METHOD FOR MATCHING IMAGES USING 3-DIMENSIONAL TRACKER AND COMPUTER VISION BASED ON AUGMENTED REALITY

摘要

PURPOSE: A method for matching images using a 3-dimensional tracker and a computer vision based on augmented reality is provided to extract an accurate position and posture by using a coarse-to-fine type searching method for extracting a feature point for a local computer vision using an error range of a tracker and detailedly correcting a position and posture of a camera. CONSTITUTION: Internal parameters of a camera for capturing a real-photograph image are determined, and a database is constructed for processing an image by a computer vision. The real-photograph image is captured, and position and posture information of the camera is extracted from a tracker. The position and posture of the camera is corrected using a feature point obtained through the computer vision process using the database. A virtual image is synthesized to the real-photograph image using the corrected position and posture value of the camera.