Sistema de detecção de Tomates (SDT): uma ferramenta de Gestão para agricultores nas fases de maturação

Abstract

This study presents the development and application of a computer vision system for the detection and classification of tomatoes at different ripening stages using the YOLO v8 model. The system was tested in a controlled agricultural environment, where images and videos were analyzed to identify ripe, green, and rotten tomatoes. To facilitate data visualization and interpretation, an interactive statistical dashboard was implemented, enabling real-time monitoring of agricultural production. In addition to fruit detection, the system was integrated with a climate monitoring module, providing information on temperature, humidity, and precipitation to support decision- making processes. The results indicate that the model demonstrated good performance in identifying ripe and rotten tomatoes, but faced challenges in accurately detecting green tomatoes due to lighting variations and the lower representation of this class in the dataset. This research highlights the potential of artificial intelligence in agriculture, fostering greater efficiency, automation, and quality control in food production.