测绘学报 ›› 2020, Vol. 49 ›› Issue (2): 225-234.doi: 10.11947/j.AGCS.2020.20180588

• 摄影测量学与遥感 • 上一篇    下一篇

局部表面参数化的实景三角网模型语义增强方法

汪利斌1, 胡翰2,3, 朱庆1, 丁雨淋1, 陈敏1   

  1. 1. 西南交通大学地球科学与环境工程学院, 四川 成都 611756;
    2. 轨道交通工程信息化国家重点实验室(中铁一院), 陕西 西安 710043;
    3. 浙江中海达空间信息技术有限公司, 浙江 湖州 313200
  • 收稿日期:2018-12-21 修回日期:2019-06-10 发布日期:2020-03-03
  • 通讯作者: 胡翰 E-mail:huhan19880715@163.com
  • 作者简介:汪利斌(1995-),男,博士生,研究方向为倾斜摄影测量。E-mail:2325748989@qq.com
  • 基金资助:
    国家自然科学基金(41631174;61602392;41871291);广东省科技计划(2017B010117006)

A semantic enhancement method for photorealistic mesh model based on local parameterization

WANG Libin1, HU Han2,3, ZHU Qing1, DING Yulin1, CHEN Min1   

  1. 1. Faculty of Geosciences and Environmental Engineering, Southwest Jiaotong University, Chengdu 611756, China;
    2. State Key Laboratory of Rail Transit Engineering Informatization(FSDI), Xi'an 710043, China;
    3. Zhejiang Hi-Target Spatial Information Technology Co. Ltd., Huzhou 313200, China
  • Received:2018-12-21 Revised:2019-06-10 Published:2020-03-03
  • Supported by:
    The National Natural Science Foundation of China (Nos. 41631174;61602392;41871291);The Science and Technology Planning Project of Guangdong Province of China (No. 2017B010117006)

摘要: 倾斜摄影测量方法已可自动获取城市规模的实景三角网模型,然而散乱的三角网缺乏精细的几何结构和功能语义信息。为克服上述问题,提出一种局部表面参数化的实景三角网模型语义增强方法:将具有语义信息的独立三维部件与实景三角网模型的无缝融合问题,通过定义三维表面结构树,转换为局部区域的三角网替换操作;在待融合区域附近,将原实景三角网模型和替换的三维语义部件,通过局部参数化表达,UV展开为二维平面三角网;在二维平面上构建约束Delaunay三角网(CDT),实现两模型的无缝拼接,逆映射至三维空间并自动重建语义部件。通过深圳某区域的倾斜影像进行的试验证明,本文方法能有效实现具有开放边界和语义信息的部件模型与表面模型的无缝融合。与商业软件Maya对比,这种基于插入、融合的手段对提高建模效率具有实用价值。

关键词: 三维表面结构树, 局部表面参数化, 三角网模型, 语义增强, 倾斜摄影测量

Abstract: With the advances in structure-from-motion and multi-view stereo, state-of-the-art oblique photogrammetric solutions can obtain city-scale photorealistic mesh models automatically. However, the mesh models are lack of fine geometric structure and semantic free. Aiming at solving this issue, it is proposed that a semantic enhancement method for photorealistic mesh models based on local surface parametrization. The basic idea behind the proposed method is that, through the representation of surface tree, it is converted that the seamless fusion of semantic components and photogrammetric mesh models to a replacing operation in a local area. The two 3D models are parametrized to 2D space in the local region and seamless merged and replaced in the UV space by 2D constrained delaunay triangulation (CDT). The replaced semantic components are then reversed transform to 3D space, which finalize the semantic enhancement automatically. Experiments on oblique images in Shenzhen reveal that the proposed method can effectively realize the automatic seamless fusion of semantic components with an open boundary and photorealistic mesh models. Compared with the commercial software Maya, based on the method of insertion and fusion, the proposed method has practical value for improving modeling efficiency.

Key words: 3D surface structure tree, local parameterization, triangulation model, semantic enhancement, oblique photogrammetry

中图分类号: