A Perth-based Company spun out from a traditional environmental consultancy has developed a cutting-edge AI software product suite being used by mining giants to better manage natural environments.
After seven years of development and in-house use, Tyton Ecological Intelligence (EI) is soon launching a SaaS version of its Tyton AI software – a machine learning classification tool, customised to ecological monitoring, which will be marketed globally.
Supporting the Company’s existing Tyton EIS analytics platform, the new software gives users access to Tyton’s machine learning vegetation ‘mega-model’ built using data accumulated from large land areas in Western Australia (WA) and other parts of the world.
Tyton AI uses vision transformers, along with other machine learning and data science techniques, to analyse aerial imagery of landscapes. This allows identification of individual plants, vegetation lifeforms and species across thousands of hectares.
Tyton EI Founder and CEO Dr Cam Mounsey said the Tyton AI machine learning tool was changing the way companies monitor the environment at scale.
“We’re enabling environmental managers to harness the power of machine learning by making it easy. They can now monitor entire study areas at individual plant scale with a high degree of accuracy, without needing a degree in machine learning,” he said.
“We have developed a machine learning model trained on billions of pixels captured from swathes of land stretching across WA and the world.
“That is our secret sauce – giving clients a data-powered, machine learning ‘mega-model’ that can be dialled in to the environments they want to gather more information on, monitor and rehabilitate.
“So once a client is using our software, they can load it up with their own aerial data and use our pre-trained ‘mega-model’ to analyse sites at enormous scale.”
Mounsey says Tyton’s software is helping land managers produce better environmental outcomes, allowing them to run a multitude of inferences on data collated and stored on Tyton’s platform.
Even in environments new to Tyton software, the AI element can fine-tune fresh data added to existing mega-models with up to 90% identification accuracy.
The technology is a vast improvement on traditional on-ground monitoring methods, and far exceeds alternative spatial analysis techniques using generic GIS, which may require hundreds of hours of manually drawn lines around vegetation and plants to create training datasets.
Tyton’s new software development has been led by CTO Alex Dunmow over the past year and is anticipated to be launched globally in coming months.
“It’s novel, there’s nothing like it on the market,” said Dunmow, who formerly ran Ninja Software.
“We’re applying advanced AI to help solve real environmental challenges. It’s satisfying to see our tech making a tangible difference in how mining companies manage and restore ecosystems.
“Doing this from Perth shows we can be at the forefront of both AI development and environmental tech. It’s a good combination of interesting work creating positive outcomes and shows how tech can have real-world impact.”
About Tyon Ecological Intelligence (EI)
Tyton EI is setting a new standard in environmental monitoring by combining state-of-the-art machine learning and artificial intelligence with an integrated modular analytics platform to improve ecological management.
Already being used by global Tier-1 mining companies, Indigenous and government land management groups, its Tyton AI software is able to identify vegetation lifeforms, genres and species at landscape scale.
This is supported by the Tyton EIS platform, which stores collated datasets and facilitates powerful analyses to assist environmental management knowledge building and decision making.
