基于图的混合加工特征识别方法  

Hybrid Recognition of Machining Features Based on Graph

作　　者：李大磊[1] 陈广飞[1] 尹跃峰[1] 

机构地区：[1]郑州大学机械工程学院,郑州450001

出　　处：《组合机床与自动化加工技术》2013年第6期81-83,共3页Modular Machine Tool & Automatic Manufacturing Technique

摘　　要：针对传统的基于图的特征识别方法难以识别相交特征和拓扑不固定特征的问题,提出了一种基于图的混合加工特征识别方法。该方法首先利用插入分割线分割贴合的面的方法拆分相交特征,然后重构扩展属性邻接图,从中分解出最小条件子图,最后根据二次曲面特征、由平面组成的特征的边界模式,分别建立相应的特征识别知识库,并应用知识树通过推理识别特征。验证结果表明该方法能合理地拆分相交特征,有效地识别常见的加工特征。Since traditional feature recognition method based on graph is difficult to recognize intersecting features and features with variable topology, this paper presents a method of hybrid feature recognition based on graph. Firstly, merged faces are split by creating split lines in order to separate intersecting feature. Secondly, the method establishes the extended attribute adjacency graph again, which is decom- posed into several minimal condition sub-graphs （MCSG）. Finally, according to boundary pattern of quadric features and features consist of planes, constructs separately feature recognition knowledge bases, and recognizes features by using the knowledge tree and reasoning. The results show the method can separate reasonably intersecting feature and recognize effectively machining features.

关 键 词：特征识别 相交特征 最小条件子图 知识树 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]