建筑物立面精细建模的多视街景影像几何纠正与拼接方法

Approaches of Geometric Correction and Stitching of Multiple St reetview Panorama for Detailed Modeling of Facades

同时搭载差分卫星定位系统、惯性测量单元和街景全景相机的移动测量系统,可高效获取具有精确位姿信息的城市环境高分辨率街景图像,为建筑物立面结构实景三维精细建模提供了重要的数据源。由于地面遮挡、视野限制、影像畸变等因素,单视街景影像难以有效覆盖较大建筑物立面结构。本文提出多视街景影像几何纠正与拼接方法,消除多视影像间的几何、曝光不一致性。利用谷歌开源街景影像数据实验,结果表明该方法可提高建筑物立面重建的纹理质量,同时也具有较高的几何精度。

The mobile mapping system equipped with the differential satellite positioning system,inertial measurement unit and street view panoramic camera can efficiently obtain high-resolution urban street panoramas with precise pose information.It has been a primary data source for 3D refined modeling of facade structure of buildings.Due to the influences of ground occlusion,visual field limitation,image distortion etc,it is difficult to effectively cover the facade structure of large buildings with single-view panorama.This article proposes a geometric correction and stitching method for multi-view panoramas to eliminate the geometric and exposure inconsistencies among multi-view images.Experiment of Google open source street-view panorama shows that this method can improve the texture quality of building facadereconstruction and has high geometric accuracy.