Ideal color space component for reconstruction of contamination plumes

There are several techniques to reconstruct a plume of nonaqueous phase liquid (NAPL) contamination in bench-scale geoenvironmental studies. Using image processing methods, the most crucial issues are to determine the optimum concentration and the best color of dye as a tracer for use in mapping a zone of contamination. This issue becomes more complicated when the original color of contaminant cannot be altered such as of crude oil. The objective of this study is to find a color space in which the relationship between transmitted signal and integrated concentration is quantifiable. In particular, the goal was to correlate the spatial concentration of contamination with color pixel information. For this purpose, a new algorithm was used to identify the best concentration for a number of dyes that can be used as tracers. Additionally, the ideal color space component for reconstruction of each dye was determined. The effectiveness of this color classification method was assessed using 10,368 color space component images within the framework of the peak signal to noise ratio for eight different dyes and six color spaces spanning a concentrations ranging from 1 to 2000 ppm, for 8 NAPL zone lengths. The effect of data filtering was also considered and a 15 × 15 pixel convolution average filter is recommended for image conditioning.