Image Orientation Recognition Based on Content and GPS Information

• Main Authors: Ming-Hung Cho, 卓明弘
• Other Authors: 陳敦裕
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/84480772403074320678

碩士 === 元智大學 === 電機工程學系 === 98 === This thesis mainly aims at discussing the methods of how to Identify landmark or building in images by using low-level image features and GPS information from EXIF tags. The positioning and recogniting system is implemented based on connect to the database management system. After loading Latitude and longitude data from the image, transfer the ellipsoidal coordinates to the TM2 coordinates for the purpose of system operation. The SURF (Speeded-Up Robust Features) algorithm similar to the SIFT (Sale Invariant Feature Transform) algorithm. We extract SURF feature points from object images as a feature database of landmarks. Then, nearest neighbor method is used to find matching features in the loaded image and image database. Total amount of matching features to decide whether or not similarity. Finally, completion of positioning and recogniting a landmark in loaded image.