Subsea Fauna Enumeration Using Vision-Based Marine Robots

Karim Koreitem, Yogesh A. Girdhar, Walter Cho, Hanumant Singh, Jesus Pineda, Gregory Dudek

Abstract: This paper describes a robotics system for population density estimation of marine organisms and vision-based algorithm for computing the associated population estimates. We focus on benthic fauna, through the use of Seabed AUV to collect benthic imagery, and then employ a support vector machine (SVM) for automated analysis of these images to estimate the population of the fauna of interest. The proposed approach is a significant improvement over existing techniques such as trawling, or manual inspection of images collected by a towed vehicle. We tested our proposed technique by first collecting benthic image data using the Seabed AUV at Hannibal seamount in Panama, and then predicting the counts of the crabs and squat lobsters in the data. We compare our predictions with ground-truth data from thousands of sample locations containing manual counts estimated by a team of experts, and found that our estimates have 94% precision and recall on held out test data.