Augmented Reality

Abstract: Augmented Reality is a technique where the physical world gets enhanced by using a device which can overlay digital information and integrate it with the real environment. Using this technique makes it possible for one to experience things that are not present or real at the moment. Good augmented reality applications requires accurate position- and motion sensing. For the application to be used by a consumer, the device which runs the application needs to be both affordable and portable.

In this thesis, the work of developing an augmented reality application using a smartphone running Android OS is described. The available sensors are examined and tested with respect to their usefulness in location-based augmented reality. The possibility of calculating linear movement using computer vision are also done.

The implementation uses LibGDX game engine for representing the virtual world and rendering the virtual objects. For computer vision, the OpenCV library was used.

Results showed that using the software defined sensor Rotation Vector gave good results when it comes to positioning of the device. For the GPS and determining the device location, it works for some cases but are limited to slow movements. When it comes to image analysis, the device could not deliver information fast enough when having high resolution at the same time.

Conclusions are that a handheld device performs well with augmented reality even though it has some limitations. For future work, using another positioning system or improving the way of analyzing the images and environment could improve the way a virtual object are augmented and positioned.