The real-time detection of context information in marker-less Augmented Reality (AR) scenarios with a large amount of key-point based descriptors still represents a scientific challenge. In this demo paper we present ...
详细信息
ISBN:
(纸本)9781479938407
The real-time detection of context information in marker-less Augmented Reality (AR) scenarios with a large amount of key-point based descriptors still represents a scientific challenge. In this demo paper we present a mobile assistive work system that uses key-point based descriptors in marker-less AR to automatically detect the current working context of an employee and display the required information for the user. Our system is used in workshops for the maintenance of train vehicles. In this use case the mobile application has to detect a large amount of different tools and train components (i.e. objects). We use marker-less detection algorithms, which require the existence of referential images, whose amount grows linear to the number of objects that have to be detected. However, a large amount of referential images makes it impossible to execute the whole detection process on the mobile device. Therefore we implemented a distributed computation of the marker-less detection, which allows to find many different objects. In this paper we describe the components of the system, the image processing technologies used, and current limitations that we have identified.
暂无评论