ArticlePublisher preview available

Hybrid vision system for online measurement of surface roughness

Optica Publishing Group
Journal of the Optical Society of America A
Authors:
Gui Yun Tian
Gui Yun Tian
Gui Yun Tian
  • This person is not on ResearchGate, or hasn't claimed this research yet.
rs lu at Hefei University of Technology
rs lu
Read publisher preview
Download citation
To read the full-text of this research, you can request a copy directly from the authors.

Abstract and Figures

A hybrid vision system for online measurement of surface roughness is introduced. The hybrid vision system applies two cameras for capturing the laser speckle pattern and scattering images simultaneously. With the help of advanced image processing, several features of texture and shape are computed for the surface roughness characterization. On the basis of experimental tests, feature fusion to improve measurement range and linearization of the measurement is also discussed.
Figures - available from: Journal of the Optical Society of America A
This content is subject to copyright. Terms and conditions apply.
A preview of this full-text is provided by Optica Publishing Group.
Learn more
Content available from Journal of the Optical Society of America A 
This content is subject to copyright. Terms and conditions apply.
Hybrid vision system for online measurement of
surface roughness
Gui Yun Tian
School of Computing and Engineering, University of Huddersfield, Queensgate, Huddersfield, HD1 3DH, UK
Rong-Sheng Lu
School of Instrument Science and Opto-Electronic Engineering, Hefei University of Technology, Hefei, Anhui, China
Received May 5, 2006; revised June 7, 2006; accepted June 23, 2006; posted July 10, 2006 (Doc. ID 70586)
A hybrid vision system for online measurement of surface roughness is introduced. The hybrid vision system
applies two cameras for capturing the laser speckle pattern and scattering images simultaneously. With the
help of advanced image processing, several features of texture and shape are computed for the surface rough-
ness characterization. On the basis of experimental tests, feature fusion to improve measurement range and
linearization of the measurement is also discussed. © 2006 Optical Society of America
OCIS codes: 290.0290, 240.0240.
1. INTRODUCTION
Surface roughness measurement is very important for
production quality control. It is essential to implement it
to ensure that the quality of a manufactured part can con-
form to its specified standard. It has been carried out us-
ing many technologies, such as contact stylus, microscopy,
ultrasonic and optical methodology, etc.1,2 It is worth no-
ticing that some modern techniques such as atomic force
microscopy and stylus can provide surface data with high
accuracy. However, these methods are difficult to apply
because they need delicate adjustment, extremely short
working distances, or contact with the surfaces measured.
Therefore, the development of noncontact optical tech-
niques to monitor surface changes is an area of high in-
terest, particularly online noncontact and high-speed
monitoring of surface roughness.
The typical methods based on light scattering are the
total integrated scattering (TIS) and angle-resolved scat-
tering (ARS) methods.2,3 In general, TIS methods have a
simple instrumentation and are more convenient to oper-
ate for surface roughness measurement, but much more
statistical information can be extracted from ARS. Over
the past decades, many efforts have been made to inves-
tigate the relationship between the surface roughness
and the angular distribution of scattered light
intensity.4–6 However, the theoretical expression of ARS
against root-mean-square (RMS) roughness height in-
volves states of polarization of incident light and light dif-
fraction. It is therefore difficult to precisely establish it in
theory. The derivation often makes a lot of assumptions
and simplifications so that the theoretical expression
sometimes has only qualitative meaning.7In practical on-
line surface roughness measurement, the statistical
methods representing scattered light intensity distribu-
tion against surface roughness are used.8Oneofthe
methods is based on the phenomenon that the intensity
distribution of the scattered light in the plane formed by
the incident light beam and surface normal depends on
the scattered light angle against the normal direction,
and there is more angular light scattering for a rougher
surface. Surface roughness in the algorithm is deter-
mined by computing the variance of the in-plane light-
scattering angle.9Another method has been carried out
with a machine-vision system, where surface roughness is
characterized by the frequency distribution of the gray-
level occurrence in a scattered light intensity image. The
scattered light frequency distribution is actually the his-
togram of the light-scattering intensity image obtained by
a CCD camera. Surface roughness in this method is
evaluated by the ratio between the RMS value of the his-
togram and its standard deviation.10 Other algorithms,
such as using the gray-level co-occurrence matrix11
(GCLM), are also used. However, these methods are often
effective for surfaces with uniform roughness distribution
and under conditions of noncoherent light illumination.
For the surfaces of steel parts manufactured by grinding,
milling, turning, etc., they often fail without any knowl-
edge of the surface micromanufacturing mark directions.
