Scientists at Novosibirsk State University (NSU) have begun field trials of a system that uses drones and neural networks to find Colorado potato beetles in potato fields. As reported by TASS in the university's press service, the development is intended to replace continuous chemical treatment of crops with targeted, more environmentally friendly and economical pest control.

The drone flies at a height of three to four meters above the bushes and is equipped with a module of four cameras arranged in a fan to capture the widest possible strip of plantings in one pass. An enhanced single-board computer is installed on board, which receives images, analyzes them using machine vision algorithms, and detects adult beetles. Unlike traditional methods, the system does not focus on damaged leaves but searches directly for insects sitting on the upper part of the plants, which allows for early detection of infestation before feeding traces appear.

After image processing, the results are transmitted via a digital channel to a ground station. The agronomist sees a web application on a laptop with a field map showing the detection points of beetles and decides on local treatment of specific areas. This reduces insecticide consumption and environmental impact, aligning with the global trend of precision agriculture. The development was carried out at the NSU Artificial Intelligence Center.