Researchers in Switzerland have developed a combined AI and radar technology for UAVs monitoring invasive plants, writes Nick Flaherty.

The system monitors invasive alien plants (invasive neophytes) that are difficult to detect and tackle. Invasive species like Japanese knotweed, which can replace native plants, or narrow-leaved ragwort, which is poisonous, are becoming increasingly common. They can damage agriculture, and infrastructure such as train tracks and signals.

A consortium of research lab CSEM, the University of Zurich and the Zurich University of Applied Sciences ZHAW, as well as Swiss Federal Railways (SBB) and ExoLabs, is taking a new approach to monitoring these species from the air.

AI is already capable of detecting certain plants or animals in high-resolution images, but these are aerial photos taken from great height, where small neophytes are just a few pixels in size, making it difficult to tell them apart from native plants. Vegetation ecologists, and data and sensor specialists are working together on the neophyte radar.

The AI is trained to read high-resolution images taken by a UAV, identify which fields are covered in invasive alien species and mark them in blue.