Automatically generated translation
On February 26, Terres Inovia previewed Berl'eyes, an innovative application that uses artificial intelligence to recognize and count flea beetle larvae in rapeseed crops, at the Salon International de l'Agriculture. The tool will be detailed in a webinar on May 21. A strategic innovation for the oilseed sector.
Developed over the last three years as part of the RESALT (Phosmet exit plan) and AgroEcophen projects, with several research partners, this tool meets a major challenge: better management of flea beetles, the main rapeseed pest, in a context of growing insecticide resistance.
Based on the Berlèse method - which involves collecting larvae that have fallen into a basin after the plants have dried - Berl'eyes automates the tedious and time-consuming counting stage, which until now had been carried out manually. The user takes a photo, sends it via the application and obtains a result in 15 to 30 seconds, with a high level of accuracy (up to 98% according to tests). The tool makes counting more reliable and harmonized across the country.
This tool is a first for the oilseed sector: " This innovative application is strategic, as it could ultimately lead to better yields with plots less affected by flea beetles, as well as fewer crop treatments ", points out Laurent Rosso, Managing Director of Terres Inovia.
Berl'eyes is in line with one of the technical institute's strategic priorities in promoting agro-ecological transition. By facilitating the evaluation of infestation levels, this tool helps farmers, technicians and advisors to reason out interventions, avoiding unnecessary treatments.
Free of charge and accessible online, it is aimed at both farmers and professionals from chambers, cooperatives and other agricultural advisory bodies. Currently in the test phase, it will be officially launched in May, and a webinar will detail the details. It will then be further developed (recognition of larval stages, mapping, extension to other pests).