HabCam Image Informatics: Scallop and Substrate Classification

Scott Gallager Woods Hole Oceanographic Institution

Abstract

The use of underwater stereo imagery in stock assessment and ecosystem analysis is becoming wide spread as an alternative to pelagic and benthic trawling. We have been collaborating with NOAA for the past six years in developing new, non-invasive technologies for cost effectively surveying sea scallops and ground fish. As such we have developed a stereo optical and acoustic imaging system that is currently being used by NOAA to survey scallops along the entire northeast continental shelf.

This proposal is in support of two WHOI/NOAA initiatives that directly relate to either further development of this technology or to its application in enhancing stock assessment of sea scallops and ground fish and assessing Essential Fish Habitat.

Project Tasks: 1) Continue development and integration of automated substrate classification into the habcam real-time workflow (S Lerner and S Gallager). 2) Support for development and testing of image processing and machine vision tools for segmentation and classification of yellowtail flounder and other difficult targets (in collaboration with Lakshman Prasad, Los Alamos National Laboratory.