Sistema de automatização de mapeamento utilizando drones

Abstract

This study presents a methodological proposal for visual and photogrammetric mapping of an area simulating a crop field, using commercial drones as tools to support spatial documentation. The research was repositioned toward the agricultural context, focusing on organized aerial image acquisition, experimental photogrammetric processing, and visual delimitation of the study area. Initially, the DJI Mavic Pro Platinum drone was used during site recognition, operation planning, flight simulation, preliminary area delimitation, and mission organization. However, due to an operational limitation related to the camera of the equipment, the effective aerial image acquisition was performed with the DJI Mini 3 drone. This adaptation did not change the main objective of the study, which remained focused on validating a visual mapping workflow using drones. The collected images were processed in WebODM as a photogrammetric testing stage, allowing the visual reconstruction of the experimental area to be evaluated. The final processing used 70 images, all of which were reconstructed by the software. However, the report indicated the absence of automatic geographic reference, which limited the use of the generated product for direct measurements. Therefore, the final measurement was performed in QGIS using an original aerial image captured by the DJI Mini 3. The area simulating a crop field was manually delimited using a polygon, with a reference scale based on the approximate distance of 22 meters between two visible points in the image. As a result, an approximate area of 174.65 m² and an approximate perimeter of 91.20 m were obtained. The results have an academic and demonstrative purpose, aimed at validating the visual mapping workflow with drones. Geofencing is addressed as an auxiliary resource for spatial delimitation and operational organization, not as the technique responsible for map generation.