Metodología de desarrollo de sistemas de detección y seguimiento de peces mediante tecnología láser y visión artificial con Inteligencia Artificial

Abstract

This PhD thesis presents a methodology to underwater fish detection, tracking and measuring, using remote sensing techniques. In most cases the applications where fish detection and size measurement is required, it is essential that they are in their hábitat, swimming freely and without any electrical or mechanical device attached to the fish itself. In this way, their behavior is not conditioned or they are subjected to stress that could cause unwanted situations. Remote sensing techniques based on visible light and infrared light allow the fish to be detected without physical contact however, it is necessary to adapt them for their correct operation under water. In this thesis, this methodology using computer vision and three practical applications with detection devices base don different physical principles are presented. In addition, this work studies the feasibility of using devices, initially designed to work in outside the water enviroments, which can be adapted by means of computer algorithms in the tracking of underwater objects, and mainly in the detection and measurement of fish in their environment.