“…computer vision techniques can be used to identify the causes of symptoms observed in plants…”
Jayme Barbedo is a researcher at Embrapa’s Digital Agriculture Unit, graduated in electrical engineering from the Federal University of the State of Mato Grosso do Sul, M.Sc. and Ph.D. from the State University of Campinas.
Jayme Barbedo, researcher at Embrapa
AgriBrasilis – What is Big Data, and how is it used in agriculture?
Jayme Barbedo – With the proliferation of different types of sensors in agriculture, a large amount of data is being created, a phenomenon commonly referred to as “Big Data.” This data often contains information that can greatly support farmers in decision-making. However, extracting valuable insights from this data is complex and usually requires sophisticated data processing techniques.
AgriBrasilis – What are the applications of computer vision in plant health?
Jayme Barbedo – There are two main applications of computer vision regarding plant health. Firstly, if there is a specific disease of interest that needs early detection, computer vision techniques, paired with cameras installed inside the farms or with agricultural machinery, can issue an alert when the disease is detected. Secondly, computer vision can identify the causes of symptoms observed in plants. This technology, for example, could be implemented as a mobile app.
AgriBrasilis – How can Artificial Intelligence (AI) identify plant diseases? Is this technology already available on the market?
Jayme Barbedo – AI efficiently supports the extraction of information present in images. With recent advancements in AI techniques, nearly all identification and classification issues can be addressed using AI. However, to be effective, the data used to train AI models must cover all possible variations of conditions that are relevant to the problem.
In agriculture, where many factors introduce variability, a large amount of image data is needed. Consequently, although AI-based agricultural technologies are still relatively limited, there have been successful applications, such as the case of automated weed detection and removal.
AgriBrasilis – Can AI replace professionals in agriculture?
Jayme Barbedo – In the short and medium terms, agronomists will remain essential. Although AI may impact certain roles, a significant labor shortage persists in agriculture in Brazil, particularly in fruit production, where recruiting enough workers for farm tasks and for harvests has been challenging. Therefore, automating certain processes will be essential for ensuring that food production meets the required quantity and quality levels.
AgriBrasilis – What projects does Embrapa’s Digital Agriculture Unit have in artificial intelligence?
Jayme Barbedo – We have several initiatives in AI. Beyond plant disease detection and recognition, we are also working on animal monitoring via drones, with the goal of tracking the number of animals and also of detecting anomalies that require attention, such as sick animals, thefts, or births. We are also working on fruit detection and biomass measurement. Recently, we have begun exploring generative AI to create technologies that provide information to users as reliably as possible.
AgriBrasilis – What are the challenges in developing this research? Can variability in the farms be controlled?
Jayme Barbedo – The variability of agricultural conditions is the biggest challenge. Some measures, such as filter screens or structures to control lighting, can manage certain image characteristics, but they increase the complexity of data collection, limiting practical applications. The best technical solution is to create image databases that adequately represent this variability, although achieving this is particularly difficult when dealing with images.
READ MORE: