Please wait a minute...
工程设计学报
Chinese Journal of Engineering Design  2021, Vol. 28 Issue (4): 407-414    DOI: 10.3785/j.issn.1006-754X.2021.00.065
Design Theory and Method     
Three-dimensional model classification and retrieval algorithm based on polar radius surface moment and HMM
WANG Hong-shen1, LIU Min1, QIANG Hui-ying2
1.School of Mechanical and Electrical Engineering， Lanzhou University of Technology， Lanzhou 730050， China
2.School of Mathematics and Software， Lanzhou Jiaotong University， Lanzhou 730070， China
Download: HTML     PDF(1815KB)
Export: BibTeX | EndNote (RIS)      

Abstract  In the computer aided design (CAD), the classification and retrieval of three-dimensional models of mechanical parts facilitates designers to reuse design information, so as to shorten the product development cycle and quickly respond to market demands. For the classification and retrieval of three-dimensional models, a classification and retrieval algorithm based on the polar radius surface moment and hidden Markov model (HMM) was proposed. Firstly, the polar radius surface moments of the three-dimensional model were calculated and the feature vector was formed. After sorting and coding, the feature vector was used as the observation sequence of the HMM. Then, a part of manually labeled three-dimensional models were taken as the training samples, and the HMM was trained by the multi-observation sequence B-W (Baum-Welch) algorithm with scale factors. Finally, the trained HMM was used to classify and retrieve three-dimensional models. Experimental results showed that， compared with the two existing classification and retrieval algorithms, the proposed algorithm had higher recognition rate and retrieval efficiency. The characteristics of the algorithm were as following: the calculation of polar radius surface moment was fast and the voxelization of three-dimensional model was not needed; the HMM had fast training and strong classification ability, and it had certain classification ability without a large number of training samples. The research shows that the proposed algorithm could better solve the problem of classification and retrieval of three-dimensional models, which had certain practical value.

Received: 03 July 2020      Published: 28 August 2021
CLC:  TP 391.7  
Service
E-mail this article
Add to my bookshelf
Add to citation manager
E-mail Alert
RSS
Articles by authors
Cite this article:

WANG Hong-shen, LIU Min, QIANG Hui-ying. Three-dimensional model classification and retrieval algorithm based on polar radius surface moment and HMM. Chinese Journal of Engineering Design, 2021, 28(4): 407-414.

URL:

https://www.zjujournals.com/gcsjxb/10.3785/j.issn.1006-754X.2021.00.065     OR     https://www.zjujournals.com/gcsjxb/Y2021/V28/I4/407


基于极半径曲面矩和HMM的三维模型分类与检索算法

在计算机辅助设计（computer aided design，CAD）中，对机械零件的三维模型进行分类和检索有利于设计人员重用设计信息，从而缩短产品的开发周期，以快速响应市场需求。针对三维模型的分类与检索，提出了一种基于极半径曲面矩和隐马尔科夫模型（hidden Markov model，HMM）的分类与检索算法。首先，计算三维模型的极半径曲面矩并组成特征向量，经排序编码后，将其作为HMM的观测序列；然后，取一部分人工标注过的三维模型作为训练样本，采用添加比例因子的多观测序列B-W（Baum-Welch）算法对HMM进行训练；最后，利用训练好的HMM对三维模型进行分类与检索。实验结果显示，与现有的2种分类与检索算法相比，所提出的算法具有更高的识别率和检索效率。该算法的特点是：极半径曲面矩计算快，不用将三维模型体素化；HMM训练快，分类能力强，且不需要大量训练样本就有一定的分类能力。研究表明，所提出的算法能较好地解决三维模型的分类与检索问题，具有一定的实用价值。
[1] JIA Qiang, GAO Yue-Fei, TONG Yan-Fa, ZHAO Peng. Design of linear eddy current buffer[J]. Chinese Journal of Engineering Design, 2011, 18(3): 209-213.
[2] XIAO Deng-Hong, WANG Jun-Yuan, ZENG Zhi-Qiang, DU Wen-Hua. Research on dynamic characteristics of an explosion-proof rubber-tyred vehicle[J]. Chinese Journal of Engineering Design, 2011, 18(1): 23-27.
[3] WANG Chun-Xiang, LI Le, WANG Jian-Guo. Geometric reverse resolution and model reconstruction of excavator dipper teeth based on feature extraction[J]. Chinese Journal of Engineering Design, 2010, 17(3): 211-214.
[4] XIAO Wei-Yue, YANG Ji-Rong, CAI Yue-Hua. Decision-making method of part surface processing plan under
manufacturing constraints based on fuzzy Petri net[J]. Chinese Journal of Engineering Design, 2010, 17(2): 86-90.