Tonini, AndreaCastelli, MauroBates, Jordan StevenLin, Nyi Nyi NyanPainho, Marco2024-10-242024-10-242024-10-172076-3417PURE: 101455505PURE UUID: 53f5d598-70be-4d37-9449-89d72ef1a8c9Scopus: 85207440600WOS: 001341381900001ORCID: /0000-0003-1136-3387/work/170276122ORCID: /0000-0002-8793-1451/work/170276213http://hdl.handle.net/10362/174020Tonini, A., Castelli, M., Bates, J. S., Lin, N. N. N., & Painho, M. (2024). Visual-Inertial Method for Localizing Aerial Vehicles in GNSS-Denied Environments. Applied Sciences, 14(20), 1-13. Article 9493. https://doi.org/10.3390/app14209493 --- This work was supported by national funds through FCT (Fundação para a Ciência e a Tecnologia) under the project UIDB/04152/2020 (DOI: 10.54499/UIDB/04152/2020)—Centro de Investigação em Gestão de Informação (MagIC)/NOVA IMS. This work was partially funded by the Deutsche Forschungsgemeinschaft under Germany’s Excellence Strategy. The authors would like to express their gratitude to the NOVA Impact Office (https://novainnovation.unl.pt/nova-impact-office/, accessed on 10 February 2024) for all the support provided.Estimating the location of unmanned aerial vehicles (UAVs) within a global coordinate system can be achieved by correlating known world points with their corresponding image projections captured by the vehicle’s camera. Reducing the number of required world points may lower the computational requirements needed for such estimation. This paper introduces a novel method for determining the absolute position of aerial vehicles using only two known coordinate points that reduce the calculation complexity and, therefore, the computation time. The essential parameters for this calculation include the camera’s focal length, detector dimensions, and the Euler angles for Pitch and Roll. The Yaw angle is not required, which is beneficial because Yaw is more susceptible to inaccuracies due to environmental factors. The vehicle’s position is determined through a sequence of straightforward rigid transformations, eliminating the need for additional points or iterative processes for verification. The proposed method was tested using a Digital Elevation Model (DEM) created via LiDAR and 11 aerial images captured by a UAV. The results were compared against Global Navigation Satellite Systems (GNSSs) data and other common image pose estimation methodologies. While the available data did not permit precise error quantification, the method demonstrated performance comparable to GNSS-based approaches.13626532engGNNS-denied environmentslocalization of aerial vehiclesvisual-inertial methodGeneral Materials ScienceInstrumentationGeneral EngineeringProcess Chemistry and TechnologyComputer Science ApplicationsFluid Flow and Transfer ProcessesSDG 9 - Industry, Innovation, and InfrastructureSDG 11 - Sustainable Cities and Communitiesjournal article10.3390/app14209493https://www.scopus.com/pages/publications/85207440600https://www.webofscience.com/wos/woscc/full-record/WOS:001341381900001