Statistical Identification Method for oil Pollutants using Laser Fluorimetry Data

We describe an algorithm for probabilistic identification of oil pollutants on the sea surface using spectra obtained by lidar scanning. Identification was made by comparing the measured discrete spectrum with a set of reference sample spectra. The result of the comparison is a probability distribution for similarity to the reference samples. We have carried out a model comparison with the familiar identification method based on minimum distances between vectors.