Application of chemometric preprocessing and achromatic lens in hyperspectral imaging for assessment of uneven wound

Abstract

Hyperspectral imaging (HSI) uses contiguous spectral channels to collect images under different wavelengths. It is known that this spectral information can be used for material differentiation or identification. However, different sample preparation methods, lighting conditions, distance, and sample morphology will cause huge variations in the collected spectra. In this study, we performed clustering analysis on a HSI data of an elbow area, which is roughly a spherical surface rather than a flat surface. Chemometric preprocessing (CMPP) developed in spectroscopy is applied on the HSI data to reduce the morphology-induced spectral variation. A hardware approach, installing an achromatic lens, is also adopted for comparison. Positive results indicated that the with CMPP and certain spectral measures in clustering could be helpful for wound assessment using HSI.