AUTOMAÇÃO ROBÓTICA: SOLUÇÕES SUSTENTÁVEIS E INCLUSIVAS: TECNOAGRO

Abstract

The object of this project consists of the development of a robot assistant sustainable agriculture, an innovative autonomous system that aims to optimize agricultural practices through advanced technologies. Aligned with the Sustainable Development Goal (SDG), a worldwide appeal to make planet Earth prosperous (UN, 2015). The proposal meets the following SDGs: Zero Hunger and Sustainable Agriculture (SDG 2), Consumption and Responsible Production (SDG 12) and Land Life (SDG 15). The robot will be able to identify and remove diseased plants with high precision, reducing the need for manual interventions and the use of chemical inputs, promoting thus a more sustainable and efficient approach to plantation management.Automation in agriculture has benefited from learning technologies machine and robotics to improve processes and optimize plant management. An example is the development of robots with advanced capabilities to identify and remove plants sick. This system follows a two-step approach: identification of diseased plants and precise removal of these plants. In the first step, the system is designed to identify diseases in plants specific, using machine learning. Convolutional neural networks (CNNs) are especially efficient for this purpose due to their ability to learn and recognize complex visual patterns (LeCun et al., 2015). These networks have been widely used in agriculture for tasks such as detecting diseases in crops, in which can analyze visual characteristics of diseases, such as spots and changes in leaf coloration (Ferentinos, 2018). The robot's cameras capture images in time real life of plants, which are then analyzed by CNNs to identify signs of diseases with precision. The training database used contains images of healthy and sick, which allows the system to distinguish with high accuracy between healthy plants and plants with signs of disease such as yellowing leaves, spots and wilting (Mohanty et al., 2016).