基于图像清晰度评价的磨削表面粗糙度检测方法

针对当前机器视觉检测粗糙度主要采用图像灰度值信息进行统计分析,没有充分利用色彩信息且忽略了人眼视觉系统主观评判的问题,提出一种基于图像清晰度评价的磨削表面粗糙度检测方法。根据色块在不同等级粗糙度表面上形成的图像清晰度不一样,采用熵函数评价算法和基于色彩相关性的彩色图像清晰度评价算法分别构建清晰度与粗糙度之间的关系模型,论证了基于图像清晰度检测磨削表面粗糙度方法的可行性。试验结果表明,基于图像清晰度检测磨削表面粗糙度是一种可行的粗糙度检测方法,清晰度与粗糙度相关性强,清晰度有随着粗糙度的增大而减小的趋势,且利用色彩相关性的彩色图像清晰度评价算法对磨削表面粗糙度检测具有较好的灵敏性,并且该方法符合人眼视觉系统的主观评价;清晰度算法和主观评价二者结合可快速简易地在线检测工件的整体表面轮廓粗糙度。     

基于支持向量回归机的外圆纵向磨削表面粗糙度预测

表面粗糙度是评价磨削加工质量的重要指标,表面粗糙度预测是实现表面粗糙度在线控制的前提。针对现有神经网络方法在预测外圆纵向磨削表面粗糙度方面存在的不足,提出了一种新的基于支持向量回归机的外圆纵向磨削表面粗糙度预测方法。在分析了影响外圆纵向磨削表面粗糙度预测主要因素的基础上,建立了基于支持向量回归机的外圆纵向磨削表面粗糙度预测模型。应用实例的仿真结果表明,所建立的预测模型具有较强的泛化能力和较高的预测精度。     

磨削粗糙度在线检测方法的实验研究

本文对基于摩擦声发射信号的磨削粗糙度在线检测方法进行了实验研究。实验结果表明，采用声发射传感器探头与磨削表面摩擦产生的声发射信号的特征可以对磨削表面粗糙度进行评价，建立了摩擦声发射信号特征与磨削表面粗糙度之间的对应关系，并通过实验对该方法的可行性进行了实测。结果表明，探针与工件表面摩擦声发射信号的FFT和RMS特征与磨削粗糙度有很好的对应关系，可用于磨削表面粗糙度的在线检测。     

基于色彩均匀敏感度的磨削表面粗糙度测量

对于纹理较弱、粗糙度较低的磨削加工表面,采用传统的粗糙度测量方法存在提取表面纹理特征困难,且测量精度受表面纹理方向的影响,对此提出了一种对纹理方向不敏感的彩色光测量方法。首先,基于Beckmann-Spizzichino光反射模型,分析了粗糙面的光反射机理;然后,将磨削加工表面作为反射面,提出了一种基于彩色光反射图像的磨削表面粗糙度测量方案,并搭建了测量平台;最后,结合磨削表面存在的纹理方向特性,设计了一种色彩均匀敏感度测量指标Cu_(sd),建立了与磨削表面粗糙度的强相关性,实现了对磨削表面的粗糙度测量。实验结果表明:指标Cu_(sd)性能明显优于其他对比指标;与粗糙度的相关性达到0.978,具有较强的正相关性,且该指标对纹理方向不敏感;对粗糙度Ra范围在0.017～0.646μm的磨削表面,指标Cu_(sd)预测均方差为0.000 84,平均测量误差为0.02μm,可满足磨削表面测量要求。     

磨削液对陶瓷结合剂CBN砂轮磨削性能影响

分析了磨削液对陶瓷结合剂CBN砂轮磨削性能的影响，使用3种磨削液在精密外圆磨床M1420E上进行了磨削加工实验，用加工表面微观形貌、表面粗糙度R。值、工件表面残余应力以及砂轮径向磨损量对磨削液效能进行评价。结果表明，轻质润滑油不仅能提高工件表面质量，降低表面粗糙度值，而且砂轮磨损量明显降低，乳化液和化学合成液对磨削性能的影响各有利弊，润滑油是陶瓷结合剂CBN砂轮磨削的优选磨削液。     

