测绘通报 ›› 2020, Vol. 0 ›› Issue (8): 28-32,38.doi: 10.13474/j.cnki.11-2246.2020.0243

• 学术研究 • 上一篇    下一篇

多源点云数据融合的建筑物精细化建模

王树臻1, 郑国强1, 王光生1, 胡玉民2, 张德怀2, 亓伟3   

  1. 1. 山东建筑大学测绘地理信息学院, 山东 济南 250102;
    2. 山东智维测绘科技有限公司, 山东 济南 250014;
    3. 山东正维勘察测绘有限公司, 山东 济南 250100
  • 收稿日期:2019-12-31 出版日期:2020-08-25 发布日期:2020-09-01
  • 通讯作者: 郑国强。E-mail:gqzheng@sdjzu.edu.cn E-mail:gqzheng@sdjzu.edu.cn
  • 作者简介:王树臻(1994-),男,硕士生,研究方向为倾斜摄影测量与遥感。E-mail:1095751340@qq.com
  • 基金资助:
    国家自然科学基金青年基金(51808320);山东省研究生创新计划(SDYY16031)

Building fine modeling based on multi-source point cloud data fusion

WANG Shuzhen1, ZHENG Guoqiang1, WANG Guangsheng1, HU Yumin2, ZHANG Dehuai2, QI Wei3   

  1. 1. School of Surveying and Geo-informatics, Shandong Jianzhu University, Jinan 250102, China;
    2. Shandong Geovey Surveying Technology Co., Ltd., Jinan 250014, China;
    3. Shandong Zhengwei Survey and Mapping Co., Ltd., Jinan 250100, China
  • Received:2019-12-31 Online:2020-08-25 Published:2020-09-01

摘要: 三维建模技术能够实现建筑物的数字化存档,在古建筑保护与修复和现代建筑规划与改造中具有不可替代的作用。针对倾斜摄影测量和三维激光扫描建模技术中建筑物模型存在的问题,本文提出了一种倾斜摄影测量和三维激光扫描生成三维点云模型相融合的建筑物精细化建模方法。选用无人机和三维激光扫描仪作为试验设备,利用ContextCapture、SCENE软件完成点云拼接、生产和编辑,通过ICP算法完成点云精细匹配,实现多源点云数据融合建模;对比单一建模方法模型,从纹理结构和模型精度两方面对融合建模模型进行质量评价。结果表明,融合建模模型纹理清晰,几何结构完整,模型距离中误差和高差中误差的均值均低于倾斜摄影测量模型的值,接近三维激光扫描模型。

关键词: 倾斜摄影测量, 三维激光扫描, ICP算法, 点云匹配, 融合建模

Abstract: The 3D modeling technology can realize the digital archiving of buildings and play an irreplaceable role in the protection and restoration of ancient buildings, the planning and reconstruction of modern buildings. Aiming at the problems of building models in oblique photogrammetry and 3D laser scanning modeling, this paper proposes a fine modeling method of building based on the fusion of obligue photogrammetry and 3D point cloud model generated by 3D laser scanning. Uav and 3D laser scanner are selected as experimental equipment. Point cloud stitching, production and editing are completed by ContextCapture and SCENE software. Point cloud fine matching is completed by ICP algorithm. And multi-source point cloud data fusion modeling is realized. Compared with the single modeling method model, the quality of the fusion modeling model is evaluated from the two aspects of texture structure and model accuracy. The results show that the fusion model has clear texture and intact geometric structure, and the mean error of the median distance and the mean error of the height difference of the model are both lower than that of the oblique photogrammetry model, which is close to the 3D laser scanning model.

Key words: oblique photography, 3D laser scanning, the ICP algorithm, point cloud matching, fusion modeling

中图分类号: