作　　者： (杨隽颖）; (吴国祥）;

机构地区： 嘉兴职业技术学院,浙江嘉兴314036

出　　处： 《激光杂志》 2017年第8期152-155,共4页

摘　　要： 服装设计系统中服装模型特征数量增长迅猛,需要借助特殊方法提取出模型特征为设计者所用。三维激光扫描可从多方面分析服装特征,为此,提出了一种基于三维激光扫描的服装模型特征提取方法,介绍了三维激光扫描建模原理,构建出服装模型。为了避免模型中空值信息对方法性能的影响,对其中的局部特征点进行了归一化处理,并将局部特征点的平均曲率设置成极限值,构造备用特征点集合,利用曲率极值法提取重要特征点。将重要特征点汇集成直线,使用最小生成树处理直线中的独立点,实现了对服装模型中特征线的提取。实验验证结果表明,基于三维激光扫描的服装模型特征提取方法检索效果佳、提取速度快且稳定。 The feature number of clothing special model in clothing design systerm is growing rapidly,designers need a method to extract the model feature. Therefore a feature extraction method of clothing mode based on 3D laser scanning is proposed which introduces the modeling principle of 3D laser scanning and builds the clothing model. In order to avoid the effects of method function caused by null value information in the model,do the normalization of the local feature and set the mean curvature as extremum of the local feature points. Construct the integration of backup feature point and extract important feature points by using curvature extremum value. The important feature points are integrated into straight line,and using the minimum spanning tree to process the individual points of it. Finally realize the characteristic line extraction of clothing model. Experiment results show that this model has good retrieval effect and the extraction is quick and stable.

关 键 词： 三维 激光扫描 服装 模型 特征 提取