• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2022, Vol. 58 ›› Issue (16): 384-390.doi: 10.3901/JME.2022.16.384

• 交叉与前沿 • 上一篇    下一篇

扫码分享

基于CAD模型参数和MHD度量的装配体检索研究

华顺刚1, 谢守广1, 刘斌2, 宋金龙1   

  1. 1. 大连理工大学机械工程学院 大连 116024;
    2. 大连理工大学中日国际信息与软件学院 大连 116620
  • 收稿日期:2021-10-08 修回日期:2022-06-18 出版日期:2022-08-20 发布日期:2022-11-03
  • 通讯作者: 宋金龙(通信作者),男,1987年出生,博士,副研究员,博士研究生导师。主要研究方向为非传统加工技术及CAD。E-mail:songjinlong@dlut.edu.cn
  • 作者简介:华顺刚,男,1964年出生,博士,教授,博士研究生导师。主要研究方向为CAD与图形学、图形图像处理。E-mail:hsgang02@dlut.edu.cn;
    谢守广,男,1995年出生,硕士研究生。主要研究方向为CAD与图形学。E-mail:xsg_1001@mail.dlut.edu.cn;
    刘斌,男,1981年出生,博士,教授,博士研究生导师。主要研究方向为3D可视化与医学图像处理。E-mail:liubin@dlut.edu.cn
  • 基金资助:
    国家自然科学基金资助项目(61972440)

Study on Assembly Retrieval Based on CAD Model Parameters and MHD Metric

HUA Shungang1, XIE Shouguang1, LIU Bin2, SONG Jinlong1   

  1. 1. School of Mechanical Engineering, Dalian University of Technology, Dalian 116024;
    2. DUT-RU Inter. School of Information Science & Engineering, Dalian University of Technology, Dalian 116620
  • Received:2021-10-08 Revised:2022-06-18 Online:2022-08-20 Published:2022-11-03

摘要: 标准交换格式的装配体CAD模型仅包含几何形状信息而缺乏装配约束信息,为适应该类模型的检索及重用,提出一种结合CAD模型参数和改进的豪斯多夫距离的装配体检索方法。从CAD系统中直接获取零件模型的几何参数和属性数据,经归一化处理后形成矢量来表示该零件。将装配体中各零件的参数矢量合成为一个矢量的集合,作为装配体的描述符。然后采用改进的豪斯多夫距离度量算法,通过多对多的零件匹配来计算两个装配体之间的非相似度。该方法支持装配体模型的全局检索和局部检索。试验结果表明,该方法快速高效,适用于多种文件格式的CAD模型,并能获得满意的装配体检索效果。

关键词: 装配体检索, CAD模型参数, 改进的豪斯多夫距离, 零件匹配, 非相似度

Abstract: Assembly CAD model in standard exchange format only contains geometric shape information but lacks assembly constraint information. In order to adapt to retrieval and reuse of such model, a method of assembly retrieval is proposed by combining CAD model parameters and the Modified Hausdorff Distance. First, the geometric parameters and attribute data of the part model are directly obtained from the CAD system, and the vector is formed to represent the part after normalization processing. The parameter vectors of each part in the assembly are synthesized into a set of vectors as the descriptor of the assembly. Then, the Modified Hausdorff Distance algorithm is used to measure the dissimilarity between two assemblies through many-to-many parts matching. This method supports global retrieval and local retrieval of assembly model. The experimental results show that the proposed method is fast and efficient, suitable for CAD models of multiple file formats, and can obtain satisfactory assembly retrieval results.

Key words: assembly retrieval, CAD model parameters, modified Hausdorff distance, part matching, dissimilarity

中图分类号: