iNavigation: an image based indoor navigation system
MetadataShow full metadata
Image-based navigation techniques attract much attention as a powerful alternative to traditional map-based navigation. The Google Street View, which gets a mixed reception is one such technique featured in Google Map that allows users to navigate through large scale outdoor environment with 360 degree imagery. However, the Google Street View can not provide timely updates because it requires immense data, and this technique involving a panoramic camera has not been extended to indoor environment due to its data collection approach. This thesis proposes a convenient and efficient image-based indoor navigation web application designed for mobile phones. It utilizes ordinary photographs and navigation information, rather than the panoramic view used in Google Street View, to guide people through the building. Our system implements SIFT feature detection and ANN search to provide timely positioning service. It also can roughly locate a query image on the map by using IPM (Inverse Perspective Mapping), if the query image is taken following certain rules. Thus it enables interactive navigation and knowledge sharing among users.