COMPARISON OF NDVI OBTAINED BY DRONE AND SATELLITE IN THE PHENOLOGICAL PHASES OF GRAPEVINE

Abstract

Several interactions of the biophysical environment, whether by biotic or abiotic factors, hinder the study of crops' behavior using standard methods. The use of platforms with high spatial resolution sensors enables detecting heterogeneity of crops such as grapevine through NDVI. This study aimed to evaluate the NDVI obtained using images from the AggieAir Minion UAV and Landsat 8 in different phenological phases of the Pinot Noir grapevine. The images were collected in Sacramento County, California, USA. Image processing was performed using Quantum GIS (QGIS). We observed an underestimation of the NDVI calculated using Landsat 8 images in the berry growth phase, presenting lower values (0.68) than those obtained by the VANT (0.87). The NDVI values should be closer to 1.0 in this phase because the plant is in full vegetative development. The details of the images depends on the spatial resolution of the sensor that the platform carries. Thus, considering that the distance between the lines of the vine was 3.35 m, and due to the pixel size of the Landsat 8, a larger soil-exposed area was detected as a consequence of the maximum similarity of veracity used by the algorithm language employed in the programming software. The resolution influences the vegetative index values in phenological phases of grapevine and, consequently, the biophysical parameters estimated by them, showing that the UAV images convey values more similar to reality for NDVI evaluation in grapevines.