Discrimination of snow/ice and cloud against various ecological region using DWT and color models

The cloud detection algorithm is important to raise the accuracy of parameters which is used in the research field related to the energy budget of earth and much more. Thus algorithm should be strong enough so that it can differentiate clouds against various surfaces (i.e. Ocean, Vegetation, Soil, and Snow). This paper covers the description of discrete wavelet transform (DWT) and HSV color model and results based on the same is realized over NOAA, VIRR and MODIS datasets in visible band. After realizing the merits and demerits of proposed algorithm over around 920 images, it is found that proposed algorithm gives better results approximately 90%, compare to existing techniques in differentiating cloudy pixels and snow/ice contaminated pixels of satellites images.