A WorldView-3 multispectral glint correction methodology and relation to hyperspectral signatures in space coast Florida waters

The type of bottom resting below the water surface may be difficult to be determined due to the contamination of pixels by glint in satellite reflectance imagery. A nearest neighbor non-glint pixel search algorithm was written in Python to correct glint contaminated pixels. An algorithm was then utilized to estimate water visibility and to analyze the results of the glint corrected pixels by using light attenuation depths and attenuation coefficients based upon glint corrected satellite reflectance in the WorldView-3 (WV-3) image channels. The aforementioned correction was undertaken for sand, mud, and seagrass bottom types. The results suggest that it is possible to estimate attenuation depths (a measure of water visibility).