Thursday , October 28 2021

Sharks and crocs port management system AI – Software


In spite of having more coral species than the Great Barrier Reef, Darwin's waters are under increasing threat of introduced species, excessive fishing and climate change.

In an attempt to control and control fish populations in the turbulent and unfounded waters of Darwin Harbor crocodiles, the Northern Territory government automated the process by feeding images of buoy cameras on an artificial intelligence platform.

Scientists from the Department of Industry and Primary Resources (DPIR) turned to the cognitive services of GitHub and Microsoft Azure to develop the open source AI solution, spending dozens of hours identifying fish species that were introduced in the vision of the cameras to train the program of Learning machines behind the platform.

The first iteration of the system began to work in a month, and during the next six months of development and deployment, its identification capabilities have been progressively improved to the point where AI can now identify a fish with 95 to 99 percent accuracy .

The speed of fish identification has also accelerated to the point where AI can analyze video hours in minutes, releasing DPIR scientists for a more valuable ecosystem management work.

DPIR's fishing industry scientist Dr Shane Penny said the system is allowing them to keep an eye on species at risk of excess fishing, such as goldfish and black fish, and improving conservation efforts.

"These are two species of commercial and recreational interest in the Northern Territory, but research has shown that they were overweight around the larger Darwin area," said Penny.

The DPIR information chief, Rowan Dollar, is keen to expand the program based on its initial success, including the supervision of wild fish in freshwater systems and commercial boat equipment with cameras.

"We could look for the creation of a camera in a trawler that is offshore and that identifies the capture by flight, so that we can begin to measure the accidental capture. We can begin to be able to identify this in real time, to help better manage these fisheries, "Dollar said.

Other fisheries around the world could also adopt the system that was uploaded to GitHub to manage fisheries in different environments.

Only in the Northern Territory, the value produced by the primary industries, including fishing, reaches more than half a billion dollars per year.

The demand for NT marine resources is expected to grow with global trends, where fish consumption has increased by 3.2 percent per year according to a 2018 United Nations report.

Other state environmental agencies are also looking for mechanical vision solutions to manage sensitive ecosystems, including the Queensland Department of Agriculture and Fisheries, monitoring invasive plant species in a link with QUT.

Source link