基于蚁群搜索的三维CAD模型相似性计算  

Similarity Calculation of 3D CAD Model Based on Ant Colony Searching

作　　者：高雪瑶[1] 李慧楠[1] 张春祥[2] 于晓洋[3] 

机构地区：[1]哈尔滨理工大学计算机科学与技术学院,黑龙江哈尔滨150080 [2]哈尔滨理工大学软件学院,黑龙江哈尔滨150080 [3]哈尔滨理工大学测控技术与仪器黑龙江省高校重点实验室,黑龙江哈尔滨150080

出　　处：《西南交通大学学报》2017年第2期416-423,共8页Journal of Southwest Jiaotong University

基　　金：国家自然科学基金资助项目(61502124;60903082);中国博士后科学基金资助项目(2014M560249);黑龙江省自然科学基金资助项目(F2015041;F201420)

摘　　要：模型相似性计算是三维CAD模型检索中的关键技术.为了更准确地区分模型之间的差异,提出了一种基于蚁群搜索的模型相似性计算方法.首先,根据边数的差异度量源模型面与目标模型面之间的形状相似程度,并引入面邻接对应关系矩阵衡量两个模型之间的结构相似性;其次,使用蚁群算法搜索源模型与目标模型之间的最优面匹配序列,以最优面匹配序列为基础来计算两个模型之间的相似性;最后,使用贪心算法和本文所提出的方法分别计算源模型与目标模型之间的相似性,并进行对比实验.实验结果表明:在计算关键模型的相似性时,本文所提出方法的计算结果比贪心算法提高了8.33%;与贪心算法相比,本文方法能够有效区分实验中的10个模型.Model similarity computation is a key technology in retrieval of 3D CAD models. In order to accurately distinguish differences between two models, a method for model similarity calculation based on ant colony search was proposed. According to the difference of edge numbers, shape similarity between source model face and target model face was calculated. A matrix which describes the correspondence relationships of face adjacency was introduced to evaluate two models’ structure similarity. At the same time, ant colony algorithm was used to find a sequence of optimal matching faces between source model and target model. Based on this sequence, the similarity of two models was computed. Greedy algorithm and the proposed method were applied to compute the similarity between source model and target model. Experimental results show that computation results of the proposed method are improved 8 . 33% than those of greedy algorithm for the key model’s similarity. Compared with greedy algorithm,the proposed method can distinguish 10 models in experiments effectively.

关 键 词：模型相似性 CAD模型 结构相似性 蚁群算法 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]