基于声发射信号的精密外圆切入磨削时间评估算法及试验研究

针对精密外圆切入磨削智能监控的需求,设计一种基于声发射信号的磨削时间在线评估方法。通过建立声发射信号方均根值曲线预测模型,获得声发射信号与磨削系统时间常数的关系,设计磨削系统时间常数在线计算方法;利用在线检测的声发射信号识别砂轮运动去除状态,推导基于声发射信号的外圆切入磨削表面粗糙度评价和工件几何精度预测模型,以此建立砂轮进给与驻留时间的评估算法;编写磨削时间分析评估软件,设计磨削时间在线评估方法,通过加工试验分析磨削时间对磨削加工精度与表面粗糙度的影响规律,并对评估算法进行验证。试验结果表明:该评估方法能够根据磨削时间有效评价加工质量,为精密外圆切入磨削智能监控与工艺优化提供决策依据。     

基于遗传算法与BP神经网络的微晶玻璃点磨削工艺参数优化

通过低膨胀微晶玻璃点磨削实验,测试了加工表面粗糙度、表面硬度,分析了实验数据变化趋势。通过最小二乘拟合,建立了关于粗糙度、表面硬度的一元数值模型,并将模型预测值与实验值进行了比较,以验证模型的精确性,结果表明模型具有较高的精度。根据正交实验结果,基于BP神经网络算法和遗传算法,建立了粗糙度、表面硬度的多元数值模型并以此作为目标函数,以表面硬度最大和表面粗糙度最小作为优化目标,基于遗传算法进行了工艺参数的双目标优化,获得了一组点磨削工艺参数的最优解范围,实验验证结果表明优化结果是合理的。     

恒转速曲轴切点跟踪磨削表面粗糙度的研究

通过恒转速曲轴切点跟踪磨削的运动学方程,结合基于磨削深度与弹性形变的磨削表面粗糙度计算模型,对不同磨削深度下的曲轴连杆颈表面粗糙度进行了求解计算。计算结果表明,曲轴转角及磨削深度对工件表面粗糙度值存在影响差值,揭示了曲轴连杆颈表面粗糙度值的变化规律,为今后恒转速曲轴切点跟踪磨削的表面粗糙度分析提供了理论依据。     

工程陶瓷磨削表面粗糙度数学模型的研究

提出利用平行于磨削方向的表面粗糙度Ｒａｐ和垂直于磨削方向的表面粗糙度Ｒａｖ两个参数同时衡量工程陶瓷磨削表面质量,建立了表面粗糙度Ｒａｐ和Ｒａｖ值的数学模型公式,并根据模型公式提出了改善磨削表面质量的措施。试验表明：根据模型公式算出的理论值和实际测得值的吻合性较好,该数学模型适用于工程陶瓷磨削表面质量的预测与估计。     

石英陶瓷磨削表面粗糙度的计算机模拟

工程陶瓷通常采用的加工方法是磨削，磨削表面粗糙度对零件使用性能具有很大的影响。由于客观条件的限制，磨削时各参数如何选择能获得低的表面粗糙度仍然是现今许多学者研究的方向。本文根据单颗磨粒的轨迹模型，利用Matlab软件模拟出不同磨削参数下工件表面的粗糙度情况，为不同磨削参数下获得表面的粗糙度情况模拟提供了一种新方法。     

细粒度金刚石砂轮超精密磨削硅片的表面质量

为实现硅片高质量表面的超精密磨削，研究了5000目、8000目和30 000目金刚石砂轮磨削硅片的表面质量。利用数学模型预测了硅片磨削表面的粗糙度Ra并对预测结果进行了试验验证，分析了硅片磨削表面的形貌特征；通过磨床主轴电机的电流变化对比分析了5000目、8000目和30 000目砂轮磨削过程中的磨削力变化趋势。研究结果表明：8000目砂轮磨削后的单晶硅表面粗糙度Ra小于10 nm,亚表面损伤深度小于150 nm,磨削过程中的磨削力稳定，磨削质量优于5000目砂轮，磨削过程的稳定性优于30 000目砂轮。     

