Leveraging band diversity for feature selection in EO data
Sadia Hussain, Brejesh Lall
TL;DR
This paper tackles high-dimensional hyperspectral imaging (HSI) data, where hundreds of narrow bands create high computational costs and limited training samples. It introduces a unified band grouping framework that combines (i) diversity-driven band selection via Determinantal Point Processes on a correlation-informed representation, (ii) spectral correlation-based grouping, and (iii) Spectral Angle Mapper (SAM) to disambiguate overlapping bands within a group, formalized with $L$, $P^{k}_{L}(Y)$, and SAM criteria. The approach enables accurate and efficient reconstruction of high-resolution HSI by operating on a reduced, non-redundant band set, e.g., reconstructing $Z$ from a reduced basis $B \in \mathbb{R}^{N \times n}$ and coefficients $M \in \mathbb{R}^{n \times WH}$. If successful, this modular framework could improve restoration and analysis workflows in earth observation by preserving spectral diversity while mitigating redundancy.
Abstract
Hyperspectral imaging (HSI) is a powerful earth observation technology that captures and processes information across a wide spectrum of wavelengths. Hyperspectral imaging provides comprehensive and detailed spectral data that is invaluable for a wide range of reconstruction problems. However due to complexity in analysis it often becomes difficult to handle this data. To address the challenge of handling large number of bands in reconstructing high quality HSI, we propose to form groups of bands. In this position paper we propose a method of selecting diverse bands using determinantal point processes in correlated bands. To address the issue of overlapping bands that may arise from grouping, we use spectral angle mapper analysis. This analysis can be fed to any Machine learning model to enable detailed analysis and monitoring with high precision and accuracy.