EN 
PT_BR An autonomous robotic system, operating at night and named LumiBot, is capable of generating data that allows the construction of models for the early diagnosis of nematodes in cotton and soybean plants, even before symptoms appear. Developed by Embrapa Instrumentação (SP) in partnership with the Cooperativa Mista de Desenvolvimento do Agronegócio (Comdeagro), from Mato Grosso, LumiBot emits ultraviolet-visible light onto the plants and analyzes the fluorescence captured in leaf images using scientific cameras.
Cotton and soybean farming are of enormous economic importance to the country, with a record harvest projected for the 2025/26 period, with 4.09 million tons of cotton lint and 177.67 million tons of soybean grains, according to estimates from the National Supply Company (Conab). However, both crops face the threat of a microscopic parasite measuring 0.3 to 3 millimeters in length.
Hit rates above 80%
The robot is a prototype, but it is already showing promising results in diagnosing nematode infections in greenhouse experiments, with approximately seven thousand images collected over three years of research.
“We were able to generate data and models with accuracy rates above 80%, in addition to differentiating water stress diseases,” says researcher Débora Milori, coordinator of the study and of the National Agrophotonics Laboratory (Lanaf).
The next stage of the study will be the development of equipment for field operation, such as adapting the optical apparatus to an agricultural vehicle like a grasshopper sprayer or a rover vehicle.
Fewer chemicals in agriculture.
According to the researcher, the conventional method of nematode control is based on the application of nematicides to the soil or seeds before planting. These products work by reducing the nematode population near the roots. However, this application is costly, can cause environmental impact, and its effectiveness varies depending on soil conditions.
Other control strategies include biological control, crop rotation, and the development of resistant cultivars. “A more efficient and economical alternative would be monitoring the planted area, applying control strategies only in the effectively infested regions. However, there is still no commercial equipment capable of detecting the presence of the disease in plants early. In this context, the use of photonic techniques emerges as a promising solution,” states the study coordinator.
According to Comdeagro consultant Sérgio Dutra, early disease diagnosis is fundamental so that farmers can act quickly and in a localized manner. “This avoids the excessive use of chemical pesticides and reduces environmental impact, an important advancement for precision agriculture in Brazil. It is also possible to improve fiber quality and guarantee greater profitability for the producer,” assures the specialist.
Proof of concept
LumiBot has the support of Embrapii Itech-Agro (Integration of Enabling Technologies in Agribusiness) from Embrapa Instrumentation, which seeks to develop a prototype of the equipment to minimize costs and boost the soybean and cotton production chain. Embrapii is a social organization that connects companies and research institutions to develop innovations, such as LumiBot.
The project involved developing a proof of concept for early disease diagnosis using photonic techniques in cotton production systems, in partnership with Comdeagro.
The physical and automation project was designed in partnership between Embrapa Instrumentação and the company Equitron Automação, from São Carlos (SP), tailor-made for operation in the greenhouse.
Photonics is a branch of physics that studies applications of light, its generation, manipulation, and detection. The techniques are widely applied in various fields due to their sensitivity, precision, and high portability potential.
