Hyperspectral Image Prediction Using a Linear Model in Different Illumination Conditions

When using a hyperspectral imaging system to capture images of objects in daylight, it’s difficult to determine the specific sun illumination under which each object in the image was captured. In order to tackle this issue, we have conducted an experiment to demonstrate the feasibility of predicting how the sun’s illumination affects various objects under different lighting conditions. We created 28 hyperspectral images and used a linear machine learning model to predict the lighting conditions in which different objects appear in those images. During the experiment, we were able to confirm a linear relationship between the intensity of illumination on objects and their light reflection. Furthermore, our findings suggest that it is possible to accurately predict the light reflection of an object under different daylight conditions.