摘要
反向工程是利用实物模型测得的数据,构造CAD模型,继而进行分析制造.几何建模是反向工程的关键技术,目前反向工程几何建模存在以下不足之处:(1)通用性差;(2)交互操作多;(3)大自由曲面拟合能力差;(4)系统集成度低.针对以上缺点,提出了反向工程几何建模自动化系统,利用柯西(Kohonen)自组织特征映射人工神经网络进行区域分割;利用具有特殊基函数的人工神经网络实现曲面几何参数的提取,网络权值代表B-样条曲线曲面的控制顶点;
The existing reverse- engineering geometry modelling ( REGM) system has some shortcomings: 1 requiring certain data formats( slice data,or ramge data,or grid data,) 2 too much inter- actions;3 disable to process large data of free- form surface;4 low integration degree.This paper discussed an REGM automatic system,which overcomes the disadvantages of the present reverse- engineering. It intro- duced systematically the applications of artificial neural network in reverse- engineering:1 segmentation by Kohonen self- organizing feature map( SOFM) metwork;2 a new approach called the adaptive basis function network is developed and investigated for the curve and surface fitting process.The control points of B- spline basis functions are defined as the weights of the network;3expressing CAD model by SOFM's topology preserving mapping.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2000年第3期392-395,共4页
Journal of Shanghai Jiaotong University
关键词
反向工程
几何建模
CAD
自动化系统
产品设计
reverse- engineering
geometry modeling
artificial neural networks
feature abstraction
surface approximation
