基于自组织特征映射区域分割的三维几何模型简化算法  

3D geometric model simplification based on region partition using self-organizing feature map

作　　者：王蕊[1] 李俊山[1] 刘玲霞 胡双演[1] 罗蓉[1] 

机构地区：[1]西安第二炮兵工程学院计算机与指挥自动化系,西安710025 [2]二炮装备部计量站,北京100085

出　　处：《大连海事大学学报》2008年第2期100-104,共5页Journal of Dalian Maritime University

基　　金：国家自然科学基金资助项目(60772151);总参装备维修项目

摘　　要：为满足模型简化后保留细节特征的需要,引入自组织特征映射(SOFM)神经网络,提出一种基于区域分割的三维几何模型简化算法:将三维几何模型划分成具有不同特征的区域,在此基础上进行多区域并行简化,利用顶点微调法对简化后的模型进行局部特征修正.结果表明,该方法可在提高模型简化速度的同时,有效保留模型的细节特征,显著改善模型因简化而产生的形变.Simplification algorithm of 3D geometric model based on region partition was developed to meet the requirements of detail-preserving after the model simplified. Self-organizing feature map （SOFM） neural network was introduced, and 3D geometric model was divided into regions with different characteristics, and then multi-region collateral simplification was executed, vertex micro-adjusting algorithm was used to modify the local detail of simplified model. Results show that the algorithm can enhance simplification speed, while preserving the minutiae feature of the model effectively and improving model distortion caused by simplifying model obviously.

关 键 词：三维几何模型 简化算法 自组织特征映射 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]