Evaluation of surface roughness based on monochromatic speckle correlation using image processing

The measurement of roughness on machined surfaces is of great importance for manufacturing industries as the roughness of a surface has a considerable influence on its quality and function of products. In this paper, an experimental approach for surface roughness measurement based on the coherent speckle scattering pattern caused by a laser beam on the machined surfaces (grinding and milling) is presented. Speckle is the random pattern of bright and dark regions that is observed when a surface is illuminated with a highly or partially coherent light beam. When the illuminating beam is reflected from a surface, the optical path difference between various wavelets with different wavelength would result in interference showing up as a granular pattern of intensity termed as speckle. The properties of this speckle pattern are used for estimation/quantification of roughness parameters. For measurement of surface roughness, initially the speckle patterns formed are filtered in the spatial frequency domain. The optical technique, namely spectral speckle correlation (autocorrelation) is utilized in this work for the measurement of roughness on machined surfaces. It has been observed that the pattern formed is dependent on the roughness of the illuminated surface. For example, a rough surface (milled) shows a small central bright region with a rapid decrease in intensity towards the edges, while a smooth surface (ground) shows a large central bright region with gradually decreasing intensity towards the edges. The complete methodology and analysis for quantification/estimation of surface finish of milled and ground surfaces based on speckle images that could be implemented in practice, is presented in this paper.     

Assessment of surface roughness based on super resolution reconstruction algorithm

In this work, an attempt has been made to use a super resolution image processing algorithm for preprocessing the images over and above existing image quality enhancement techniques. The improved quality images processed using a machine vision system have been used to assess the quality of the surfaces. To ensure the validity of the approach the roughness values quantified using these images are then compared with widely accepted standard mechanical stylus instrument values. Quantification of digital images for surface roughness is performed using two Fourier transform parameters (major peak frequency and principal component magnitude squared value) and the standard deviation of gray level intensity values. Then the group method of data handling (GMDH) technique was used to obtain an analytical relationship of the roughness parameters calculated using the digital surface image and the stylus instrument values. We present in this paper an analysis based on the comparison to make sure that the present approach of estimation of surface finish based on the digital processed image could be implemented in practice.     

评价彩色图像质量的四元数矩阵最大奇异值方法

针对传统的图像质量评价方法忽略颜色信息以及与人眼感知信息一致性差的问题,提出了一种全面利用彩色图像颜色信息,突出人眼敏感图像结构的彩色图像质量客观评价方法。将图像中人眼敏感的结构分为细节,亮度和颜色3方面因素,将四元数矩阵作为载体,构造了一种用于彩色图像质量评价的四元数矩阵,并对其进行奇异值分解。将最大奇异值作为度量图像结构相似性的主要参数,通过分析图像结构差异映射图谱得到了最终的量化评价结果。采用LIVE数据库中包含5种失真类型的982张测试图片验证了提出的算法,得到的交叉失真实验非线性拟合均方根误差（RMSE）值为9.176,Spearman等级相关系数（SROCC）值为0.929 6,而结构相似度（SSIM）方法的RMSE值为9.299,SROCC值为0.925 6。试验结果表明,该方法采用四元数矩阵描述彩色图像的结构信息,考虑了彩色图像的多方面结构特征,与人眼视觉感知特性的一致性优于传统方法。     

Rapid measurement of surface roughness for face-milling aluminum using laser scattering and the Taguchi method

