I’m colorblind, and that has gotten in the way when trying when doing photography - from trying to figure out which camera color preset is the best to sucking at raw file editing. The problem is that most of filters suck, actually (they probably were not even made by colorblind people in the first place!). So, my idea is to make a filter that adapts to your level of colorblindness focusing on fidelity to real life (so it is actually useful for photography).
FLOWS: Flood Likelihood from pre-compiled Watershed Structure. Make flood-risk assessment with low-end hardware! FLOWS finds bowl-like depressions and analyses them. Why is that good: We now know were water accumulates, and for a fixed amount of rainfall max water height in theory should stay about the same. This means we can run a one time computation that analyses different scenarios of rainfall and soil humidity and store the values so we can check them later anytime!