Uma aplicação para suporte a atividades de manutenção em campo baseada em Realidade Aumentada sem uso de marcadores

Abstract

In recent years, several Augmented Reality (AR)-based applications have been developed for use on mobile devices, along with the large offer of more accessible devices with significant embedded technology. More recently, Augmented Reality has also been sought by companies, which have envisioned the possibility of its application, obtaining cost savings in several operational activities, such as maintenance tasks. The motivation for this work arose from the observation of a current problem for telecommunications operators, which is the maintenance of an updated optical network register available online, with information regarding geographic location and use of network elements. By evaluating publications and commercial applications related to this specific topic, it is possible to find some initiatives that apply technologies along with AR, such as image recognition or geolocation. However, the developed applications do not explore the benefit of using all those resources combined. The objective of this work is the implementation of a new application, called “ARNet”, based on Augmented Reality, geographical data and image recognition on mobile devices, in order to allow a user to receive the identification of a desired optical network equipment directly on the screen of his smartphone, as well as additional information to assist his tasks of interacting with the equipment. The proposal of this work brings a novel solution for the use of AR in maintenance activities, as it fetches information about a desired object in real time, in an existing database, integrating image recognition techniques with information regarding the geographical position of the object and allowing the user to update information related to it, at the end of his activities. A systematic review was carried out to compare approaches already implemented that could contribute to the development of the solution. Tests were also carried out in a commercial application, and case studies related to this theme were evaluated. The proposed application was then developed and validated by a case study, applying the system to fiber optics installation activities, performed by field technicians. The image recognition worked properly, with the correct identification of the georeferenced object and displaying the requested information through AR. It was verified, therefore, that the proposed application can be used for tasks performed by field technicians and can reduce the average maintenance time and provide greater accuracy of the information needed by the technicians.