IOBSERVER: species recognition via computer vision

This paper is about the design of an automated computer vision system
that is able to recognize the species of fish individuals that are classified into a fishing
vessel and produces a report file with that information. This system is called iObserver
and it is a part of project Life-iSEAS (Life program).A very first version of the
system has been tested at the oceanographic vessel “Miguel Oliver”. At the time of
writing a more advanced prototype is being tested onboard other oceanographic
vessel: “Vizconde de Eza”. We will describe the hardware design and the algorithms
used by the computer vision software.