Despite the fact that they have more coral species than the Big Barrier Reef, Darwin's waters are increasingly threatened by established species, overfishing and climate change.
In an effort to better manage and monitor fish populations in Darwin Harbor's dark, crocodile-infested waters, the government of the Northern Territory automated this process by making shots from buoy mounted cameras into the platform of artificial intelligence.
Researchers from the Department of Primary Industry and Resources (DPIR) turned to GitHub and Microsoft Azure Cognitive Services to develop open source AI solutions and spent tens of hours identifying the species of fish that crawled into the camera platform.
The first iteration of the system began within a month, and during the next six months of development and deployment, its identification skills gradually increased until AI can now identify fish with an accuracy of 95 to 99 percent.
The speed of fish identification has also been accelerated until AI can analyze video lessons in a matter of minutes and release DPIR scientists for more valuable work with ecosystems.
Fisheries scientist DPIR Dr Shane Penny said that the system allows them to monitor species that are at risk of overfishing such as golden blades and blacks and better adapt conservation efforts.
"These are two commercial and recreationally important species on the northern territory, but research has shown they were overcrowded around a larger area of Darwin," Penny said.
Chief Executive Officer of DPIR Rowan Dollar already wants to extend the program on the basis of its early success, including watching wild fish in freshwater systems and commercial ship equipment by cameras.
"We could look at building a camera on a trawler that is at sea and make catch identification, so we can begin to measure the by-catch. We can begin to identify in real time to better manage these fish," said Dollar.
Other fishing around the world could accept the system that was sent to GitHub to manage fisheries resources in different environments.
Only in the northern territory produces more than half a billion dollars annually in the production of primary industries, including fishing.
The demand for NT marine resources is expected to grow with global trends where fish consumption is rising by 3.2 percent per year according to the UN report of 2018.
Other state environmental agencies are also working to address machine visions for managing sensitive ecosystems, including the Queensland Ministry of Agriculture and Fisheries, which monitors invasive plant species in conjunction with QUT.