Automatic segmentation of coral reefs implementing textures analysis and color features with gaussian mixtures models

摘要

The applications of digital image processing techniques over coral reefs are emerging research field due to coral reefs importance in marine ecosystems. One important step to make a correct analysis of these reefs is segmentation stage. Few automatic segmentation techniques in corals have been used to solve this problem, which is why in this paper we propose a texture segmentation algorithm with applications to coral reefs segmentation. Our goal is reached using a combination of texture analysis, color features and Gaussian Mixture Model, varying the parameters for each one. The results show that automatic segmentation of corals using texture and color is a good approach, in most cases the corals are correctly segmented, even in a real environment the images are successfully segmented.