Mapeamento de copas de cajá (Spondias mombin L.) utilizando imagens de veículo aéreo não tripulado (VANT)

Abstract

Precision agriculture employs advanced technologies to monitor and manage crops efficiently and sustainably. The study aimed to extract and estimate individual biometric variables, diameter and crown area of a cajazeira orchard in Teresina-PI, using an unmanned aerial vehicle (UAV). The aerial images were captured using a DJI Air 2S drone on flights at 30, 60 and 90 meters high. The data was processed using QGIS and Orfeo Toolbox software, using manual canopy delineation, vectorization and object-based analysis (OBIA). The average diameter and crown area were measured on site and compared to the estimates obtained by analyzing the UAV images. The best results in terms of average diameter and crown area measured by axes were in flights at 90 m height. Manual vectorization showed a high correlation between field measurements and those obtained from UAV images, especially for flights at 30 and 90 m height. OBIA segmentation showed great potential for automation, correlating with manual vectorization at the different heights, being more efficient at 30 and 90 m and with lower errors at 30 m. It can be concluded that the use of UAV images is an efficient tool for estimating biometric variables of cajá (Spondias mombin L.) canopies.