Sistema com baixo custo de captura de imagens aéreas e detecção de anomalias em plantações em tempo real

Abstract

This work presents a low-cost system for aerial image capture and anomaly detection in real time. A fully autonomous fixed-wing UAV (Unnamed Aerial Vehicle) was developed for overflying the plantations. As system validation, two case studies are presented. Both happen in sugar cane plantations near the coast of Brazilian state of Paraíba. The first is a mill producer of distilled cachaça and the second is a controlled environment plantation. The NDVI (Normalized Difference Vegetation Index) support the anomaly detection. In addition to planting vigor, water insufficiency and exposed soil were identified by the system. It was found that it is possible to detect regions of interest during flight, and based on the NDVI to predict planting productivity.