Surface roughness measurement by laser speckle
methods is used. Using the different properties of speckle
fields and the different setups of optical systems, re-
searchers have also developed a variety of speckle meth-
ods for surface roughness measurement. For instance,
surface roughness measurement may be implemented by
the speckle pattern illumination methods12,13 the speckle
contrast methods,14 and the speckle correlation
methods.15 Surface roughness measurement by means of
speckle pattern illumination is convenient to determine
RMS roughness height in the submicrometer range but
acquires a a complicated optical illumination system con-
sisting of a diffuser and lens.13 The speckle contrast
methods, which are based on the first-order statistics of
surface speckle patterns, usually can evaluate surface
roughness value less than Ra⬍0.3
␮
m. Speckle correla-
3072 J. Opt. Soc. Am. A/Vol. 23, No. 12/ December 2006 G. Y. Tian and R.-S. Lu
1084-7529/06/123072-8/$15.00 © 2006 Optical Society of America
... Five Loss function are analyzed to improve the accuracy of the prediction model. Tian et al [17,18] . characterized surface roughness by calculating texture and shape features, and demonstrated the feasibility of measuring roughness parameters through machine vision through experiments. ...
... Inverse moment represents the complexity of the spatial distribution of the image, and the calculation formula is shown in equation(17). ...
A new surface roughness measurement method based on QR-SVM
Preprint
Full-text available
  • Sep 2023
This study proposes a method for detecting surface roughness in machining, which solves the problem of low detection accuracy caused by a small sample size based on machine vision detection. The fusion of QR and Support Vector Machine (SVM) methods is used to detect surface roughness. Firstly, a contact roughness detector is used to measure the surface roughness value, and a CCD is used to obtain the processed surface image to obtain the sample. Secondly, the QR decomposition method is improved to generate virtual samples and expand the sample size. Extract the texture feature values of the image using the gray level co-occurrence matrix, and establish the correlation between roughness and texture features. Finally, support vector machines are used to classify the surface roughness of mechanical machining. The experimental results show that the accuracy of the surface roughness detection method based on machine vision has increased from 80.6–96.5%, proving the feasibility of this method and providing a theoretical basis for on-site detection of small sample surface roughness. This method has certain engineering application potential.
View
... They analyzed five loss functions to improve the prediction accuracy of the model. Tian et al. [14,15] characterized surface roughness by calculating texture and shape features and demonstrated the feasibility of machine vision-based roughness parameter measurement. Nammi et al. [16] demonstrated the importance of considering workpiece orientation in machine vision-based roughness measurement by extracting image features from low-carbon steel milling samples at different angles. ...
A new surface roughness measurement method based on QR-SVM
Article
Full-text available
  • Jun 2024
  • INT J ADV MANUF TECH
In this paper, a method for detecting surface roughness in machining processes is proposed to solve the problem of low detection accuracy caused by a small sample size in machine vision detection. The proposed method combines QR decomposition with the support vector machine (SVM) classifier to accurately assess surface roughness. First, a contact roughness detector is used to measure the surface roughness value, and a CCD camera is used to capture the processed surface image to obtain the sample. Subsequently, an improved QR decomposition method is employed to generate virtual samples and expand the sample size. Texture feature values of the image are then extracted using the grayscale level co-occurrence matrix, and the correlation between roughness and texture features is determined. Finally, SVM is employed to classify the surface roughness of machined components. Experimental results demonstrated that the accuracy of the machine vision-based surface roughness detection method increased from 80.6 to 96.5%, thus validating the feasibility of the proposed method and providing a theoretical basis for on-site detection of small-sample surface roughness. This method has potential for practical engineering applications.
View
... The spatial correlation among the pixels on the surface image is taken into account by this statistical technique. Surface roughness is collected by investigating the relationships between average surface roughness (Ra) and the GLCM features of the surface image [119,120]. The procedure of the computer vision system for measuring the surface roughness [121], is shown in Fig. 36. ...
Surface Characteristics Measurement Using Computer Vision: A Review
Article
Full-text available
Computer vision provides image-based solutions to inspect and investigate the quality of the surface to be measured. For any components to execute their intended functions and operations, surface quality is considered equally significant to dimensional quality. Surface Roughness (Ra) is a widely recognized measure to evaluate and investigate the surface quality of machined parts. Various conventional methods and approaches to measure the surface roughness are not feasible and appropriate in industries claiming 100% inspection and examination because of the time and efforts involved in performing the measurement. However, Machine vision has emerged as the innovative approach to executing the surface roughness measurement. It can provide economic, automated, quick, and reliable solutions. This paper discusses the characterization of the surface texture of surfaces of traditional or non-traditional manufactured parts through a computer/machine vision approach and assessment of the surface characteristics, i.e., surface roughness, waviness, flatness, surface texture, etc., machine vision parameters. This paper will also discuss multiple machine vision techniques for different manufacturing processes to perform the surface characterization measurement.
