Infrastructure-Free Localization System for Augmented Reality Registration in Indoor Environments: A First Accuracy Assessment

Navigation systems combined with Augmented Reality (AR) constitute an effective solution for helping user accessing unfamiliar environments, like indoor public facilities. Implementing AR navigation systems requires determining the 6-Degrees-of-Freedom (6-DoF) localization of the user. In indoor environments, lack of Global Navigation Satellite System (GNSS) signals makes localization more challenging compared to outdoor environments. A variety of positioning systems have emerged for indoor localization which are based on several system strategies, location methods, and technologies. High-accuracy, low-cost, easy to use, and no need for any technical expertise are key features to ensure large-scale application of AR navigation systems. This study provides an answer to these key requirements by proposing a markerless infrastructure-free localization system for AR registration in indoor environments based on the comparison between a query image and a 3D mapping assumed as a reference. Since most people in public facilities are equipped with tablets or smartphone devices, they should be provided with further functionalities that will help them in day-to-day life and work. To this purpose, the proposed localization system was implemented as a web-service and tested in a university campus to assess applicability for AR registration purposes. Experiment results showed very promising localization accuracy and computational efficiency satisfying the “fine-precision” accuracy threshold (i.e., 1°/0.100 m) for near real-time AR applications.