摘要
提出了一种改进的 BP人工神经网络的图形识别算法 ,以达到加快识别过程和收敛速度的目的 ,可有效地处理断线、模糊图线、字线粘连等多种退化图纸的情况 ,既可获得线的逼近 ,又使识别具有自学习与自适应的智能性。
This paper proposes an accelerating algorithm for engineering drawing symbol recognition on the basis of BP artificial neutral network,in order to speed recognition process and convergence speed,and effectively process degenerate drawings,such as broken line,dim drawing line,word line and adjoin,etc.which can obtain line approximation and implement intelligence of selflearning and selfadaptability.
出处
《电脑开发与应用》
2003年第9期28-29,共2页
Computer Development & Applications
