Creating and sharing knowledge for telecommunications

Underwater video analysis for species abundance control: the crustacean Norway lobster – a case study

Correia, P.L. ; Yee, L. ; Fonseca, P.F. ; Campos, A.C.

Underwater video analysis for species abundance control: the crustacean Norway lobster – a case study, Proc International Workshop on Marine Technology - Martech, Lisbon, Portugal, Vol. ., pp. . - ., October, 2014.

Digital Object Identifier: 0

Image and video analysis techniques currently play a relevant role in marine science, namely for habitat mapping or for the estimation of commercial species abundance. Most often, it is still done the manual counting from photographs or video footages obtained from a wide array of cameras placed on underwater vehicles, on fishing gear or carried by scuba-divers. The authors have proposed a methodology based on video analysis tools to perform the automatic estimation of bottom species abundance by counting of the single specimens. Additionally, given the specificity of the species constituting the case-study (the Norway lobster, Nephrops norvegicus, is a burrowing crustacean), the methodology had to be extended to allow for counting burrows, a more reliable proxy of abundance. For such studies the experimental set-up typically consists of a video camera mounted on a sledge, towed from a research vessel during targeted surveys. Alternatively, as was the case, a video system can be mounted on a survey trawl headline. To test the methodology, a set of video sequences were captured at the Portuguese continental slope, at high depths, using a monochrome camera. The corresponding video analysis had to deal with challenges raised by non-uniform artificial illumination, artifacts at image border caused by the container used to withstand the high water pressure, noise and marine snow. The developed software prototype is able to provide objective and comprehensive information, to complement a manual evaluation of lobster and burrow density estimation.