论文部分内容阅读
本文介绍一个最近由我们研制成功的电子电路图自动输入系统的构成与特点。利用本文提出的内、外边缘快速判决法及在“非环形”符号各端点加“黑点”的辅助措施提高了模式分离的可靠性。为了最大限度地利用原始信息，我们采用直接从光栅图中抽取特征的办法去识别字符与符号。在手写符号的识别中，我们提出了四边带投影的特征抽取方法并把它了引入到“决策树”的判决过程中，使识别率与识别速度都得到提高。当系统对整图的拓扑关系进行理解之后，还实现了与当前国际流行的ＣＡＤ软件ＯＲＣＡＤ的链接、使系统能以自动输入的方式支持自动布线与电路模拟。
This article describes a recently developed by our successful electronic circuit diagram of the automatic input system components and characteristics. The reliability of pattern separation is improved by using the fast inner-outer edge decision method proposed in this paper and the addition of “black dots” at each end of the “non-circular” symbol. To get the most out of the original information, we use characters extracted directly from raster images to identify characters and symbols. In the recognition of handwritten symbols, we propose the feature extraction method of four-sideband projection and introduce it into the decision tree of decision tree, which improves the recognition rate and recognition speed. After the system understands the topological relationship of the whole graph, it also links with the ORCAD, a CAD software popular in the world, so that the system can support automatic wiring and circuit simulation with automatic input.