View
A Triangulation-Based Method for Complex Mold Parts Surface Wear Assessment
Article
  • Aug 2022
This work concerns the assessment of mold parts with multi-view scanning point cloud data. With the soaring demand for complex geometry parts, multi-axis electrical discharge machining (EDM) manufacturing also faces the challenge of high-cost and the requirement of high-accuracy. As a part of EDM, well-judged life-time and wear condition of mold parts are critical to ensuring product quality. In this work, we present a novel non-contact surface inspection solution for mold part 3D reconstruction and inspection based on light-section scanning. After calibration and registration, multi-view stereo (MVS) point cloud data for the parts can be obtained, which can then be used to generate 2D section profiles. Comparison method is designed to examine if the design tolerance is met by comparing the section profiles against those based on the CAD model of the parts. Our method is verified by assessing worn electrode parts, which are commonly used in machining centers. The method proposed in this paper ensures the detection accuracy and improves the detection speed. It overcomes the problem of slow detection speed of mold part wear using 3D point cloud.
View
View
View
Dual-mode snapshot interferometric system for on-machine metrology
Article
  • Apr 2019
For the first time, we present a dual-mode snapshot interferometric system for measuring both surface shape and surface roughness to meet the need for on-machine metrology in optical fabrication. Two different modes, interferometer mode and interference microscopy mode, are achieved using Linnik configuration. To realize snapshot measurement, a pixelated polarization camera is used to capture four phase-shifted interferograms simultaneously. We have demonstrated its performance for off-line metrology and on-machine metrology by mounting it on a diamond turning machine. © 2019 Society of Photo-Optical Instrumentation Engineers (SPIE).
View
Effect of Illumination Systems on Statistical Texture Parameters Based Clustering and Discrimination of Machined Surfaces Using Machine Vision
Article
  • Jul 2018
Inspection of surface quality using machine vision is based on the principle of surface characterization using texture parameters deduced from intensity values of images captured by the system. Illumination system plays significant role in deciding the performance and robustness of machine vision by controlling quality of image acquisition. This paper presents an experimental study of the effect of various illumination systems on image acquisition. Images of flat machined surfaces under five different illumination setups: ambient light, dark field, partial bright field with spotlight, partial bright field with tubelight and partial bright field with diffuse surface light; are grouped into three surface classes based on first and second order statistical texture parameters. The result showed that partial bright field with diffuse surface light provides maximum performance during image acquisition providing highest clustering accuracy. The images under this optimum setup are further analyzed using multiple discriminant analysis for determining the parameters significantly contributing to discrimination. The results showed that average height departure, root mean square, maximum peak to valley, skewness based on line samples; maximum peak to valley, skewness, kurtosis for surface and gray level co-occurrence matrix based contrast, correlation, energy, homogeneity effectively contributed to characterization of texture for discrimination.
View
Apport de la mesure in-situ pour la maîtrise de la qualité des surfaces fabriquées
Thesis
Full-text available
  • Oct 2016
Actuellement, les pièces de formes complexes sont réalisées par des procédés de plus en plus automatisés, à l'aide de machines outil à 5 axes ou de robots. Le processus de fabrication s'appuie sur la définition d'une gamme de fabrication et se décompose en différentes étapes à partir de la définition du cahier des charges. À chacune de ces étapes, des écarts géométriques entre la forme souhaitée et la forme obtenue peuvent apparaître. La surveillance des procédés de fabrication est donc nécessaire, nécessitant même une mesure à 100% des pièces produites lors de la phase de la stabilisation du processus. Ainsi, mes travaux portent sur maîtrise de la qualité des surfaces produites en vue d'améliorer le procédé de fabrication. Ainsi, la définition des paramètres de la stratégie de fabrication est un élément déterminant dans le processus de fabrication. La première partie de mes travaux est consacrée à l'amélioration de la prédiction de la topographie des surfaces générées par le procédé d'usinage. Un autre aspect des travaux sur l'amélioration de la qualité des surfaces concerne l'intégration des opérations de contrôle au sein du moyen de production, afin de réaliser une mesure dite in- situ. Enfin, la dernière partie des travaux réalisés ont pour objectifs de proposer des stratégies de numérisation minimisant le nombre de configurations capteur et la quantité de données acquises tout en contrôlant la qualité.
