Detecção de linhas e falhas de plantio em cana-de-açúcar utilizando vants em diferentes alturas de voo

Abstract

Sugarcane is a crop of great economic importance, but its productivity can be compromised by planting failures. This study evaluated the use of Unmanned Aerial Vehicles (UAVs) in detecting such failures in an 8.3-hectare plot at COMVAP, in União–PI, using RGB images captured at altitudes of 90 m (2.13 cm/pixel) and 120 m (3.14 cm/pixel). The orthomosaics were resampled to 10 cm/pixel to compare operational efficiency. Image processing was performed in QGIS using the Green Leaf Index (GLI) and the Difference plugin. Most of the detected gaps ranged from 0.5 to 1.5 meters. Images captured at 90 m showed a lower percentage of detected gaps (8.77%) compared to those at 120 m (9.41%). The resampling to 10 cm/pixel reduced sensitivity to smaller gaps but maintained accuracy for larger ones, while also optimizing processing time and cost. It is concluded that the use of UAVs with RGB cameras, combined with the GLI index and manual refinement, is effective for mapping planting failures, supporting monitoring and decision-making for correction and crop management.