Comparison of matching algorithms for virtual tour

Electronic tourism is one of the newest types of tourism. The virtual tour is part of the e-tourism, which aims to eliminate temporal and spatial constraints. Panoramic photos are one of the main components for making virtual tours. Nowadays, given the fact that technology has made a lot of progress, there are few panoramic photos from different regions, and usually special people create panoramic photos for the particular usage. On the other hand, given the high quality of photos by advanced technology, these photos have created the potential for producing panoramic photos. Panoramic photos can create by combining multiple images together by matching algorithms. Each of these algorithms has a different accuracy in different situations, and most of these methods are not evaluated for panoramic photos. The purpose of this paper is to investigate various methods and algorithms for categorization of these methods and to find the appropriate method for the preparation of panoramas in tourism. In this regard, how these algorithms function in different situations between images such as transition, rotation and scaling is investigated in order to get the panoramic photos of the same region and choose the best method of evaluation. For this purpose, different matching methods were first investigated. Then, four algorithms were evaluated on panoramic photographs in imposed by different conditions such as fish eye distortion, salt and pepper noise and various scaling. The results show that the SURF algorithm in most situations obtained better results in less time compare to the other three algorithms.