The NSU Artificial Intelligence Center has started testing the “Colorado Detector” system, which uses drones to precisely locate pests in potato fields. This innovation enables a shift from blanket chemical spraying of land to targeted treatment—an approach that is both more environmentally friendly and cost-effective.
According to the university’s press office, the project is being carried out on behalf of the company “Dary Ordynska.” Currently, agronomists monitor crops by conducting spot checks and walking through fields diagonally. Experts note that this method only provides an overall assessment and does not guarantee that insect breeding hotspots will be detected and eliminated at an early stage.
“If you pinpoint the exact area, you can treat it locally without dousing the entire field with chemicals,” explains Maxim Magerov, an engineer at the NSU AI Center.
In developing the technology, the authors drew on previous experience. Specifically, the team had worked on locating people using drones: in that project, neural networks processed video footage onboard, sending already analyzed information back to the ground.
The same principle was applied to agricultural challenges. A fan-shaped module with four cameras is mounted on the drone, providing the widest possible field of view. The images are analyzed by a high-performance single-board computer. Using machine vision algorithms, it identifies adult Colorado potato beetles.
The developers explain that the system’s key feature is its focus on detecting arthropods rather than damaged plants. In the early stages, leaf beetles are already hiding on the bushes but leave no obvious traces.
“It’s harder to spot larvae from the air because they are often under the leaves, and the current version hasn’t been trained to recognize them yet. However, adult beetles sitting on the upper parts of plants can be detected individually,” says the center’s employee.
After processing, the results are sent to a ground station. The field worker opens a special program on a computer and sees a map showing the pest’s locations.
Based on this analysis, localized treatment can be carried out on specific areas, thereby reducing insecticide use and minimizing the environmental impact. The innovators claim that this approach aligns with the global trend of precision agriculture.
The system is currently being tested in the fields. Researchers are collecting new data and evaluating the AI in real time. Additionally, efforts are being directed toward fine-tuning the hardware setup and improving its accuracy. Afterward, the technology will be patented and handed over to agricultural producers.
About the Author
