A visual-GPS fusion based outdoor augmented reality method

Virtual objects can be overlaid with real scenes through outdoor augmented reality technology, which bring about prominent experience. The existing outdoor augmented reality methods are usually limited in accuracy and scalability. To solve this problem, a novel method combining computer vision and Global Positioning System is proposed in this paper. The Geohash method is introduced to stimulate the retrieval of nearby locations. The vocabulary tree is built to recognize the current scene from the reference library. Faster R-CNN based object detection method is combined with AKAZE feature detection and image matching algorithm to realize the scene recognition and target tracking. The results show that our method can realize efficient and scalable outdoor augmented reality.