数字图像刀具测量机

【作者】 侯学智；

【导师】 杨平；

【作者基本信息】 电子科技大学 ， 机械电子工程， 2004， 硕士

【摘要】 在数控加工中心中,为了减少非生产时间,提高加工效率,有必要研制高精度、高效率的刀具测量系统。本文通过对各种刀具测量技术的比较和分析,提出基于图像处理技术的数字图像刀具测量机方案。数字图像刀具测量机由机械运动平台、光学照明装置、CCD成像单元、数字图像处理软件模块和硬件接口模块组成。系统的工作原理是采用CCD相机摄取刀具图像,利用USB接口将图像传输到计算机中,并由数字图像处理软件完成刀具图像检测,最后刀具的几何参数通过RS-232接口输出。 本文重点讨论了刀具图像二值化、边缘检测、轮廓特征点提取、亚像素边缘定位等图像处理算法。采用最大类间方差阈值法将灰度图像分割为二值图像,比较了图像形态学与经典微分算子在边缘检测中的特点。提出一种新的轮廓特征点提取方法,它利用最小二乘法拟合轮廓曲线,结合斜率、曲率信息提取特征点。详述了各种亚像素边缘定位方法,采用Zernike 正交矩法、B样条拟合法以及灰度重心法,实现刀具几何参数(长度、角度、半径)的亚像素测量。 本文论述了系统的软硬件实现,分析了测量中各种可能的误差来源和理论测量精度,给出了实际测量结果。试验表明,测量系统可快速、精确、非接触地测量刀具尺寸,满足数控加工的需要。更多还原

【Abstract】 It is necessary to develop a cutting tools measurement system with high precision and high efficiency in the NC(numerical control) machining center, so that it can decrease non-production time and improve machining efficiency. In this paper all kinds of cutting tools measurement technology are compared and analyzed before the project of a digital-image cutting tools measurement machine based on digital image processing technology is presented.The digital-image cutting tools measurement machine is made of two-dimension movement platform、optical lighting equipment、CCD imaging unit、digital image processing software module and hardware interface module. The working principle of this system is that an image shooted by CCD camera is transported into the computer through USB interface, then the digital image processing software can measure it precisely, finally the geometrical parameters of cutting tools is outputted through RS-232 interface.Several image processing algorithms are mainly discussed, including cutting tools image binarization、edge detection、contour feature-points extraction、subpixel edge location. Intensity image is segmented to Bi-value image through adopting a method based on maximum between-cluster variance and image morphology compares with classical differential operators in edge detection. A new feature-points extraction algorithm is proposed, which uses least square method fitting contour curve and combines slope、curvature information to extract feature-points. By discussing many subpixel edge location methods in detail, ZOM(zernike orthogonal moments) method、B-spline fitting method and gray center method are adopted and finally the geometrical parameters(lengths、angles、radiuses) of cutting tools are measured in subpixel level.The system software and hardware realization are discussed. All types of possible error sources are analyzed and the actual measurement results are presented. Experiments show the system is capable of realizing quick、accurate、non-contact measurement of cutting tools size and can meet the NC machining need.更多还原