SimonTopp
commented
6 years ago After reviewing papers, we've realized there is no simple threshold approach to filtering out sun glint. Most approaches involve some sort of radiative transfer function that's beyond the scope of the data pull. Instead we're pulling down standard deviation of band reflectances in order to filter by heterogenous water surface.
We have noticed an issue with pulling down data that is clearly a product of wind chatter or sun glint where reflectance values are way above a reasonable value for over water. This is especially obvious in values with high SWIR. We need to read papers like this one and update our reflectance pull to automatically mask out the worst sun glint.