Font Size: a A A

Research On The Restoration Of Ancient Sites By Augmented Reality

Posted on:2017-09-27Degree:MasterType:Thesis
Country:ChinaCandidate:J ZhangFull Text:PDF
GTID:2335330512469380Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
China is a country rich in cultural heritage resources. Due to time lapse, the natural erosion and some man-made factors, these precious heritage was heavily worn or has ceased to exist. So the site protection and development attaches great attention by the people. In the popularity of the computer, people use the technology of computer vision to protect the relics. In recent years, augmented reality technology is a hot research topic in the field of computer engineering. Many researchers apply the augmented reality technology to the protection of monuments, which makes people learn the historical significance of the sites conveniently. This thesis uses the tracking principle of planar scenes for the virtual recovery of the xi'an Wei yang Palace and then adopts the combination of Harris-SIFT algorithm and RANSAC algorithm to calculate the homography and ascertain the position. With augmented reality technology to realize a multi-angle augmented reality system by the virtual-real fusion. This thesis is summarized as follows:1. This thesis analysis the traditional camera calibration algorithm. For the camera distortion, research on the plane lines algorithm, which can calculate the distortion coefficient. Through the experiment analysis, it can demonstrate the effectiveness of the method.2. This thesis adopts the combination of Harris-SIFT algorithm to detect and match the plane interest points, then use the RANSAC algorithm, through experiment analysis, the RANSAC algorithm can purify the matching points and improve the accurate of the matching interest points.3. Designing and realizing the augmented reality system by the planar scene tracking. Detecting interest points of the area firstly and then matching points with this algorithm, select four points of the plane to calculate a homography matrix and estimate the pose, under knowing the camera completely, thus this can avoid setting artificial markers work. Fusing the three-dimensional model on the real scene by OpenGL interface, showing a multi-angle by the virtual-real fusion.
Keywords/Search Tags:Augmented reality, camera calibration, tracking registration, site recovery
PDF Full Text Request
Related items