I2N2: A software for the classification of benthic habitats characteristics

Underwater video cameras mounted into towed platforms (e.g. sledges) have been increasingly used for the assessment of commercial crustacean stocks and also for more ecology-directed studies, including the impact of human activities in marine habitats. In this study a video camera was mounted on a trawl headline, to acquire footages at about 500 meters depth in Norway lobster (Nephrops norvegicus) fishing grounds, to automatically quantifying the species abundance and density of burrows, and assess the impact of fishing activities in these fishing grounds. Six complementary features are proposed to identify the lobsters and their burrows: average intensity, slant angle, run-length histogram, shape ranking, co-occurrence matrices and cross-counting. A prototype system, IT-IPIMAR Nephrops Norvegicus (I²N²) is presented and experimental results show that the proposed features, when used in combination, are able to effectively classify segmented regions as lobsters or burrows.