TrueColorHSI: Perceptually Faithful Visualization of Hyperspectral Imagery

Principal Investigator(s)

David Messinger

Research Team Members

Fei Zhang

Project Description

TrueColorHSI is an open-source Python toolkit for converting hyperspectral imagery into natural-color visualizations. Rather than representing the visible spectrum using only a few selected bands, it applies colorimetric principles across the full visible wavelength range to produce more perceptually faithful and interpretable RGB images. The toolkit supports multiple standard lighting conditions and basic image-enhancement options, making hyperspectral data easier to inspect, communicate, and use across research applications. A functional version has been developed, packaged, and publicly released through GitHub and PyPI, and is currently available for use by the broader hyperspectral imaging community at https://hsi.yale.edu/.

Figures and Images

True Color Visualizations

True color visualization using different standard illuminants (D50, D65, D75). Adjusting the chosen illuminant allows for tuning the color temperature.