VisNed and several fishing vessels are developing a digital tool to implement a fully documented fishery (FDF)
Dutch researchers and fishers are developing a digital tool to implement fully documented fisheries (FDF), with support from the EU.
When you go out fishing, you can never be sure what you’ll catch. It could be anything – from a prize specimen to an old shoe. Imagine, then, the scale of the problem if you’re using fishing nets and end up bringing hundreds of kilograms of catch on board?
Unwanted catches have long been a problem in fishery, because not everything that is caught can be sold or is fit for human consumption. Catching fish and then throwing it back into the sea is both harmful and wasteful. To combat this problem, the European Union introduced the landing obligation – conservation measures that require all catches to be registered and brought to shore (landed).
But this requires time and space and can be very costly. A Dutch fishers’ organisation VisNed calculates that, on average, a vessel would need around four more people on board to process and land the unwanted catches. However, there is no room for more people on most Dutch fishing vessels and a larger crew would mean lower wages for each crew member.
Technology to the rescue
This is the point at which science and fisheries intersect to develop innovative solutions. In the Netherlands, researchers at two institutes of Wageningen University (Wageningen Marine Research and Agro Food Robotics), in cooperation with VisNed and several fishing vessels are developing a digital tool to implement a fully documented fishery (FDF). The project has received just under €3 million in support from the European Maritime and Fisheries Fund (EMFF).
FDF introduces automated recognition of the size and species of each fish, distinguishing between catches fit for human consumption (above minimum size) and unwanted catches (below minimum size). Using remote electronic monitoring (REM) systems, the tool can determine the weight of the total catch.
Developed by Agro Food Robotics, the process uses complex algorithms, artificial intelligence (AI) and spectral learning and vision technologies to recognise fish species and classify the fish by size (above or below the allowed minimum conservation reference size). The same technology is already in use for measuring and assessing the quality of agricultural products, such as broccoli and tomatoes.
To test the process in a life-like environment, researches at the Den Helder fish market first constructed an exact copy of a fish sorting line used on board fishing vessels and equipped it with two high-definition cameras: one to recognise the species and one for 3D visualisation and determination of volume.
After a successful trial run, the prototype was then tested on board a fishing vessel for the effects of vessel movement and salt-water spray on the camera lenses. The second trial with more vessels at sea was launched in 2021.
The preliminary results show that the process can be applied without major changes to the way fish is normally handled on board. It currently recognises four species – plaice, sole, turbot and brill – with speed and accuracy. The technology calculates the weight of each fish by determining its volume and can also distinguish if a fish is above or below the minimum size.
An earlier project (Innoray) has demonstrated that the same technology can be used to distinguish between the three main commercial species of rays, even when they are lying upside down, which is very difficult for the human eye.
The project also meets privacy and data protection (GDPR) requirements, as the system can process data without human interaction. In fact, the footage can be destroyed as soon as the computer has counted the fish and the data belongs to the fishers themselves.
Promising perspectives for future use
There is a lot of potential for this technology. For example, the cameras could be linked to electronic catch reporting systems (e-logbook) and reduce the administrative burden on the skipper. It also allows real-time monitoring of fishery, which could improve data collection and transparency.
Because it collects data on catch quantity and composition of both target species and unwanted catches, this technology could also become an important source for scientists evaluating the state of fish stocks, boosting the efforts for the sustainable management of marine resources.
Thanks to this project, the Dutch flatfish fishery received an exemption from the landing obligation until the end of 2021. While the project is yet to be fully evaluated, the FDF technology could become a mainstay of EU and global fisheries, all of whom are facing the same challenges of registering catches by species and size.