Diseño de un sistema integral de monitorización de la actividad pesquera

  1. Galotto Tébar, María del Mar
Supervised by:
  1. Juan Carlos Gutiérrez Estrada Director
  2. Ivone A. Czerwinski Director

Defence university: Universidad de Huelva

Fecha de defensa: 01 July 2022


Type: Thesis


Fisheries management is generally based on the regulation of fishing effort, limiting fishing capacity and fishing activity. Knowing the activity of fishing vessels accurately and in real time represents a leap in quality in the management of fishing activity. Fishing capacity can be objectively quantified, however, the calculation of fishing activity requires knowledge of the effective fishing time, for which the monitoring of vessel activity is essential. Current vessel monitoring systems do not provide sufficient information to accurately determine fishing activity, they are so expensive that they leave almost the entire artisanal fleet out of control, they only provide the information recorded by the GPS installed on the vessel, the frequency of sending information is very low and they do not send information between samples. This thesis presents the development of a new integrated system for monitoring fishing activity that can complement and overcome the limitations of current monitoring systems. In order to assess fishing activity more accurately, it is proposed to incorporate new sensors on vessels to provide additional information. The proposed system is developed on the basis of a low-cost mobile device with GPS, accelerometer, gyroscope and magnetic field sensors; it also has processing capacity that allows the incorporation of artificial intelligence algorithms to identify in real time when the vessel is fishing. This work evaluates the capacity of sensors integrated in current low-cost mobile devices to detect significant changes in the static and dynamic behaviour of the vessel during trawling activity. The ability of different statistical and heuristic methods to classify and identify the phase of the haul in which the vessel is during its fishing activity is analysed. It also identifies the combination of sensors and the set structure that provides the best response in the classification task of the learning machines. The results obtained indicate that, in general, the heuristic techniques have a high degree of discrimination of each of the phases of the fishing operation and that, in particular, the multilayer perceptron (MLP) is capable of correctly identifying 96.3% of the trawl phase samples using only the GPS and gyroscope sensors.