摘要
在充分分析电子元器件图像特征及物理特性的基础上,进行元器件自动分类的算法研究与系统开发。本文将最小二乘法和Hough变换相结合,得到一种改进的直线段精确检测算法,准确地提取了元件边缘直线;而设计的快速圆检测算法,对电感(蜂鸣器)这一特定元件具有显著的效果;提出了新的管脚数检测算法,较好解决了元件管脚密度高、噪声敏感的问题。实验证明,本文的算法快速有效易于实时处理,且系统平均识别率达到90%,有良好的容错功能,能够满足用户需求。
This paper research on the algorithmic of components auto-classification and system development based on analyzing the graph and physics characters of electronic components. A accuracy detection algorithmic for the straight line is been developed by combining the least square method and Hough Switch, which can pick up exactly the borderline of components. We also designed a rapid circle detection algorithmic which can detect the buzzer directly, and a new detection algorithmic of pin number for solving the problem of high density pin and noise sensitivity. It is proved by many experiments , the algorithmic have so many advantage as followed： real-time processing, the average discrimination of 90% and fault tolerance, etc. it can satisfy the needs of users.
出处
《机电产品开发与创新》
2008年第6期133-135,共3页
Development & Innovation of Machinery & Electrical Products
基金
北京市科技新星计划(2005B45)
国家科技基础条件平台建设(2005DKA32900)
