Hyperspectral cameras face challenging spatial-spectral resolution trade-offsand are more affected by shot noise than RGB photos taken over the same totalexposure time. Here, we present a colorization algorithm to reconstructhyperspectral images from a grayscale guide image and spatially sparse spectralclues. We demonstrate that our algorithm generalizes to varying spectraldimensions for hyperspectral images, and show that colorizing in a low-rankspace reduces compute time and the impact of shot noise. To enhance robustness,we incorporate guided sampling, edge-aware filtering, and dimensionalityestimation techniques. Our method surpasses previous algorithms in variousperformance metrics, including SSIM, PSNR, GFC, and EMD, which we analyze asmetrics for characterizing hyperspectral image quality. Collectively, thesefindings provide a promising avenue for overcoming the time-space-wavelengthresolution trade-off by reconstructing a dense hyperspectral image from samplesobtained by whisk or push broom scanners, as well as hybrid spatial-spectralcomputational imaging systems.