Abstract: To recognize 2.5D interacting feature effectively,a new feature recognize algorithm was proposed based on graphics and hint.The part model was handled with this algorithm,and the machining feature faces were extracted which could reduce the search space of part graphics but not affect the recognition accuracy.Through establishing the interacting feature split factor queue and using the combined method of automatic and interactive compute,the optimal dependent feather faces were confirmed.According to the feature face,the manufacturing features defined by STandard for the Exchange of Product model data for Numeric Control (STEP-NC) were encoded which could match the feature accurately and effectively by Comparing with sub-graphic and rule matching.The effectiveness and feasibility of proposed algorithm was verified by practical experiment.

Key words: feature recognition, interacting feature, attributed adjacency graph, standard for the exchange of product model data, numerical control

摘要： 为有效识别2.5维相交特征,提出一种新的基于图和痕迹的相交特征识别算法。该算法首先对零件模型进行处理,提取加工特征面,识别时只考虑加工特征面,降低了零件图的搜索空间且不影响特征识别的正确率;通过建立相交特征拆分因子队列,并采用自动计算和交互式相结合的方式确定最优依赖特征面,能够比较准确地分离相交特征;根据特征基面对STEP-NC定义的制造特征进行编码,相对于子图匹配和规则匹配,该方法能够快速和有效地匹配特征。通过具体实例验证了算法的有效性和可行性。

关键词: 特征识别, 相交特征, 属性邻接图, 产品模型数据交换标准, 数控