首页 | 官方网站   微博 | 高级检索  
     

一种适用于规则形体的三维重建方法
引用本文:王佳生,唐好选,仝建国,苏小红.一种适用于规则形体的三维重建方法[J].哈尔滨工业大学学报,2009,41(3):152-156.
作者姓名:王佳生  唐好选  仝建国  苏小红
作者单位:王佳生,WANG Jia-sheng(哈尔滨工程大学,计算机学院,哈尔滨,150001);唐好选,仝建国,苏小红,TANG Hao-xuan,TONG Jian-guo,SU Xiao-hong(哈尔滨工业大学,计算机学院,哈尔滨,150001)  
摘    要:给出了一种适用于规则形体的三维重建方法,从待建模对象的若干不同角度的照片出发,先通过特征提取和交叉匹配获得足够多的特征点,并以协方差作为相似性度量依据,采取分布匹配策略剔除错误匹配对,然后通过空间点重建和表面重建法重建出目标对象的三维模型,并从采集的图像中提取真实纹理,对图像进行二维三角化,三角化的二维点列再映射到三维空间,为三维模型添加纹理,得到具有真实感的外观模型.实际应用验证了所提方法的正确性和有效性.

关 键 词:三维重建  特征提取  特征匹配  规则形体

A 3D reconstruction method adapted to regular object
WANG Jia-sheng,TANG Hao-xuan,TONG Jian-guo,SU Xiao-hong.A 3D reconstruction method adapted to regular object[J].Journal of Harbin Institute of Technology,2009,41(3):152-156.
Authors:WANG Jia-sheng  TANG Hao-xuan  TONG Jian-guo  SU Xiao-hong
Affiliation:1.School of Computer Science and Technology,Harbin Engineering University,Harbin 150001,China;2.School of Computer Science and Technology,Harbin Institute of Technology,Harbin 150001,China)
Abstract:A 3D reconstruction method adapted to regular objects was proposed.Firstly,sufficient feature points were obtained through feature extracting and cross matching process of some photos at different angles taken from an object.In order to delete some mis-matched points generated in matching process,a distribution matching strategy was used based on similarity measurement of covariance.Then the space point reconstruction and surface reconstruction method were used to reconstruct a 3D model of the object.At the same time,the realistic texture was extracted from some images;the 2D images were triangularized and then mapped into 3D space.Finally,the texture was added to the 3D model and a realistic exterior model was obtained.The correctness and validity of this modeling method was confirmed through practical application.
Keywords:image-based 3D reconstruction  feature extraction  feature matching  regular object
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号