Conference Papers by Richard Leach
To address the future challenges in quality monitoring of metal laser powder bed fusion, a novel ... more To address the future challenges in quality monitoring of metal laser powder bed fusion, a novel method is proposed to detect topographic anomalies on layer surfaces, which may appear during the manufacturing process. The method combines light scattering and principal component analysis. Scattering patterns, experimentally generated from real surfaces representative of in-control processes and encoded as digital images, are collected and used to build a reference set, which is then further populated by simulation. Principal component analysis is then applied to the set. A certain number of principal components is extracted and used to define a transform to map any scattering pattern to principal component space. Using the created transform, any new scattering pattern can be transformed to principal component space and then back into the original space (reconstruction), with some reconstruction error. The error is expected to be low if a pattern from the reference set is processed. However, if a different pattern is processed, e.g. generated by an out-of-control layer topography, then the reconstruction error is larger. In this work, a layer monitoring system is proposed, capable of detecting out-of-control topographies through observation of the reconstruction error. The system was implemented and experimentally validated through application to a selected test case. Measurement, laser powder bed fusion, light scattering, principal component analysis
We propose an instrument model for coherence scanning interferometry using familiar Fourier optic... more We propose an instrument model for coherence scanning interferometry using familiar Fourier optics methods, the spectrum of plane waves, and the assumption that the light source spectral bandwidth is the dominant factor in determining fringe contrast as a function of optical path length. The model is straightforward to implement, is computationally efficient, and reveals many of the common error sources related to the optical filtering properties of the imaging system. We quantify the limits of applicability of the model related to the geometrical approximations for conventional Fourier optics, particularly for high numerical apertures, and when using the fringe contrast for determining surface heights. These limitations can be overcome by using a three-dimensional imaging model.
Please cite this article as: Jitka Metelkova, Lars Vanmunster, Han Haitjema and Brecht Van Hoorew... more Please cite this article as: Jitka Metelkova, Lars Vanmunster, Han Haitjema and Brecht Van Hooreweder, Texture of inclined up-facing surfaces in laser powder bed fusion of metals, Additive Manufacturing, (2020)
Close-range photogrammetry is an optical metrology technique for measuring the three-dimensional ... more Close-range photogrammetry is an optical metrology technique for measuring the three-dimensional form of an object. Photogrammetry has many advantages over competing techniques such as structured light projection-the primary being the relative inexpense of the equipment required. However, photogrammetry is currently slow and user dependant at many stages. In order to facilitate automation of the measurement process, the spatial relationship between the camera and the measurand must be known. We present a method of establishing this relationship from a single initial image via a convolutional neural network (CNN). The CNN is trained on synthetic images of an object which are generated from the computer aided design (CAD) data. A simulated texture is applied to the CAD data which is representative of the observed surface created by the manufacturing process and material. By using this representative texture combined with accurately characterised camera intrinsic parameters, near photo-realistic images can be generated. By using the CAD data to generate the training data, training can be conducted in parallel with manufacture removing any additional time penalty that would be otherwise incurred. This method also removes any need for manual labelling of training data. We test this approach on a range of different CNN architectures and select the best performing network based on the model loss and prediction error.
To address the increase in tight tolerance requirements for small parts produced by precision man... more To address the increase in tight tolerance requirements for small parts produced by precision manufacturing, on-machine optical areal surface topography instruments are emerging. To calibrate these instruments and estimate their measurement uncertainty, their metrological characteristics need to be determined according to ISO 25178 part 600. In this paper, the amplification coefficient and linearity deviation metrological characteristics in the vertical axis of a prototype compact on-machine focus variation areal surface texture and form measurement sensor are determined. With a series of experiments in different positions of the vertical axis using calibrated materials measures with heights from 0.2 µm to 1000 µm, we determine the amplification coefficient and linearity deviation for the vertical axis. In addition, with a procedure derived from ISO 10360 part 8, the maximum permissible unidirectional stationary error of the vertical axis is determined.
