Light-Cone Feature Selection in Methane Hyperspectral Images

Hyperspectral images (HSIs) capture detailed spectral information across numerous contiguous bands, enabling the extraction of intrinsic characteristics of scanned objects and areas. This study focuses on the application of light-cone feature selection in quantum machine learning for methane detection and localization using HSIs. The proposed method leverages quantum methods to enhance feature selection and classification accuracy. The dataset used includes HSIs collected by the AVIRIS-NG instrument captured in geographically diverse locations. In this study, we investigate the performance of support vector machine classifiers with different classic and quantum kernels. The results indicate that the quantum kernel classifier, combined with light-cone feature selection, provides in one metric, superior performance when compared to the classic techniques. It demonstrates the potential of quantum machine learning in improving the remote sensing data analysis for environmental monitoring.