A new AI system will survey for invasive species such as the Japanese knotweed (pictured above).

Invasive plant species cost the U.K. economy an estimated £1.7 billion annually. Though about 3,000 alien plant species are found in the country, only 234 of them are currently considered a threat to its ecosystem. These harmful species have properties such as the ability to reproduce rampantly or to spread themselves across a range of habitats. In so doing, they develop an edge over native flora. The U.K. is now taking recourse to AI to counter these disruptive aliens called Invasive Non-Native Species (INNS). (BBC Science Focus)

Tracking INNS is expensive

Surveyors currently conduct an assessment of the proliferation of INNS alongside roads, railways, and highways.

Employing AI reduces the cost and time for these surveys. It also generates more accurate, location-specific data about the growth of the INNS.

The UK Centre for Ecology and Hydrology (UKCEH) and Birmingham-based company Keen AI are collaborating on the development of an AI system to survey roadsides for INNS.

Says the Keen-ai website: “Such work currently requires the temporary closure of roads to ensure the safety of surveyors. We estimate costs associated with this to be up to £5800 per week covering 50 miles. An ecologist supported by our system could remotely survey 120 miles (20 mph for 6 hrs) without road closures. This is a cost per mile of £9 vs. £116.”

How it works

High-speed cameras mounted atop vehicles will survey up to 120 miles of roadside vegetation daily.

Keen AI uploads these images, along with their GPS coordinates, to an online platform for identification of the species by UKCEH scientists.

The tagged images would be fed to a machine-learning algorithm for training in the identification of INNS species.

These include Japanese knotweed, rhododendrons, Himalayan balsam, and cherry laurel.

Once trained, the AI model would be able to rapidly analyze roadside images for INNS, reducing the effort, time and cost for controlling these plants.

The project focuses on roadside surveys because invasive plant species tend to grow in corridors.

The UK government’s Innovate UK agency is funding a 10-month pilot to survey north Wales and Birmingham to locate INNS using AI.

Related Story:  Goldspot’s AI Boost’s Success in Mineral Exploration