We present the novel design of an all-optical dimensional measuring system (AODMS) for measuring ... more We present the novel design of an all-optical dimensional measuring system (AODMS) for measuring the geometry and surface texture of small-scale components. The system is designed to operate in a cube of 100 mm sides, with micrometre or sub-micrometre measurement uncertainties. The AODMS includes a four-axis motion system for mounting and moving the sample to be measured, a photogrammetric system for coordinate measurement and motion system tracking, a combination of coherence scanning interferometry and focus variation microscopy for texture measurement and a metrology frame fabricated using additive manufactured lattice structures with internal resonating bandgaps for vibration isolation. The paper will discuss the development of the AODMS, the experimental realisation of the instrument and the first steps in its validation. Coordinate metrology, surface topography, information-rich metrology
We introduce a novel method for the validation of surface texture filtration. The method utilises... more We introduce a novel method for the validation of surface texture filtration. The method utilises a combination of a mathematically defined surfaces and an accurate evaluation of the linear Gaussian filter transmission characteristic function to produce a mathematically traceable reference pair suitable for the performance assessment for surface texture analysis software. The method is suitable for both profile and areal surfaces and filtration methods. We showcase the method with two example reference pairs; one profile and one areal, which are used to demonstrate how the performance of a surface texture analysis software package can be assessed. Results are presented for a variety of Gaussian filter options, including end effect removal widths and end effect management.
In this paper, we propose a light scattering method to identify classes of structured surface top... more In this paper, we propose a light scattering method to identify classes of structured surface topographies and estimate their main geometric properties. The method is based on a cascaded machine learning model, designed as a two-layer architecture implemented using neural networks. The first layer consists of a classification model designed to determine which type/class of surface is being observed amongst a set of predefined surfaces The second layer, cascaded to the first one, is designed to infer geometric properties specific to the individual structured surface being measured within each class, for example, pitch and height for a grating-type surface. The training datasets for the cascaded machine learning model, i.e. scattering signals from different surfaces, are generated through rigorous scattering simulation applied to computer-generated surfaces and based on a boundary element method. Once the model is trained, any scattering signal obtained from a real surface belonging to the considered classes can be fed into the model, and both the surface class and specific values for its geometric properties can be quickly estimated. For validation, we developed a prototype experimental apparatus to generate light scattering data from real surface samples. Different grating patterns (classes) were considered, as well as different values for the main geometric properties specific to each class. Validation consisted both in the assessment of classification performance in recognising instances of each specific class and in quantification of estimation accuracy in determining the geometric properties of each instance, by comparison with measurements performed with atomic force microscopy.
Additive manufactured parts have complex geometries featuring high slope angles and occlusions th... more Additive manufactured parts have complex geometries featuring high slope angles and occlusions that can be difficult or even impossible to measure; in this scenario, photogrammetry presents itself as an attractive, low-cost candidate technology to acquire digital form data. In this paper, we propose a pipeline to optimise, automate and accelerate the photogrammetric measurement process. The first step is to detect the optimum camera positions which maximise surface coverage and measurement quality, while minimising the total number of images required. This is achieved through a global optimisation approach using a genetic algorithm. In parallel to the view optimisation, a convolutional neural network (CNN) is trained on rendered images of the CAD data of the part to predict the pose of the object relative to the camera from a single image. Once trained, the CNN can be used to find the initial alignment between object and camera allowing full automation of the optimised measurement procedure. These techniques are verified on a sample part showing good coverage of the object and accurate pose estimation. The procedure presented in this work simplifies the measurement process and represents a step towards a fully automated measurement and inspection pipeline.
Optical topography measurement of steeply-sloped surfaces beyond the specular numerical aperture ... more Optical topography measurement of steeply-sloped surfaces beyond the specular numerical aperture limit," Proc. ABSTRACT Engineered functional surfaces often feature varying slopes on macro-and micro-scales. When surfaces are mirror-like, the highest surface slope that can be measured by a far-field 3D imaging optical surface measuring instrument is the arcsine of the numerical aperture (NA) of the objective lens, i.e. the acceptance angle of the lens. However, progress in instrument design has allowed for measurement of non-specular surfaces with slopes steeper than this "traditional" NA limit. Nonetheless, there is currently a lack of understanding about the instrument response to surfaces with steep slopes beyond this limit. It is unclear over what surface spatial frequencies we can expect to accurately report fine surface-feature details. Here we present results demonstrating the capability of a commercial coherence scanning interferometer for measuring surface topography of a roughened flat and a blazed grating with tilt angles greater than the NA slope limit. We show that the surface form, i.e. the tilted plane, can be measured correctly. But, while surface texture information that can appear useful is also obtained, tilting significantly influences the measurement accuracy of micro-scale texture, and for asymmetric gratings, can depend on the tilting direction. A simplified surface scattering model suggests that the loss of scattered power captured by the instrument and a low signal-to-noise ratio causes the reduction of measurement accuracy. However, a rigorous three-dimensional instrument model is needed for a full understanding; we will develop this in our future work.
