The Brazilian Agricultural Research Corporation (EMBRAPA), in partnership with Macnica DHW and InnerEye, is testing an algorithm that can accurately identify plant problems and suggest solutions. To do this, they use human brainwaves as a reference for the project, and two experts look at 1,500 images of diseased and healthy leaves to determine patterns.

The goal is to increase the flexibility of decision-making and reduce the losses in this regard. By improving the use of natural resources, both agriculture and animal husbandry get higher yields, and the population gets their hands on higher quality products. The experiment was carried out with soybean leaves, studying two diseases that are often found on plantations: Asian rust and powdery mildew.

One of the most interesting parts of the whole project is that they intend to apply artificial intelligence to various devices for farmers. According to EMBRAPA, the app can be inserted into branch machines, as well as mobile phones and drones. In this way, they will increase the supervision and care of plantations.

EMBRAPA is already planning the next steps

According to the disclosures, initially they want to identify the most common issues for commercial significance. They are also evaluating the start of corn and coffee research to improve the use of artificial intelligence and create even more options over time.

The system also has tools that allow it to learn, fix common bugs as it runs, and suggest improvements as it understands its environment. Researcher Jaime Barbedo says this step is critical to their safe continuation. “It makes the trained model more robust.”

Steps taken during testing

The AI ​​capture went through several steps to make the test pass, analyzing the factors so that experts can prove its effectiveness.

• Initially, the neural signals of the experts were synchronized in the helmet with electrodes;
• The two evaluated over 1,500 images of soybean leaves in a variety of settings;
• When captured, the AI ​​was able to determine experience, judgment, and image classification;
• The software trained the algorithm with models generated by spontaneous electrical current emitted by the brain;
• When they finished capturing, the software automated the diagnostics and made the move faster and more efficient;
• The results show that artificial intelligence is currently on par with experts.

The research team claims the technology will help with pasture rotation strategies for dairy farming, an area that lacks a specialized workforce. In doing so, they should quickly determine where to insert the pickets and proceed with the installation.

Mundo Conectado Deal Center: Selection of Discounts and Lowest Prices
Best deals on electronics, cell phones, TVs, soundbars, drones and more

Source: G1

…..