FISH_ROCK : a tool for identifying and counting benthic organisms in bottom photographs

Abstract

Recent advances in underwater robotics and imaging technology now enable the rapid acquisition of large datasets of near-bottom high-resolution digital imagery. These images provide the potential for developing a non-invasive technique for fisheries data acquisition that reveals the organisms in their natural habitat and can be used to identify important habitat characteristics. Using these large datasets effectively, however, requires the development of computer-based techniques that increase the efficiency of data analysis. This document describes one such tool, FISH_ROCK, which was developed for a group of fisheries researchers using the SeaBED AUV during a research cruise in October 2005. FISH_ROCK is a graphical user interface (GUI) that is executed within Matlab, and allows users digitally generate a database that includes organism identification, quantity, size and distribution as well as details about their habitat. Further development of this GUI will enable its use in different oceanographic environments including the deep sea, and will include modules that perform data analysis.