Augmented reality (AR), despite its great potential, still struggles to be widely used in real construction processes due to difficulties in registering holograms in mixed indoor-outdoor scenarios. Since a definitive technological solution for inside-out AR registration does not exist yet, a seamless markerless augmented reality registration system, integrating real-time kinematic positioning (RTK), inertial measurement units (IMU) technologies and image comparison based on convolutional neural networks (CNN), has been proposed by the authors. Experiment results have shown the need to continuously register AR at regular temporal interval within 1 seconds and/or 1 meter to achieve “fine-precision” positional accuracy (i.e., 0.10 m).
A first evaluation of the seamless markerless augmented reality registration system supporting facility management / Messi, Leonardo; Spegni, Francesco; Vaccarini, Massimo; Corneli, Alessandra; Binni, Leonardo. - (2024). (Intervento presentato al convegno 2024 European Conference on Computing in Construction tenutosi a Chania, Greece nel 14-17 July, 2024) [10.35490/EC3.2024.324].
A first evaluation of the seamless markerless augmented reality registration system supporting facility management
Messi, Leonardo
Primo
;Spegni, FrancescoSecondo
;Vaccarini, Massimo;Corneli, AlessandraPenultimo
;Binni, LeonardoUltimo
2024-01-01
Abstract
Augmented reality (AR), despite its great potential, still struggles to be widely used in real construction processes due to difficulties in registering holograms in mixed indoor-outdoor scenarios. Since a definitive technological solution for inside-out AR registration does not exist yet, a seamless markerless augmented reality registration system, integrating real-time kinematic positioning (RTK), inertial measurement units (IMU) technologies and image comparison based on convolutional neural networks (CNN), has been proposed by the authors. Experiment results have shown the need to continuously register AR at regular temporal interval within 1 seconds and/or 1 meter to achieve “fine-precision” positional accuracy (i.e., 0.10 m).I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.