Object Recognition Based on ORB Descriptor for Markerless Augmented Reality

This paper investigate a binary local image descriptor for Augmented Reality (AR) applications. Recently, various fields are benefit from AR. This technique can enhance the real environment by inserting virtual objects generated by computer. Temporal coherence between virtual and real objects must be ensure in AR system realization. In this paper, object recognition based on extracted natural features is presented. Experimental results indicate better performance of object recognition using ORB(Oriented FAST and Rotated BRIEF) descriptor compared to the SURF(Speed Up Robust Features) descriptor in AR applications. Keywords—Augmented reality; object recognition; local binary descriptor; feauture matching; ORB; SURF.