They might be small, but bark beetles can ruin a forest. In the US, tens of millions of acres have been devastated by them in the past decade alone. In Europe, however, a combination of drones and artificial intelligence might be giving trees a fighting chance.
Bark beetles burrow into trees and lay eggs under the surface of the bark where the larvae feed on the tree’s inner layers  and eventually chew their way out. All of this damages the tree’s vascular system, fatally weakening it.
Within months of a bark beetle infestation, great swathes of a forest can be irreversibly damaged, leaving acres of grey leaves and dead wood. If the problem is spotted in time foresters can tackle it, by cutting away infected trees or applying pesticides. But they have to know it’s happening first.
Advertisement
Skylab, a drone and software company based in Hamburg, Germany, is using drones fitted with multispectral cameras to scan spruce forests in Germany. They can image the ground below at a resolution of two centimetres and are mounted on large fixed wing drones that can fly for an hour per charge. The drones skim above the canopy, covering up to 150 kilometres a day, looking for signs of infestation.
Healthy hue
Thriving leaves with plenty of chlorophyll absorb most of the blue and red light but reflect more green light and infrared. But stressed plants produce less chlorophyll and reflect more light across the spectrum, while absorbing a little more in the near-infrared. The cameras can pick up on these differences, creating maps of the forest’s health.
Machine learning algorithms that are trained on the camera data can then churn out an infestation analysis – a disease risk index for each tree. The data is calibrated using spot tests of trees.
The idea is that workers can focus on disease hotspots first. “When you detect these stress signals you can work out the densest infestation spot, “ says Skylab’s Rene Heim.
It’s not just bark beetles. The team is also monitoring apple orchards for early signs of apple scab, a disease that blights the crop and makes unpleasant spots on the fruit, which can knock 90 per cent off its value. Skylab is running a trial where fungicides are applied to different areas to see how it affects the spread of the disease as picked up by the camera. “If you can spot the disease you can apply the fungicides in a more targeted manner, saving money,” says Heim. Skylab is experimenting with different sensor types to see which gives the best results, including temperature detection.
Another Sklab project is looking at vineyards in Italy, using the drones and cameras to hunt for signs fungus associated with a number of diseases, some of which can strike suddenly and destroy plants within days.
Skylab is not the only player in this field. Drones paired with machine learning and spectral imaging equipment are already helping to identify farmland that needs more watering or specialist treatment, while a team in Denmark are using drone-mounted cameras to spot weeds from above, and marking them for precision treatment with herbicides.
Danny Donoghue, at Durham University in the UK, says that UAVs really lend themselves to these kinds of applications. “This technology opens up new opportunities for environmental monitoring at the scale of the individual plant that was not previously possible,” he says.
