Synthesizing the Whole View of a Large-Scale Object

A robust method for panorama view reconstruction from multiple viewpoints of the large-scale object is presented. The desired object must be acquired from multiple viewpoints. The information of the feature points of an object is obtained by establishing correspondence between views. We present a robust method based on image mosaicing approach and panorama synthesizing. First, the feature of each object is extracted. Then the corresponding feature points of an object between the images are estimated. The different of minimum threshold value between two images is computed. After that, the full-view of an object is reconstructed by integrating two images based on image mosaicing approach.This research describes how to establish feature correspondences between images accurately and effectively. Image registration technique provides an initial estimation for establishing feature correspondences of point features. The linear solution with the reliable correspondences makes the computation of the geometric transformation between two images, called homography.