Line of action:

Status:

Execution date:

The overall objective of the AutoNatura2000tres project is the implementation and automation of a coupled system for monitoring fishing pressure and fishery resources based on the acquisition of high-definition images and their analysis using artificial intelligence tools (Deep learning) in a marine Natura 2000 Network space.

The specific objectives are:
-Development and automation of the autonomous system for the acquisition of high-resolution images for fisheries monitoring developed in AutoNatura2000 (Pleamar 2017 Call) in a space of the marine Natura 2000 Network (the marine reserve of the bay of Palma);
– Development and automation of the autonomous system for the acquisition of high-resolution images for fishery resources developed in AutoNatura2000dos (Pleamar 2018 Call) in a space of the marine Natura 2000 Network (the marine reserve of the bay of Palma);
– Development of two mathematical algorithms based on artificial intelligence (Deep learning): extension of the algorithm developed in AutoNatura2000 for the automatic identification and tracking of fishing vessels using deep learning techniques and extension of the algorithm developed in AutoNatura2000dos for the identification
automatic and tracking of fish using deep learning techniques;
– Coupling and validation of the new monitoring of fisheries and fishery resources: error in the identification of fishing vessels and fish, percentage of success, false positives and accuracy of monitoring;
– Transfer of the knowledge generated in AutoNatura2000three: generation of useful output for the administrations of the Natura 2000 network: fishing effort and number of recreational fishermen, maps of resource distribution. Outreach activities.

The following actions will be carried out within the project:

• A1. Installation of the autonomous high-resolution image acquisition system for the automatic monitoring of fishing boats.
• A2. Installation and anchoring of an underwater recording system to assess the state of the resources.
• A3. Development of a mathematical algorithm for the automatic detection and monitoring of fishing boats based on artificial intelligence (Deep learning).
• A4. Development of a mathematical algorithm for the automatic detection of fish species based on artificial intelligence (Deep learning).
• A5. Coupling and validation of fisheries and fishery resources monitoring systems.
• A6. Generation of AutoNatura2000three Output to cover FEMP priorities, the application of the CFP and the PMI.
• A7. Transfer of AutoNatura2000three results and awareness-raising activities.

The marine Natura 2000 Network is a network of protected areas that aims to ensure the long-term survival of marine biodiversity, making this objective compatible with the development of human activity in the coastal zone. To this end, it is necessary to follow up and monitor both biodiversity and human activities carried out in these areas, to determine the state of their conservation. IMEDEA (CSIC-UIB) has completed the AutoNatura2000tres project, developing specific artificial intelligence tools for monitoring fishing pressure and conservation status of fish stocks in the protected area of the Bay of Palma.

Within the framework of AutoNtaura2000tres, a system based on a computer and a camera has been developed that is capable of counting and locating fishing vessels and fish species on its own. Thanks to this computer and the implementation of artificial intelligence algorithms, information on the distribution of fishing effort and the state of the reproductive biomass of fishery resources is obtained, in an autonomous and unsupervised way.

The project has been developed in two large blocks. On the one hand, a study of fishing boats has been carried out. The project has developed an automatic monitoring system for boats thanks to a high-definition image acquisition system and the development of algorithms that allow the computer to recognize the type of vessel (fishing, sailboat) and its location. This system, based on artificial intelligence, has allowed continuous monitoring of boat traffic in the protected area, as well as the development of a zoning of the recreational activities that take place in the bay of Palma. Among others, the system has monitored in high resolution the peak of fishing pressure caused by the opening of the ban on raor (Xyrichtys novacula), information that helps to know and analyse the status of its exploitation.

On the other hand, the project has allowed the development of tools that facilitate the acquisition of information on the conservation status of fish stocks vulnerable to fishing and invasive algae. To this end, a series of underwater cameras have been anchored in the marine reserve of the bay of Palma. The images are automatically classified by a system previously trained with images of fish of different species, which allows the abundance of fish to be automatically quantified. The system has been able to quantify the state of the fish populations that inhabit the sandy bottoms of the marine reserve. As a result of this monitoring, a decrease in the population of raor has been detected, which could be caused by a decrease in the presence of an invasive algae of tropical origin that colonized a large part of the protected area. Over the next few years, the automatic monitoring system will allow studying whether this trend is confirmed, but fishermen have already detected a notable decrease in their catches in the 2021 raor fishing season.

During the development of the AutoNatura2000three project, the great potential of the automatic monitoring system to cover EMFF priorities has been demonstrated, among others, to improve the implementation of the CFP and the PMI. The pilot project of the AutoNatura2000three system is presented as a very interesting alternative for the collection of data on the distribution of vessels and fish species exploited by recreational fishing, as well as for the monitoring of these areas, therefore, which has demonstrated the potential for the achievement of several of the objectives of the EMFF.

The AutoNatura2000 monitoring system allows:

• Obtaining spatial and temporal data on the recreational effort carried out in the Natura 2000 Network similar to those obtained by commercial fleet monitoring systems such as blue boxes or AIS and with a high level of detail,
• It allows automatic monitoring and surveillance of the Natura 2000 Network by a low-cost method, which offers a multitude of opportunities.
• It allows automatic monitoring of the distribution of fish exploited by recreational fishing, both temporally and spatially.

The pilot project of the AutoNatura2000three system is presented as a very interesting alternative for the collection of data on fishing and recreational activities, as well as for the monitoring of these areas, therefore, it has demonstrated the potential for the achievement of several of the objectives of the EMFF. The following points should be highlighted within the general assessment of the project:

1. AutoNatura2000tres is capable of characterizing the coastal fishing effort, following the movements of recreational fishing boats and identifying their activity autonomously, and identifying fish species of fishery resources.
2. The system improves the optimal design of MPAs and Natura 2000 Network, and covers multiple EMFF priorities, including improving the implementation of the Common Fisheries Policy (CFP) and enabling integrated management of the spatial coastal zone (PMI) – coupled with catches, which can contribute significantly to the improvement of fisheries management.