In this paper, an experimental model for the rapid measurement of surface roughness (R_rms) in CNC face-milling specimens using the laser speckle method and digital image processing is established. The specimens used in this study were made of 6061-aluminum alloys through the high-speed face-milling process. In order to evaluate the effect of machining conditions, such as the feed rate, the spindle speed, the depth of cut, and the material of the cutting tool on the roughness of the specimens, the Taguchi method was used to determine the optimal parameters for machining. The laser radiation results in the speckle structure formed in the space when coherent light is scattered through an optically rough surface. The features of the speckles depend on the characteristics of the rough surfaces. Hence, the experimental work for the roughness measurement is based on the speckle effect. The experimental setup in this study consisted of a He-Ne laser, a ground glass, a CCD camera, and a digital image processing system developed using the Virtual Basic language. Computer evaluation of the speckle images revealed the values of R_rms rapidly. This study proposed a precise and non-contact optical method for evaluating the surface roughness from 0.20 to 0.60 μm.     

Potential use of image analysis techniques in surface texture studies

This paper discusses a possible new approach to surface texture measurement based on the computer analysis of pictorial surface images. The method involves the evaluation of image textural features which are then compared with prototype standards derived from surfaces with known roughness properties. Preliminary results show that the technique is capable of distinguishing between different manufacturing methods and roughness values     

Surface roughness and texture analysis in microscale

The capacity of various instruments in roughness measurements and analysis is compared. Review of various models of roughness is made and the models of contact mechanics are presented, when taking account the nanometer scale roughness and relating phenomena of adhesion and surface forces. The concept of multi-level models of roughness and contact area is presented. Analysis of surface topography as a spatial pattern is given, when using the approaches of image recognition theory operating with the 3D digital images processing. Qualitatively the spatial structure is often characterized in terms of texture features such as random, linear, wavy etc., and some national standards introduce spatial structure of machined surfaces. However, texture characteristics are not adequately investigated. AFM images of different surfaces were used as initial data and multi-dimensional scaling technique was used for the data analysis. The study has shown that there are at least four types surface textures on nanoscale level. The correlation was found between texture types and reasons of their formation.     

马尔科夫随机场模型下的Retinex夜间彩色图像增强

由于Retinex算法在处理夜间彩色图像时容易出现光晕、颜色失真、细节丢失与噪声干扰等问题,本文基于马尔科夫随机场（MRF）提出了一种针对单幅图像的Retinex图像增强算法。该算法在HSV颜色空间下采用线性引导滤波估计图像照度分量;在MRF模型下求解仅包含物体本身特性的反射分量,并通过颜色恢复函数与增益补偿方法进行颜色恢复与校正,最终实现了夜间彩色图像的增强。实验结果表明,利用本文算法处理后图像的均值（整体亮度）可以提高2倍以上,标准差、熵、峰值信噪比（PSNR）等参数均有5%以上的提升。与其它基于Retinex原理的算法相比,本文提出的算法增强效果显著,具有消除“光晕伪影”现象、抑制噪声、颜色保真和有效地凸显边缘细节信息等能力。     

Automatic defect inspection for LCDs using singular value decomposition

Thin film transistor liquid crystal displays (TFT-LCDs) have become increasingly popular and dominant as display devices. Surface defects on TFT panels not only cause visual failure, but result in electrical failure and loss of LCD operational functionally. In this paper, we propose a global approach for automatic visual inspection of micro defects on TFT panel surfaces. Since the geometrical structure of a TFT panel surface involves repetitive horizontal and vertical elements, it can be classified as a structural texture in the image. The proposed method does not rely on local features of textures. It is based on a global image reconstruction scheme using the singular value decomposition (SVD). Taking the image as a matrix of pixels, the singular values on the decomposed diagonal matrix represent different degrees of detail in the textured image. By selecting the proper singular values that represent the background texture of the surface and reconstructing the matrix without the selected singular values, we can eliminate periodical, repetitive patterns of the textured image, and preserve the anomalies in the restored image. In the experiments, we have evaluated a variety of micro defects including pinholes, scratches, particles and fingerprints on TFT panel surfaces, and the result reveals that the proposed method is effective for LCD defect inspections.