Augmented reality (AR) systems are getting more and more popular in several application fields, such as medicine, education, cultural heritage, etc. The recent hardware development of AR-oriented smartglasses extends these applications to hands-free contexts, in which the user cannot hold a tablet or a smartphone device. An example of this scenario is the AR-supported technical training of human operators. In this paper, we propose an AR smartglasses-based system for the training of technical staff working on ship engine parts. The size of components and the impossibility to use fiducial markers led to the development of an ad-hoc detection and tracking algorithm to align virtual parts over the real-world images. The system is currently under evaluation by a large multinational company leader in the field of marine market.
