基于网格的球面虚拟实景空间搭建方法  

A Mesh-based Method of Building Spherical Virtual Imaging Space

作　　者：叶强强[1] 余烨[1] 张静[1] 刘晓平[1] 

机构地区：[1]合肥工业大学计算机与信息学院,安徽合肥230009

出　　处：《华南师范大学学报（自然科学版）》2012年第4期56-60,共5页Journal of South China Normal University(Natural Science Edition)

基　　金：国家自然科学基金项目(61070124);安徽省自然科学基金项目(11040606Q43)

摘　　要：提出了一种基于网格的球面虚拟实景空间搭建方法.首先针对采集的图像进行畸变校正,获得无畸变的图像;然后结合二维像素坐标和三维空间坐标之间的关系,构造初始模型的球面网格,并利用纹理映射的方法将图像投影到球面上;最后利用网格调整的方法调节图像边缘处场景缺失和场景错位现象,实现球面虚拟实景空间的搭建.与传统球面虚拟实景空间搭建方法相比,本方法只对网格点进行调整,时间代价低且场景具有较高的清晰度.A method is presented to build the spherical virtual imaging space. Firstly, the distortion of images is calibrated to get distortionless images. Secondly, based on the relationship between the pixel coordinates and the three dimensional space coordinates, the spherical meshes of the initial model are constructed, and the texture is mapped onto meshes at the same time. Finally, a mesh tuning algorithm is proposed to solve scene missing and scene dislocation problem. Compared to the traditional methods, this method just deals with points in the meshes, so it costs less time and displays a good description of the scene.

关 键 词：球面虚拟实景空间 网格调整 场景缺失 场景错位 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]