基于数字图像的机械加工表面质量检测技术研究

发布时间：2018-04-22 17:35

本文选题：工件表面质量 + 数字图像　；参考：《兰州理工大学》2017年硕士论文

【摘要】：表面粗糙度是指加工零部件表面上存在的,由较小的间距和峰谷所构成的微观几何形状特征。随着产品精度和零部件表面粗糙度要求的不断提高,接触式检测工件表面粗糙度的仪器仪表的测量精度和测量速度等方面要求的不断提高,推动了现在的非接触式无损检测技术的发展。本文将数字图像的相关理论与技术应用于工件表面粗糙度的检测研究,主要研究内容和结果包括:(1)国内外关于检测工件表面粗糙度的发展状况进行了表述,并且对数字图像处理技术在零部件表面粗糙度检测中的应用以及意义进行探讨,而且介绍了数字图像处理的基本方法以及原理并且运用其对工件表面的图像进行分析处理,用不同的数字图像的处理方法,针对相同的图片,从而获得不能的图像特性;(2)将灰度变换、灰度拉伸等方法应用到零部件加工表面的图像中分析,经过图像处理后,根据其灰度直方图及图像显示效果选择适合表面粗糙度检测的图像处理方法,研究了零部件的表面粗糙度Ra、Rz值的测量原理及其实现的数字化技术;(3)介绍了图像噪声的分类,并且根据零部件表面图像中噪声的不同,运用不同的滤波方法来处理图像中不同类型的噪声,并且对其进行了降噪处理,最终获得不同的降噪效果;(4)以工件加工表面的图片为对象,采用传统的表面粗糙度测量方法,对特征纹理的提取方法进行研究;采用最小二乘法,通过对工件加工表面图中的特征纹理的分布情况进行分析、对加工工件的纹理特征的统计,计算得出零部件表面粗糙度值,并给出零部件表面粗糙度值的最终测量方法;(5)基于Open CV平台,借助C++语言编出测量加工工件表面粗糙度的程序,实现了对工件加工表面图片的处理,从图像的读入到表面的纹理特征提取,再到表面粗糙度值的自动测量,最后将测得数据与表面粗糙度试验得到的实际值相比较,将两组数据进行参数回归分析,得到最终的运算结果,也就实现了本系统的标定。本文将工件表面粗糙度检测与数字图像处理技术有机结合起来,对加工工件的表面图片进行预处理,在此基础上,提取了图像中的纹理特征并编程计算,最后得到的工件表面的粗糙度值,从而证明了此方法的可行性,研究结果对工件表面粗糙度的检测具有非常重要的实践意义。
[Abstract]:Surface roughness refers to the micro-geometric characteristics of machining parts which are composed of small spacing and peak and valley. With the continuous improvement of product precision and surface roughness requirements of parts and components, the measurement accuracy and measuring speed of the instruments and instruments used to detect the surface roughness of workpieces have been continuously improved. It promotes the development of non-contact non-destructive testing technology. In this paper, the theory and technology of digital image are applied to the measurement of workpiece surface roughness. The main research contents and results include the development of measuring workpiece surface roughness both at home and abroad. The paper also discusses the application and significance of digital image processing technology in measuring the surface roughness of parts, and introduces the basic method and principle of digital image processing, and uses it to analyze and process the image of workpiece surface. Using different digital image processing methods, aiming at the same picture, we can obtain the image characteristics that can not be obtained. (2) the methods of gray level transformation and gray scale stretching are applied to the analysis of the image of the machined surface of the parts and components. After image processing, the image processing is carried out. According to the grayscale histogram and image display effect, the image processing method suitable for surface roughness detection is selected. The measuring principle of surface roughness Rz value of parts and components and its digitization technology are studied. The classification of image noise is introduced. According to the different noise in the surface image of parts, different filtering methods are used to deal with the different types of noise in the image, and the noise reduction is carried out. Finally, different noise reduction effects are obtained. (1) taking the images of workpiece machined surface as the object, the traditional surface roughness measurement method is used to study the feature texture extraction method, and the least square method is used. By analyzing the distribution of the feature texture in the workpiece machining surface map and counting the texture features of the machined workpiece, the surface roughness value of the parts is calculated. The final measurement method of surface roughness value of parts is given. Based on Open CV platform, the program of measuring workpiece surface roughness is compiled with C language, and the processing of workpiece surface picture is realized. From the image reading to the texture feature extraction of the surface, and then to the automatic measurement of the surface roughness value, the measured data are compared with the actual values obtained from the surface roughness test, and the two groups of data are regressed by parameter regression. Finally, the calibration of the system is realized. In this paper, the surface roughness detection of workpiece is combined with digital image processing technology to preprocess the surface image of the machined workpiece. On this basis, the texture feature of the image is extracted and calculated by programming. Finally, the roughness value of the workpiece surface is obtained, which proves the feasibility of this method. The research results are of great practical significance for the measurement of workpiece surface roughness.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TH161;TP391.41

【参考文献】