View
Surface roughness measurement based on digital holography
Article
  • Nov 2014
Surface roughness measurement based on digital holography was carried out. The tested surface roughness parameters were calculated from the surface profile which was mapped from the reconstructed phase distribution of digital hologram. Firstly, the standard calibration plate and high-resolution panel were taken as the tested samples to reconstruct errors and test the repeatability of the digital holography measurement system, involving the lateral size error and the height error. The reconstruction error and repeatability error for lateral size were 1.11%, 0.61%, respectively. And the errors for height were 11%, 1.8%, respectively. Then an ordinary coated reflector was taken as a sample, and its surface roughness values along 3 segments of evaluation length were 0.010 37 μm, 0.010 33 μm, and 0.009 67 μm, respectively(15 sampling lengths). ©, 2014, The No.205 Research Institute of China Ordnance Industry. All right reserved.
View
Computer Vision Based Non-contact Surface Roughness Assessment Using Wavelet Transform and Response Surface Methodology
Article
Full-text available
  • Jul 2005
This paper describes a new non-contact measurement approach in characterizing manufactured surfaces. Computer vision is applied to capture digital images of three types of anisotropic steel specimen surfaces from shaping, grinding, and polishing processes. Multiresolution wavelet decomposition is used to obtain signatures of surface profiles from the digital images. Relationships between these signatures and surface roughness parameters (Ra and Rq) are built by response surface methodology (RSM). The proposed models thus developed are suitable for predicting roughness in terms of the roughness parameters. Experimental results show that the proposed approach successfully correlates wavelet signals to Ra and Rq values. In addition, they also show repeatable gage capabilities. The proposed method is a good candidate for on-line, real-time surface roughness inspection when specimens of known surface roughness are available.
View
Application of machine vision in the automated inspection of engineering surfaces
Article
Full-text available
  • Feb 1992
Recent work has shown that computer vision has real potential when applied to the automated measurement of engineering component silhouettes, internal contours and profiles. The present study considers the detailed examination of surface textures using the 3D data arrays which are available using this vision system. Progress has been made towards establishing a model of texture based on vision data and it has been shown that a parameter based on both amplitude and space features could provide an additional control for use in on-line automated inspection of manufactured components. Corresponding stylus data has been employed as the basis for comparison in the current evaluation process. In the field of tribology the automated examination of a wear track using computer vision has demonstrated that the approaches can be integrated so as to provide full information about the overall dimensions, area and texture of a wear scar.
View
Surface-roughness dependence of the intensity correlation function under speckle-pattern illumination
Article
Full-text available
The surface-roughness dependence of the intensity correlation function of the speckle pattern, produced in the Fresnel region with fully developed speckle-pattern illumination, has been theoretically investigated and has been discussed as follows. This correlation function is represented by two correlation functions of scattered and unscattered components. As the diffuse object becomes rough, the speckle size varies from the speckle size of the illumination light to that obtained with the condition, so that the object is a deep phase screen. The speckle contrast, however, is always one.
