Block segregation of coral mosaic via gray level co-occurrence matrix

Abstract

Current coral survey methods are tedious and time consuming. In this study, we propose a survey method incorporating the use of video transect and block segmentation by Gray Level Co-occurrence-matrix (GLCM). Reef mosaic images will be partitioned into blocks and from each block GLCM are calculated. Blocks are then categorized as either coral or sand based from the texture values of contrast, energy and homogeneity calculated from the GLCM. As such, this paper shows the potential of GLCM for binary segregation of sand and coral, with homogeneity incurring an accuracy of 79% under direct application of a threshold.