摘要
研究用感知器和三层BP神经网络识别二维零件图的形状特征信息。提出了零件图扩展属性矩阵的概念及求法，对AutoCAD R14进行了二次开发，完善了笔者开发的二维回转类零件图特征识别和提取系统。
The paper gives a new approach to solve the problem of features recognition and extraction from 2D part drawing by means of perceptron and three layers BP neural network. We bring forward a concept and arithmetic of part drawing expand propertiy matrix. We development a system of features recognition and extraction from 2D revolution part drawing by means of developing AutoCAD R14 applications.
出处
《工程图学学报》
CSCD
2001年第1期13-18,共6页
Journal of Engineering Graphics
基金
中国工程物理研究院院外基金资助项目!(96060629)
