Estimation of fishing net underwater geometry with embedded deformation graphs

An innovative proposition to model fishing nets is presented in this paper. The novelty stems from the use of embedded deformation theory to infer the geometry of fishing gear as a graphical problem. The technique developed is proven suitable to solve the task in a practical fishing environment such as purse-seine fisheries, as it considers real-time observations from sparse 6D positioning sensors, which can be easily embedded prior in pre-set pockets in the fishing gear. The effectiveness is validated by simulated experiments of fishing gear deformation scenarios with 12 measurement sensors. Real deployments are also being conducted at the time of writing on a fishing vessel with the equivalent instrumented net and will be readily available for presentation at the time of the conference.