论文部分内容阅读
提出了一种基于人工神经元网络对机械工程图纸标注符号识别的方法，并对基于特征输入和基于点阵输入两种神经网络分类器的特点进行了比较．研究结果表明，在模糊特征下，神经网络方法对工程图纸标注符号的智能识别完全能达到实用化要求．
A method based on artificial neural network is presented to mark the symbol of mechanical engineering drawings. The characteristics of two neural network classifiers based on feature input and dot matrix input are compared. The research results show that under the fuzzy feature, the neural network method can fully meet the practical requirements for the intelligent identification of the drawing symbols of engineering drawings.