Estimativa da produtividade de grãos de híbridos de milho a partir de imagens multiespectrais obtidas por drone

Abstract

Remote sensing using Unmanned Aerial Vehicles (UAVs) has emerged as an efficient tool for estimating agricultural yield. Therefore, the present study aimed to estimate maize grain yield from multispectral aerial images. The experiment was conducted in a randomized block design (RBD) with a split-plot scheme, combining two planting systems (single row and double row) and 12 maize hybrids, totaling 72 experimental units. The images were acquired using a DJI Mavic 3M drone and processed in Agisoft Metashape and QGIS software. The accuracy of regression models based on 14 vegetation indices (VIs) was evaluated using the root mean square error (RMSE) and relative root mean square error (RMSE%). The results indicated a higher average yield in the single-row planting system. The TCARI index stood out by showing the highest positive correlation with grain yield and accuracy classified as excellent (lowest RMSE and RMSE% values). It is concluded that the TCARI index, associated with UAV multispectral images, constitutes an effective and highly accurate alternative for monitoring and estimating maize yield in both evaluated planting systems.