Someday, it’s possible that AI will handle all monitoring and fisheries evaluation tasks, but for now these are still human endeavors for the most part. At Big Fin Scientific, we’re certainly evaluating AI to help with several aspects of the required workload but despite all the hyperbole and wish-lists (at least for some), truly dependable, robust AI solutions are a long way off. For those of the readership who feel humans present a near insurmountable advantage over bots for the foreseeable, BFS sends you a solid “huzzah”! Humans aren’t going to be replaced anytime soon.
The key factor to almost every compromise struck between government oversight and fisheries corporations lies in the cost of the requirements imposed by the oversight body. The ferocity and duration of the battles, along with the number of combatants, is generally in direct proportion to the perceived cost. Relief from the cost pressure might come from educating consumers as to the value that fisheries monitoring work brings to the fish on their plate but the reality is that, for the overwhelming majority of the world population, cost at the counter is by far the most important factor in sales and thus the focus for every fisheries corporation. Perceived quality is certainly high on the list, too; accountability and sustainability rate very low for all but a small percentage of consumers.
The takeaway is that we need to minimize the costs of reliable fisheries stock assessments and monitoring. By decreasing assessment costs, we don’t just decrease the number and intensity of oversight battles: we also create new opportunities for research organizations to pursue new studies, to do new science. AI cannot currently come close to delivering the depth, accuracy and reliability required for scientific studies. This is human work, done with boots on the deck and waders in the water.
The Big Fin Scientific system is made for humans; it minimizes the work, costs and delays commonly associated with procuring large amounts of field data. The system is comprised of a hardware component, a length and data-entry fish-board and a software application, which is adaptable to whatever data is required. Everyone’s needs are different, so the software is easy to change, on the fly, by the user. And now, because of our 3d-printing ecosystem, we’re able to provide varying mechanical configurations, as well.
Automated Measurements with AI
There have been some exciting innovations recently related to optical measurements of fish using cameras. This is an obvious advancement of technology, similar to what’s been undertaken for construction and building operations. Of course, fish data is much more complicated than simple measurements: location data, weights, sexes, size codes and real-time feedback remain critical.
Optical measurements are currently inadequate for scientific research and monitoring because of their inherent inaccuracy and repeatability issues. Variability of lighting of the subject remains a problem. Measurements are only good to at best about a 1⁄4 inch or about 3mm and not repeatable even at that resolution and species recognition remains low.
That said, as with any tech, things can progress rapidly and we’re hopeful we’ll be able to use features afforded by this type of technology soon in our field-data application. We’re keeping a close eye on when this tech might be viable for more accurate and reliable work.
EM Innovations using 3D-Printing
Big Fin Sci has been around for almost 15 years now, so we’ve seen a lot of innovation and we’re anticipating things to move even more rapidly in the future. Nowadays, we 3d-print the boards, which lets us innovate quickly and adapt to a variety of needs, and we’re looking at how AI might help design new things even faster.
We can now supply trough sections and custom profiles, which were previously impractical due to volume considerations. The 3d-printed approach allows for high modularity, leading to a custom approach and inexpensive repairs.
Over the years, we’ve been able to decrease the cost of electronic fish boards from $15k to about $3.5k and we feel that AI might help us create new designs that decrease costs even further. Hopefully, that gets the tech to a feasible cost for a high percentage outfits that can use them. We’re looking forward to exploring how 3d-printed innovations will continue to make equipment like ours increasingly useful, adaptable and affordable. Until the ‘bots take over, we all certainly have some amazing tools to minimize the field-work drudgery and costs for us humans!
Chris Carroll is Managing Director at Big Fin Scientific. He welcomes your comments and questions via email.
Industry Focus is an ongoing series featuring fishtech companies and the proojects they’re engaged in. To learn more or discuss coverage of your company, please drop us a line at info@em4.fish.
