Brazilian technology accelerates soybean quality classification.

Brazil is among the world's largest producers and exporters of soybeans, which represent approximately 5% to 7% of the country's GDP, generating around R$ 360 billion annually. Grain quality classification is a crucial step in commercial relationships between producers, storage facilities, trading companies, processors, and exporters. The Grain Analytic System (Gras) project is a pioneering technology for the automatic classification of soybean grain defects, using photonics, computer vision, and artificial intelligence (AI). The goal is to make soybean grain quality assessment faster, more objective, and standardized, reducing analysis time from 90 to four minutes.

Gras is the result of a partnership between the Brazilian Industrial Research and Innovation Company (Embrapii) and the startup Brasil Agritest – the first company to sign a contract with the Embrapii Itech-Agro Unit. The machine's operation is very simple. A sample of soybeans is placed in the equipment, which analyzes the grains individually using special lighting, cameras, and AI. The system then identifies and classifies different characteristics and defects in the grains without the need to cut them.

Although soybeans are of great importance, determining their quality worldwide is still based on visual inspection, carried out by the few existing classification professionals through a completely manual process. "It's like transforming a predominantly manual inspection into an automated, fast analysis based on objective criteria," summarizes researcher Débora Milori, coordinator of the Embrapii Itech-Agro Unit.

According to Thomas Martins Ceglia, partner, co-founder, and head of business development at Brasil Agritest, Embrapii was fundamental in enabling the technological development of the project, bringing the company closer to the research environment. “The partnership allowed for the advancement of the solution and the development of the technological components necessary to transform the concept into equipment applicable to the market. Furthermore, the project allows for the improvement and parameterization of the equipment, adjusting its systems and models, contributing to greater precision, standardization, and reliability of the results,” Ceglia points out.

Gras's main innovation is the integration of photonics, imaging systems, computer vision, and artificial intelligence into an automated solution for evaluating grain quality. "In addition to significantly reducing analysis time, the technology helps to decrease the subjectivity inherent in visual assessment and increase the standardization, repeatability, and traceability of results," emphasizes Débora.

Ready-to-scale solution

In 2023, the startup received support from VLI – a logistics solutions company that operates terminals, railways, and ports – to boost the completion of the equipment. Today, Gras is in the final stages of development and validation in an operational environment. The equipment has already been tested in real-world conditions and is being improved based on the results obtained in tests with industry partners. "Currently, the company is seeking investments to make the final adjustments and prepare the equipment for large-scale production," says Ceglia.

For the industry, the technology represents an opportunity for digitization and automation of an activity still heavily dependent on manual procedures. For society and the production chain, the solution can contribute to greater transparency, reliability, and traceability in grain quality assessment, strengthening the competitiveness and digital transformation of Brazilian agribusiness.