Evaluating and Proposing of an Image-based Location System Suitable for Ubiquitous GIS

Nowadays image acquisitions devices such as cameras are common between different peoples. Therefore most of peoples can record visual information at any time and place. There fore different ubiquitous services could be developed based on these devices. In this article a vision-based location system has been presented. In this proposed system, each user can access the map of the region by sending an image of a noticeable nearby building in the town. At the first step we need a database consisting of the images of noticeable buildings in the town. In addition to the building images, the location of the building must be saved in the database. Received images should be compared then with the images in the database, using a complex image processing algorithms. In this paper the Scale Invariant Feature Transform algorithm have been used. The most similar image to the received image is detected from the database by comparing the descriptors of key points extracted from image using the algorithm. The location of user is assumed to be the location of identified building in the database. At the last step the map of the region can be send to the user.