Segmentación iterativa basada en conocimiento del afloramiento de aguas frías en la costa Sahariana

This paper presents a technique for the automated detection of cool upwelling in saharian coast. Classic segmentation methods based on thermal gradients give poor results due to a high variability of the image pixels. The use of isothermal lines yields a better segmentation for this problem. An iterative, knowledge based technique is proposed for the threshold selection. This technique works with morphologic and statistical features extracted from the obtained regions of the image.