretation, striking a balance between information preservation and interpretability. The method is evaluated using objective and subjective criteria, demonstrating its effectiveness in facilitating anomaly detection and class recognition.