[ad_1]
(Image provided)
The Department of Primary Industries and Resources (DPIR) of the Northern Territory turned to artificial intelligence to accelerate its fish counting exercises.
The DPIR scientists had the habit of completing the exercise by watching hours of underwater footage filmed on reefs around Darwin, but AI is now able to finish the task in a few minutes. minutes. Scientists could not get into the water themselves because crocodiles and saltwater sharks did not fully agree that humans are entering the water.
Developed with Microsoft and built on Azure Cognitive Service, the open source solution is published on GitHub and had its first iteration in a month. It was then widely deployed in six months and, thanks to the tuning and improvement of its machine learning, it is now possible to identify the fish with a success rate of 95 to 99%.
DPIR stated that the solution has already shown that golden snapper and black Jew are overexploited and that the solution could be applied elsewhere in the territory.
"We could consider setting up a camera on a trawler at sea and identifying the catch on the fly, so that we can start measuring by catch, so we can start to identify this catch in real time. "said Rowan Dollar, information officer at DPIR.
Dollar said it was important that the government operate effectively.
"It's important that every government, regardless of its identity and where it's used, uses technology to collect and analyze data," Dollar said.
"This will help you be more efficient and offer better value to your stakeholders."
In Spain, Satlink also uses artificial intelligence to observe fishing activities and fishermen are not satisfied with it.
"We already have AIS and GPS ships on board, we do not need cameras," said Paco, a Catalan fisherman with nearly 30 years of experience, last week.
(Image provided)
Earlier this month, the Australian Commonwealth Scientific and Industrial Research Organization (CSIRO) released A $ 35 million for research on AI and space technologies, including on the development of research and development. Advanced imagery of Earth from satellites and data science via AI and machine learning.
A $ 19 million will be used to target AI-based solutions in areas such as food safety and quality, health and well-being, sustainable energy and resources, resilient and valuable environments, as well as Australian and regional security, explained CSIRO.
Key areas of research include platforms to "improve prediction and understanding of complex data, platforms for reliable inference and risk-based decision-making, and data systems for ethical IA, robust and scalable ".
This investment is part of CSIRO's Future Science Platform (FSP) portfolio, which aims to dedicate research to new and emerging opportunities for Australia.
Related coverage
LG creates new divisions of robotics and autonomous vehicles
LG Electronics has created new robotics and autonomous vehicle divisions directly under the CEO's guidance as part of its year-end restructuring.
Satya Nadella: first futuristic CEO of Microsoft
Is it time for the technology sector to grow and realize how much it affects the world?
UNSW has found a way to access the information stored in the atoms
Researchers have developed a compact sensor to access information stored in the electrons of individual atoms.
Garry Kasparov is surprisingly optimistic about our future lords of IA
Interview: The chess grandmaster, beaten by a computer, said that smart machines would make us smarter.
Coalition replaces Commonwealth Scientific Council as unions promise scientific inquiry
The Australian Council of Science and Technology was created by the Australian Government as the Commonwealth Scientific Council is thrown into the sea.
Why media rigged videos pose a danger to national and professional security (TechRepublic)
If you thought the bad news was bad, wait until hackers get their hands on artificial intelligence-driven face swapping technology, says Michael Fauscette, G2 Crowd representative.
Source link
