In this paper, we present an image-based indoor positioning system for digital museum application using omniderectional images to which location information is embedded. By using robust image matching by PCA-SIFT, fast nearest neighbor search algorithm based Locaity Sensitive Hashing (LSH), and the confidence measure which takes into consideration of the location of the objects in the query image, our system can estimate users' locations with high accuracy and in a short time. We conducted some experiments of image matching and arguments about the results.