Unmanned Systems Technology 021 | Robot Aviation FX450 l Imaging Sensors focus l UAVs Insight l Liquid-Piston X-Mini l Riptide l Eurosatory 2018 show report l Zipline l Electric Motors focus l ASTS show report

14 Platform one August/September 2018 | Unmanned Systems Technology Researchers from the University of Haifa in Israel and the IMDEA Networks Institute in Madrid, Spain, have developed an autonomous system to monitor schools of fish in real time (writes Nick Flaherty). The Symbiosis system combines an optical camera with an underwater acoustic sensor to provide reliable information about the condition of marine fish stocks, something that at the moment is practically impossible without investing enormous resources. The system collects data over long periods and transmits it to a coastal centre. The researchers are focusing on identifying six large fish species that are in especially high demand – two species of tuna, scad (a type of mackerel), Atlantic mackerel’, mahi-mahi (also known as the common dolphinfish) and swordfish. The 2 x 3 array of acoustic sensors measure the size of the fish and their total biomass in the area. Once the array identifies one of the six selected species from their movement, it activates the optical system, which has several cameras, and data processing with various image identification algorithms using deep learning. When the optical system confirms the identification of one of the six species, it transmits the information via underwater acoustic comms to a surface station and then by radio comms to a coastal station. “Using acoustics to localise specific fish species is very challenging,” said Paolo Casari at IMDEA. “First, the acoustic processing chain has to incorporate cost-effective components, and it needs to be highly energy-efficient. The signal processing algorithms in the system also have to strike a good trade-off between complexity and accuracy. “On top of this, the underwater environment contains many background acoustic noise sources and reflectors, and the signals from fish around the Symbiosis system will be much weaker than the acoustic interference coming from the environment. The algorithms need to be robust enough to cope with those shortcomings.” Underwater operations must contend with low visibility, and the water distorts the images, so the big challenge is to secure good detection performances and minimise false alarms. That needs to happen autonomously in a deep-sea environment. The optical classification of fish species has its own particular challenges as well. There are very few pre-classified images available with which to train the deep-learning classifier, and the images that are available were taken under very different visibility conditions to those the system will encounter. The researchers are tackling this by using databases of fish pictures from scuba divers and underwater photographers. As part of the project, a prototype is being developed that will include a system of acoustic sensors, a network of cameras, image processing units and an energy unit permitting autonomous activity. The project will run until 2020. Marine vehicles Fishing by the numbers Schematic of the Symbiosis fish stocks monitoring system