Menu
Applied science for agribusiness optimization: LATU, Latitud, and URUPOV partner on AI project for soybeans
The Technological Laboratory of Uruguay (LATU) and its research, development, and innovation (R&D&I) foundation, Latitud, have signed an agreement with URUPOV, an association of agricultural companies that researches and promotes new plant varieties. This new partnership will launch an innovative project to identify soybean crops throughout Uruguay. This agreement is the result of a synergy between these institutions and the private sector, and it seeks to develop an artificial intelligence (AI) tool to optimize the efficiency and productivity of soybean production.
The event began with remarks from LATU and Latitud President Lucila Arboleya, who stated, "This project is a response to a specific request from the production sector to LATU and Latitud." She also noted that "we are in a context where the country needs to grow more, and the export sector is the variable we are betting on. Helping this sector make a technological leap is the reason our institution exists."
Following her, URUPOV President Jorge Erro spoke, describing the project as "a change of era, like thinking about the industrial era when we were embroidering, and now everything will be automatic." Regarding the partnership with LATU and Latitud, he expressed being "surprised by two things: the quick response and the rapid adaptation to our vision and our need to move from manual, polygon-based satellite image recognition to an automated solution."
How the project started and what it involves
The project centers on developing an algorithm using neural networks and satellite images to detect the presence of soybeans in Uruguayan fields with 97% accuracy. A second phase will aim to identify the specific varieties being cultivated.
During the presentation, URUPOV Executive Director Diego Risso reviewed the project's background, which includes 10 years of prior work in processing satellite images. This history led to international partners becoming aware of their work and wanting to "develop an algorithm with URUPOV using artificial intelligence and neural networks to process satellite images to detect soybeans in Uruguay and use our country as a training ground for the region."
With this momentum, Risso noted, "it was considered opportune to find our partners at LATU to carry out this process. We believe this has many advantages for us, for our members, but also for Uruguay as a country and for the entire sector."
The neural network designed for this purpose will be a series of information nodes that allows a program to recognize patterns and solve common AI problems. This represents a significant technological leap, moving from a manual identification system for soybean fields to a fully automated one.
Blanca Gómez, a Latitud researcher and the project leader, explained that the team will start with a database of nine years of satellite information and 250 GPS points with DNA-based variety identification, which URUPOV already has. This data will be used to train the neural networks to create a reliable and accurate tool.
A team combining knowledge
Gómez highlighted the project's multidisciplinary team, which includes experts in systems, biology, genetics, agricultural engineering, and data science, ensuring its technical robustness.
She also praised the network of technical partners linked to the organizations, including the National Institute of Agricultural Research and the National Seed Institute, as well as Microsoft, whose only AI laboratory in Latin America is located at LATU's Innovation Park.
