Abstract:

A CAD model retrieval method based on region feature decomposition was presented herein in order to resolve low retrieval efficiency problem for complex models.First,according to the salient geometric features of the mechanical part,the surface boundary of a solid model was divided into local
convex,concave and planar regions with the minimal number.Then,a kind of region
codes was given to surface region and their links in CAD model.And the similarity between two models was measured by the comparison of their
region codes.Experimental results show that this method is able to support local retrieval and its efficiency meets the requirements of  practical applications.

Key words: region feature, model decomposition, local retrieval, attributed relational graph

摘要：

为弥补现有的CAD模型局部搜索方法应用于结构复杂模型时存在的数据量大和搜索效率低下的缺陷,提出一种基于区域特征分割的CAD模型局部搜索方法。首先依据模型的边界将其分割为一组数量最少的、有一定工程意义的、由一些相互连接的面组成的区域特征集合;接着对分割形成的区域特征及其邻接关系属性进行编码,由属性编码值的相似度度量得到相比较CAD模型的相似度。实验结果表明,该方法能够搜索到相关的CAD模型局部结构,并且搜索效率和精准程度能够满足实际需要。

关键词: 区域特征, 模型分割, 局部搜索, 属性邻接图