Development of a PC-based markerless augmented reality

Abstract

AR applications are divided into two major types; marker-based and markerless AR. Some applications of AR such as in education and training are still being implemented on PC where the use of mobile and wearable devices are not possible. This paper presents the development of a PC-based markerless AR application. Four main components in markerless AR development; detector, descriptor, matcher and pose estimation are employed. This application used FAST as the detector, FREAK as the descriptor while BruteForce Matcher is used as the matcher algorithm. The reference image used is Graffiti from Mikolajczyk's dataset. The tracking process is tested for its robustness to rotation, scale, brightness and also blur changes. The application successfully rendered a 3D object on top of the reference image during the robustness test.