Nonperiodic metamaterials with appropriately designed resonator distributions can have superior v... more Nonperiodic metamaterials with appropriately designed resonator distributions can have superior vibration attenuation capabilities compared to periodic metamaterials. In this study, we present an optimization scheme for the resonator distribution in rainbow metamaterials that are constitutive of a Π-shaped beam with parallel plate insertions and spatially varying cantilever-mass resonators. To improve the vibration attenuation of rainbow metamaterials at specific design frequencies, two optimization strategies are proposed, aiming at minimizing the maximum and average receptance values. Objective functions are set up with the frequency response functions predicted by the displacement transfer matrix model. The masses of the two sets of resonators, clamped at different side walls of the Π-shaped beams, constitute the set of design variables. Optimization functions are solved using a genetic algorithm method. Results of case studies showed that the receptance values of the nonperiodic metamaterial is greatly reduced within the optimization frequency range, in comparison to the periodic metamaterial.
INSTRUCTION Advanced manufacturing techniques, such as additive manufacturing, are shaping the ve... more INSTRUCTION Advanced manufacturing techniques, such as additive manufacturing, are shaping the very nature of present-day industry and continue to redirect future research and developments. Simultaneously, 3D measurement is being integrated into manufacturing processes beyond quality control. 3D measurements are required for in-line and in-process applications to enhance condition monitoring, processing decisions, machine control and real-time corrections [1]. Subject to these demands, the preference for fast, portable and flexible solutions applicable to complex geometries in metrology is dominating. With additive manufacturing having form accuracy in the order of tens to hundreds of micrometres [2], feasible optical 3D measurement methods, such as photogrammetry and fringe projection, become viable options for quality assessment. However, the use of these methods requires testing to ascertain the limits of their capability and reliability. Photogrammetry provides compact 3D measurement solutions that are integrable and suitable for in-situ applications [3]. Common photogrammetric solutions involve space triangulation and multi-vision imaging. Multi-vision systems use one or more additional camera viewpoints. As the basis of multi-vision systems, stereo vision uses one additional camera to establish triangulation of detected image points with a 3D scene. In principle, more cameras provide further triangulation constraints of a scene which can increase accuracy [4]. To relate image distances to metric scales of objects, either known dimensions in the acquired images or known relative camera positions are used. In most commercial products, there is a requirement to use markers or feature-based gauges [5] for higher accuracy in photogrammetry. Marker-less methods, such as the structure from motion method, lack the necessary accuracy and are, therefore, limited to preliminary measurements and collision avoidance in industrial applications [6]. As an initial step to improve their accuracy, it is essential to evaluate of the performance of the multi-imaging systems used in these photogrammetric methods. This paper is focused on evaluating the performance of a stereo vision system in measuring the kinematic error of a translation stage. To accomplish this, motion of a linear stage is measured using both stereo vision and laser interferometry. The laser interferometer measurements provide a traceable benchmark for the performance evaluation of the stereo vision system. The main contribution to the overall kinematic error of a motion system usually comes from quasi-static sources [7]. Quasi-static errors are time-invariant and include geometric errors, misalignment and other errors due to manufacturing defects. Errors in the stage degrees of freedom are measured and analysed in the following sections. EXPERIMENTAL SETUP The kinematics of a motorised single-axis linear stage, with 100 mm traveling range, have been studied by measuring its linear and angular degrees of freedom. The moving plate of the linear stage was aligned to a linear rail using tight guiding wheels, that were collectively driven by a lead screw mechanism. A maximum load of 115 N and a positioning accuracy of 26 μm are specified by the stage's manufacturer. An optical target (a checkerboard pattern 48 mm × 44 mm) was used as a stage tracking target. The optical target and a retroreflector were mounted on the stage to provide positional feedback to the stereo system and the interferometer respectively. The stereo vision system is composed of two CMOS machine vision cameras with sensors of 1.85 μm pixel size and 12.2 megapixel resolution.
