The Global Wetlands Project team won a Microsoft AI for Earth Grant. The grant was awarded to continue the development of our smart remote monitoring tool FishID.
FishID is an artificial intelligence platform that can automatically detect and classify fish species from underwater footage.
We have been developing a FishID model for the Moreton Bay Environmental Education Centre (MBEEC). FishID will automatically analyse all their underwater videos collected through multiple years and multiple locations in Moreton Bay, Queensland, Australia.
We have now processed several underwater videos collected by MBEEC with a network of fast and robust virtual machines available at FishID.
With more than 31,000 manual annotations of fish, the current FishID MBEEC model can detect and count (MaxN) 8 target species across a variety of environmental and spatial conditions. The 8 target species are a variety of common fish species with commercial and/or ecological value in Moreton Bay.
Here are some examples of FishID detecting some of the target species
Australasian Snapper detected in low visibility conditions
Detections of Blue Catfish and Smooth Golden Toadfish
If you would like to follow our progress, follow me at @seabassphd or check future updates in the blog.
Sebastian Lopez (Seabass), is a PhD Candidate at the Australian Rivers Institute where he is developing and applying artificial intelligence tools to monitor fish populations in marine ecosystems.
Comments