View
View
Introduction to Surface Roughness and Scattering
Book
  • Jan 1999
The subjects of this book are surface roughness, primarily of optical surfaces, and light scattering. The type of scattering is classical scattering, not inelastic scattering such as Raman scattering. We have chosen to use the word roughness rather than the more general term texture, which is used for metal surfaces made by conventional machining processes and encompasses surface roughness, waviness, and lay. For optical surfaces we are concerned primarily with the surface roughness that causes light scattering. The roughness features that produce light scattering are typically separated by submicrometers to fractions of a millimeter. Features separated by larger distances, from hundreds of micrometers to several millimeters, are usually termed waviness and contribute to small-angle scattering. Features whose separations are still larger make up the so-called optical figure or deviation from the ideal geometrical-optics shape of a surface. The term lay refers to the directionality of surface features and is commonly used for engineering surfaces. If surface features are the same in all directions, the surface is referred to as isotropic. A small fraction of all optical surfaces such as diffraction gratings, holographic gratings, or those made by single-point diamond turning have a pronounced directionality and are thus said to have a lay; most surfaces, however, are isotropic. This book deals almost exclusively with isotropic surfaces whose features are separated by fractions of a millimeter. Measurement techniques and theories in the book can also be applied to surfaces having a lay (with appropriate modifications). Surface topography can be visualized in many ways. Looking at a surface with the unaided eye or with a simple magnifying lens can reveal many important surface features. Various optical or electron microscopes or one of the family of scanning probe microscopes will show much finer surface detail. Optical and mechanical surface profilers complement imaging techniques by giving numerical values of surface roughness and other statistical quantities. The atomic force microscope, the most useful scanning probe microscope for surface topography studies, maps the surface on a nearly atomic scale; the digitized data can give quantitative surface statistics. Light scattering from the surface microroughness gives information about the surface indirectly because a theory is required to relate the surface roughness to the scattered light. If the surface has the type of topography assumed in the light scattering theory, statistical quantities such as the root mean square roughness and power spectral density function can be obtained, but not a topographic map. Because all scattering theories contain a term for the wavelength dependence of the scattering, by measuring the scattering at several different wavelengths in the ultraviolet, visible, or infrared spectral regions, information can be obtained about different aspects of the roughness, such as the structure of a thin film coating, polishing or machining marks on the substrate, or the long-range surface waviness (mid-spatial frequency roughness). In any case, rough surfaces scatter much more light than do smooth ones. This book introduces the reader to various surface imaging and measurement techniques as well as to related topics such as cleaning of surfaces and standards for surface roughness measurements.
View
Machine vision assisted characterization of machined surfaces
Article
  • Nov 2010
This paper demonstrates the feasibility of using an inexpensive machine vision system to compute non-contact, optical parameters for the characterization of surface roughness of machined surfaces. Two parameters were selected for online analysis, where surface roughness is measured during the rotation of a specimen on a lathe. The sensitivity of the vision-based optical parameters to differences in surface roughness, ambient light and spindle speed of a lathe during measurement was evaluated. Statistical analysis of data collected through experimentation revealed that the vision parameters can discriminate different surface roughness heights and are insensitive to changes in ambient lighting and speed of rotation during measurement. The results of the experimental analysis were used to conclude the feasibility of using machine vision for the evaluation of surfaces.
View
Measuring the roughness of high-precision quartz substrates and laser mirrors by angle-resolved scattering
Article
  • Feb 2002
  • J OPT TECHNOL+
This paper presents the results of measuring the roughness of high-precision quartz substrates and of the mirrors of laser gyroscopes by angle-resolved scattering (ARS). The calculated one-dimensional and two-dimensional spectral-power-area (SPA) functions and the effective rms roughness sigma(eff) are compared with the results of measurements using atomic-force microscopy (AFM) and white-light interferometry (WLI). It is found that the roughness statistics are close to one-dimensional exponential statistics for the polished substrates and to two-dimensional exponential statistics for the laser mirrors. It is found that the best agreement of the experimental results using ARS and AFM is observed for the mirrors, which is most likely because the roughness of the sputtered coating is more homogeneous than that of the polished substrate. The sigma(eff) value strongly depends on the range of spatial frequencies of the roughness measured by the different methods. The good agreement of the results of the measurements using ARS, AFM, and WLI shows that it is suitable to use the ARS method for high-precision optical surfaces and coatings. (C) 2002 Optical Society of America.
View
View
Determination of the basic parameters characterizing the roughness of metal surfaces by laser light scattering
Article
  • Feb 1999
A method of determination of the basic morphology parameters of rough metal surfaces with periodic and random components of surface roughness by means of laser light scattering is presented. The method is based on interpreting the angular distribution of the laser light intensity scattered from such surfaces. The theory of the method is based on the scalar theory of diffraction and basic theorems of data processing. The new, more general way of interpreting experimental data is specified. It can also be used for the analysis of samples with quasiperiodic and random components of surface roughness; this is the case when standard measurement by means of a stylus profilometer can give incorrect results.
View
Roughness measurement of metallic surfaces based on the laser speckle contrast method
Article
  • Nov 1998
  • OPT LASER ENG
A method of measuring surface roughness of flat lapped, ground and polished metallic surfaces, by the far-field speckle contrast method is presented in this paper. The laser speckle contrast technique depends on the existence of an approximately linear relationship between the speckle contrast and the roughness of the illuminated surface. Initially it was shown that the linear relationship existed up to 0.1μm Ra (centre-line average) roughness using Helium–Neon light, after which a saturation effect was observed. The effect of varying the incident angle of illumination was investigated with a view to extending the measurement range. The use of high incident angles of illumination has been found to increase the surface roughness range up to 0.4μm measurement Ra.
View