Surfaces featuring complex topographies, such as high slope angles, large curvatures and high asp... more Surfaces featuring complex topographies, such as high slope angles, large curvatures and high aspect-ratio structures on both macro-and micro-scales, present significant challenges to optical measuring instruments. Here we demonstrate a method to characterise and correct the three-dimensional surface transfer function (3D STF) of a coherence scanning interferometer (CSI). Slope-dependent errors present in the original measurements are reduced after phase inversion of the 3D STF, and the final results agree with traceable contact stylus measurements within the 30 nm reproducibility of the stylus measurements. This method enables in-situ compensation for errors related to aberrations, defocus and diffraction.
Surfaces featuring complex topographies, such as high slope angles, large curvatures and high asp... more Surfaces featuring complex topographies, such as high slope angles, large curvatures and high aspect-ratio structures on both macro-and micro-scales, present significant challenges to optical measuring instruments. Here we demonstrate a method to characterise and correct the three-dimensional surface transfer function (3D STF) of a coherence scanning interferometer (CSI). Slope-dependent errors present in the original measurements are reduced after phase inversion of the 3D STF, and the final results agree with traceable contact stylus measurements within the 30 nm reproducibility of the stylus measurements. This method enables in-situ compensation for errors related to aberrations, defocus and diffraction.
Digital fringe projection is a non-contact method that is widely used for the dimensional charact... more Digital fringe projection is a non-contact method that is widely used for the dimensional characterisation of complex manufactured parts. However, single camera-projector fringe projection systems struggle to acquire the full three-dimensional point cloud in one acquisition due to their relatively small field-of-view, and the typically freeform geometry, potentially with multiple occlusions, of additively manufactured parts. In this paper, we demonstrate that a multi-view fringe projection system is an effective solution to address form measurement of complex additively manufactured parts. However, the global geometric characterisation of multiple sets of cameras and projectors is a challenge due to the lack of a common field-of-view and overlapping of the projected fringes. We use a cost-effective multi-view fringe projection system to characterise an assembly of multiple sets of cameras and projectors with different perspectives. We present an automated characterisation method that uses a checkerboard which is moved in the measurement volume. The absolute phase information from the captured phase-stepped images is used to establish the global geometric properties by automated image processing and parameter optimisation. The geometric characterisation method is implemented and the multi-view system has been used to measure a range of additive parts. In this paper, we present the three-dimensional reconstruction results from different views that are combined to optimise the global geometric parameters.
Feature-based characterisation approaches, based on assessment of individual topographic formatio... more Feature-based characterisation approaches, based on assessment of individual topographic formations (features), are increasingly being applied to characterise complex surfaces. Feature-based characterisation consists of segmenting (partitioning) a topography in order to isolate interesting regions (features) that can then be assessed via dedicated procedures, for example via dimensional characterisation. Segmentation, and its ability to isolate the feature of interest, are at the core of any feature-based characterisation approach. In this work, three segmentation approaches are compared and validated as they are applied to the isolation of particles and spatter on various powder bed fusion surfaces. The investigated segmentation approaches are morphological segmentation on edges, contour stability analysis and active contours. A manual segmentation is performed to generate a reference result to assess the performance of the investigated segmentation methods. The methods are assessed based on identification of performance (capability of detecting the features) and accuracy of feature boundary detection (capability of identifying the correct feature boundaries). The assessment is based on computing a series of custom performance indicators developed for the purpose of the comparison and derived from the theory of binary classifiers. The proposed comparison method allows for the qualification and quantification of segmentation methods used for feature-based characterisation and can help determine the efficacy of a segmentation approach when applied to a certain test case. In future, it may be possible to use this methodology to investigate and compare how changing parameters for feature-based segmentation algorithms can result in more effective segmentation. Feature-based characterisation, topography segmentation, surface metrology, additive manufacturing
To know whether a part is fit-for-purpose; for example, to determine whether or not a shaft will ... more To know whether a part is fit-for-purpose; for example, to determine whether or not a shaft will fit within a hole, but still give enough clearance to allow the flow of lubricating fluids
Uploads
Conference Papers by Richard Leach