An Automated Morphological Image Processing Based Methodology for Quantifying Coral Cover in Deeper-Reef Zones

With the advent of new robotic technologies such as AUVs, a number of end user communities are being inundated with large amounts of data. The traditional techniques of manually counting and sorting out organisms in individual images are just not scaleable to the large datasets that are now being acquired. This paper examines the use of morphological image operators in the automated analysis of imagery for studies associated with coral reef ecology. We propose a texture-based algorithm to segment out areas of coral cover in these images. Results show percent cover values competitive with the existing human methods