【作者】 王薇；

【导师】 谭庆昌；

【作者基本信息】 吉林大学 ， 机械设计及理论， 2006， 硕士

【摘要】 在机械工业中,表面粗糙度是工件表面质量的一个重要技术指标。随着表面加工质量的不断提高,对表面粗糙度的测量技术和评定方法提出了更高的要求。本文通过对图像处理技术及分形几何理论系统的学习和研究,提出了一种基于分形几何理论和图像处理技术的表面粗糙度非接触测量方法。图像处理是新兴的针对图像信息的非接触式测量技术;分形几何的主要研究对象是像表面粗糙形貌这样不规则的、复杂的、具有自相似性的几何结构。本文通过对图像处理技术的研究,设计并搭建了图像采集系统。利用图像采集系统采集工件表面形貌的图像。采用适当的图像预处理方法以改善原始图像的质量问题和突出图像中我们感兴趣的信息。然后,应用分形几何的相关理论针对图像预处理后的表面形貌的图像数据建立数学模型,得到可以表征表面粗糙程度的一组分形参数。通过对几组不同加工方法、不同粗糙程度的样块的试验分析,讨论并总结分形参数表征表面粗糙度的性质和特点。( D, G)更多还原

【Abstract】 In the mechanical industry, the roughness is an important technical parameterof the surface quality. With the unceasing improvement of the quality of surfacemanufacture, the method of the evaluation and the measure to the surfaceroughness sets a higher request.At present, the stylus type measurement applies more widely in themechanical industry, its characteristic is intuitionistic, reliable and easy foroperation. It can satisfy the need of the general measure. But because the stylustype measure technique is contact type, it has its certain limitation: 1) The shapeand the size of the tip affects the measuring accuracy. 2) The tip contacts the objectsurface direct, so it is easy to scratch the surface. 3) The measure result has certainone-sidedness and fortuitousness. 4) It is unfit for the on-line and real-timeinspection. Therefore, people have launched extensive research to each non-contact type measuring technique. In recent years, along with the rapiddevelopment of photoelectricity technology, the image processing technology andthe computer technology, the roughness measuring technique based on the imageprocessing technology receives people’s more attention. It has the characteristic ofnon-contact, fast, convenient and powerful in data processing etc.Applies what kind of attribute parameter to describe and to evaluate the roughsurface is also count for much. The traditional roughness parameter like tcare mostly based on statistics, they are scale-dependent. The sampling length, thesampling spacing and the instrument resolution etc affect the result heavily. Thefractal was born in 20 centuries 70’s.It is different from the traditional geometry, itis used to indicate one kind of design which consists of many fragmentary graphs,perhaps one kind of behavior or phenomenon which is composed by many events.The fractal is scale-independent. Therefore fractal geometry is extremely suitablefor describing and analyzing the geometry structure which has the characteristic ofirregular, complex and autocorrelation like the surface roughness.R a , RqeThis paper has researched and studied the image processing technology, thendesigned and built the image capture system. Using the system to capture the imageof surface appearance, and then this primitive image is send to the computer imageprocessing system after digitized to carry on next step of processing. In the part ofimage pretreatment, this paper has analyzed the defect of the primitive image,chose the spatial filtering enhancement, removing the uneven caused byillumination, and histogram transformation enhancement to improve the primitiveimage, and it is proved by experiment that these methods take effect on the imagedefect. Then use the theory related to the fractal geometry to establish themathematical model of the surface appearance image data which has beenprocessed by image pretreatment system, obtain the parameters( D, G) based onthe fractal theory. Finally during measuring and analyzing several groups ofspecimen which were produced by different processing methods and have differentrough degree, discussed and summarized the nature and the characteristic of thefractal parameter:1) The fractal dimension is a measurement parameter about how thesuperficial outline is similar, therefore under the identical machining, the value ofis basically same, regardless of its rough rank.DD2) The fractal dimension is scale-independent, the sampling length, thesampling spacing and the instrument resolution doesn’t affect it.D3) The fractal dimension denote the complex degree and the irregularityof the outline curve, therefore the superficial appearance produced by differentmachining way has different fractal dimension even if their traditional roughnessparameter were same;If the machining way is different, which surface outline issmoother and less irregular, its fractal dimension is bigger.DD4) Regarding the same machining way and the same fractal dimension ,the proportion coefficient G increases along with the surface roughness rankincreases, and reduces along with the surface roughness rank reduces.DThis paper is composed of six chapters. The first chapter introduced thebackground and the significance of this work, and summarized several existingroughness measuring technique, finally elaborated the main content of this paper;The second chapter have carried on a more detailed introduction on the theory andthe basic concept of fractal;The third chapter first has made a synopsis for theimage processing technology, then emphatically introduced the theory and thehardware equipment related to the image capture system, finally has specified theequipment of the experiment system and its correlation parameter;The fourthchapter elaborated the image pretreatment methods of this paper and their effect;The fifth chapter first analyzed the limitation of the traditional roughness,proposed a mathematical model of the surface appearance’s image data based onthe fractal theory, obtains the surface roughness parameter (D, G) based on fractal.Finally carries on the experiment to analyze and confirm it;The sixth chapter is thepart of conclusion, it summarize this paper and point out the insufficiencies of thispaper.更多还原