摘要
为解决航空整体结构件数控加工过程中因换刀频繁而导致刀具出错的问题,提出一种连通域外接矩形的刀具几何参数检测算法,开发基于机器视觉的刀具几何参数在位快速检测系统。采用工业相机、双远心镜头等采集刀具图像,利用MATLAB软件对刀具图像进行灰度变换、二值化等预处理,详细分析连通域标记算法和外接矩形算法的处理过程。结果表明,基于机器视觉的刀具几何参数在位快速检测系统的直径误差小于0.01 mm,悬长误差小于0.08 mm、响应速度快、测量精度高和抗噪性强。
A bounding rectangle algorithm of connected component was investigated for measuring tool geometrical parameters. An in-position detection system for measuring tool geometrical parameters was developed based on machine vision,which were to solve the problem of wrong tool because of tools＇ change frequently in NC machining process of aeronautical monolithic components. The system was carried out by means of using industrial cameras,dual-telephoto lens and others to collect tool image,and using image processing software MATLAB to have image pretreatment including grayscale transformation and binarization. The processes of connected component labeling algorithm and bounding rectangle algorithm were analyzed. The results showed that diameter error was less than 0. 01 mm and hanging length error was less than 0. 08 mm in the in-position detection system for measuring tool geometrical parameters based on machine vision. The system had a high response speed,high measuring accuracy and anti-noise performance.
出处
《山东大学学报（工学版）》
CAS
北大核心
2017年第4期77-82,共6页
Journal of Shandong University（Engineering Science）
基金
数控机床自动验刀系统开发资助项目(司科技(14)-063)
关键词
机器视觉
刀具几何参数
连通域
外接矩形
误差分析
machine vision
tool geometrical parameters
connected component
bounding rectangle
error analysis
