No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Comparison and validation of segmentation methods for feature-based characterisation of metal powder bed fusion surfaces
Abstract
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
ResearchGate has not been able to resolve any citations for this publication.
Powder bed fusion (PBF) is a popular additive manufacturing (AM) process with wide applications in key industrial sectors, including aerospace, automotive, healthcare, defence. However, a deficiency of PBF is its low quality of surface finish. A number of PBF process variables and other factors (e.g. powders, recoater) can influence the surface quality. It is of significant importance to measure and characterise PBF surfaces for the benefits of process optimisation, product performance evaluation and also product design. A state-of-the-art review is given to summarise the current research work on the characterisation of AM surfaces, particularly PBF surfaces. It is recognised that AM processes are different from conventional manufacturing processes and their produced surface topographies are different as well. In this paper, the surface characterisation framework is updated to reflect the unique characteristics of PBF processes. The surface spatial wavelength components and other process signature features are described and their production mechanisms are elaborated. A bespoke surface characterisation procedure is developed based on the updated framework. The robust Gaussian regression filter and the morphological filters are proposed to be used for the separation of the waviness component due to their robustness. The watershed segmentation is enhanced to extract globules from the residual surface. Two AM components produced by electron beam melting (EBM) and selective laser melting (SLM), are measured and characterised by the proposed methodology. Both of the two filters are qualified for the extraction of melted tracks. The watershed segmentation can enable the extraction of globules. The standard surface texture parameters of different surface wavelength components are compared. A set of bespoke parameters are intentionally developed to offer a quantitative evaluation of the globules.
The use of state-of-the-art areal topography measurement instrumentation allows for a high level of detail in the acquisition of topographic information at micrometric scales. The three-dimensional geometric models of surface topography obtained from measured data create new opportunities for the investigation of manufacturing processes through characterisation of the surfaces of manufactured parts. Conventional methods for quantitative assessment of topography usually only involve the computation of texture parameters; summary indicators of topography-related characteristics that are computed over the investigated area. However, further useful information may be obtained through characterisation of signature topographic formations, as more direct indicators of manufacturing process behaviour and performance. In this work, laser powder bed fusion of metals is considered. An original algorithmic method is proposed to isolate relevant topographic formations and to quantify their dimensional and geometric properties, using areal topography data acquired by state-of-the-art areal topography measurement instrumentation.
In this work, the performance of a focus variation instrument for measurement of areal topography of metal additive surfaces was investigated. Samples were produced using both laser and electron beam powder bed fusion processes with some of the most common additive materials: Al-Si-10Mg, Inconel 718 and Ti-6Al-4V. Surfaces parallel and orthogonal to the build direction were investigated. Measurement performance was qualified by visually inspecting the topographic models obtained from measurement and quantified by computing the number of non-measured data points, by estimating local repeatability error in topography height determination and by computing the value of the areal field texture parameter Sa. Variations captured through such indicators were investigated as focus variation-specific measurement control parameters were varied. Changes in magnification, illumination type, vertical resolution and lateral resolution were investigated. The experimental campaign was created through full factorial design of experiments, and regression models were used to link the selected measurement process control parameters to the measured performance indicators. The results indicate that focus variation microscopy measurement of metal additive surfaces is robust to changes of the measurement control parameters when the Sa texture parameter is considered, with variations confined to sub-micrometre scales and within 5% of the average parameter value for the same surface and objective. The number of non-measured points and the local repeatability error were more affected by the choice of measurement control parameters. However, such changes could be predicted by the regression models, and proved consistent once material, type of additive process and orientation of the measured surface are set.
Along with the characterisation of areal surface topography through texture parameters, surface metrology is increasingly facing challenges related to the dimensional and geometric characterisation of individual surface features. Typical scenarios range from the inspection of individual elements of micro-parts and devices, to the characterisation of pattern units in structured surfaces, and to the analysis of scratches, pores, bumps and other singularities either generated by the manufacturing process, or originated during the operational life of the surface. The characterisation of individual surface features opens up a wide array of new application scenarios and creates novel challenges for surface metrology. Early approaches are not as consolidated as what is available for the characterisation of surface texture and see the convergence of mathematical models, methods and algorithmic solutions coming from heterogeneous disciplines such as image processing, computer vision, coordinate metrology, reverse engineering, and computer-aided design. In this chapter an overview of the tools available in current surface metrology software is provided first. Then, the main challenges and open issues of achieving full metrological characterisation of individual surface features are introduced and discussed, as well as the current research approaches addressing them. © 2013 Springer-Verlag Berlin Heidelberg. All rights are reserved.
Tessellated surfaces are structured surfaces where the dominant features are organized in periodic patterns. Tessellated surfaces are becoming increasingly popular for a wide variety of industrial uses. However, their uptake is hindered by a lack of suitable metrological techniques to categorize and verify their properties. Areal surface texture parameters, commonly used for characterization of surface texture , may not provide relevant information for characterizing the periodicity and the other functionally relevant geometric attributes of the pattern. A recently proposed solution is to directly retrieve the individual features of the pattern and to determine their relevant geometric and dimensional properties. To identify individual features it is necessary to segment the surface in order to extract the pattern units (tiles) and the individual features contained within, so that their properties can be assessed. This paper reviews a number of different techniques to segment tessellated surfaces and compares their ability to accurately identify features and their boundaries. The ability to identify features is important as it has a direct impact on the computation of dimensional and geometric properties of the features.
Areal segmentation, i.e. the partitioning of areal surface topography data into regions, has recently attracted significant research interest in surface metrology. In particular morphologic segmentation , i.e. partitioning into Maxwellian hills and dales—currently the only segmentation approach endorsed by ISO specification standards—has shown potential for capturing the salient traits of a surface, so that its surface texture can be better encoded by parameters. However, recent developments in dimensional metrology applied to structured surfaces with features of dimensions on the order of micrometres (micro-electromechanical system, microfluidics, etc), and many other studies aimed at characterizing individual features in unstructured surfaces (scratches, bumps, holes, etc), are showing the importance of segmentation for extracting localized features from areal data. In this work, morphologic segmentation is applied to a selected set of case studies of industrial relevance, involving structured, semi-structured and unstructured surfaces, where the main goal is not the assessment of surface texture, but the extraction of individual surface features. The examples are designed to provide an overview of the main advantages and issues when applying morphologic segmentation in a comprehensive set of application scenarios.
To extract patterns from observable measurements we need to be able to define and identify stable features in observable measurements that persist in the presence of small artificial features such as noise, measurement errors, etc. The representational theory of measurement is used to define the stability of a measurement procedure. A technique, 'motif analysis', is defined to identify and remove 'insignificant' features while leaving 'significant' features. This technique is formalized and three properties identified that ensure stability. The connection of motif analysis with morphological closing filters is established and used to prove the stability of motif analysis. Finally, a practical metrology example is given of motif analysis in surface texture. Here motif analysis is used to segment a surface into its significant features.