摘要
基于影像的三维立体模型重建时,如何自动地将部分模型拼接成完整模型是个亟待解决的问题。目前关于三维立体模型拼接大多限制在手动取两个模型上的公共点完成拼接。提出一种方法,能够全自动地、快速地、高精度地实现三维立体模型的拼接。与图像拼接类似,三维立体模型拼接也要经过特征点提取、特征点匹配、空间变换三个步骤。值得一提的是,在特征点提取时利用了三维重建时的点云数据、特征点匹配时利用了点与点间距客观不变、空间转换参数RT矩阵计算时提出了仅用四对精匹配点计算等方法,完成三维立体模型全自动、快速、高精度地拼接,拼接精度达到毫米(mm)级。
When it comes to 3D Model rebuilding, which is based on the images, it＇s still a urgent problem to be solved that how to mosaic incomplete 3D models to complete one. Now, a method was presented, which is people select the common points on the incomplete 3D models, then match them and accomplish mosaic. Here will show you how to accomplish mosaic automatically, fast, accurately. The accuracy can be mm level. As similar with images mosaic, 3D models mosaic need to pick up characters, then match them, and then calculate the parameters. What is must to say is that making use of 3D cloud data when picking up characters, taking advantage of the distance of two points when matching the characters, only using 4 pairs of high precision matched characters to calculate the parameters.
出处
《科学技术与工程》
北大核心
2015年第12期75-80,共6页
Science Technology and Engineering
基金
国家“973计划”项目(2012CB725301)
测绘地理信息公益性行业科研专项项目(201412015)资助
