Drone uses AI to detect invading plants

CSEM, the Swiss technology innovation centre is part of a consortium which is developing an AI-equipped drone to identify invasive and potentially damaging plants.

Buddleia, or butterfly bushes, and Japanese knotweed can replace native plant species and others like narrow-leaved ragwort are poisonous. They can cause damage to the infrastructure, damaging railway tracks and signal installations or reduce sight lines.

The consortium which also includes the University of Zurich and the Zurich University of Applied Sciences ZHAW, SBB (Swiss Federal Railways) and ExoLabs, is supported by Swiss innovation agency, Innosuisse. The partners have developed this neophyte radar drone which uses AI to detect these plants from aerial images.

More ground can be covered using drone images which also allows sight of less accessible areas in which the ‘invaders’ have taken root. Work is focused on using AI to analyse the images taken from great heights to identify neophytes and distinguish them from native plants to track and monitor growth to understand how these plants spread and then put measures in place to target and repress them.