Search

In this study, images of the rails captured by a track recording car are investigated. The images are processed in order to asses and predict the rail quality by a simple method, which gives the possibility of fast and easy implementation. The first step of the method is detecting the rails from the images with an algorithm. The second step of the method is finding visually noticeable defects on the detected rail with automatized algorithms for defect detection. One of the algorithms for defect detection shows promising results and investigating the rail images could be a promising complementary method to the already used manual ultrasound measurements.

Understanding the diversity of eyes is crucial to unravel how different animals use vision to interact with their respective environments. To date, comparative studies of eye anatomy are scarce because they often involve time-consuming or inefficient methods. X-ray micro-tomography is a promising high-throughput imaging technique that enables to reconstruct the 3D anatomy of eyes, but powerful tools are needed to perform fast conversions of anatomical reconstructions into functional eye models. We developed a computing method named InSegtCone to automatically segment the crystalline cones in the apposition compound eyes of arthropods. Here, we describe the full auto-segmentation process, showcase its application to three different insect compound eyes and evaluate its performance. The auto-segmentation could successfully label the full individual shapes of 60%-80% of the crystalline cones, and is about as accurate and 250 times faster than manual labelling of the individual cones. We believe that InSegtCone can be an important tool for peer scientists to enable extensive comparisons of the diversity of eyes and vision in arthropods.

In high temperature polymer electrolyte fuel cells, polybenzimidazole membranes doped in phosphoric acid are used as electrolyte material. The membrane thickness directly relate to the amount of absorbed acid which relate to performance. In this study, we compare scanning electron microscopy and X-ray micro-computed tomography (CT) regarding suitability for determining the structure of electrolyte membranes. Semi-automated layer identification and thickness estimation was used to reduce human errors and data processing time. Scanning electron microscopy was found reliable for membrane thickness characterization despite necessary destructive sample preparation. With X-ray CT it was possible to study the cells non-destructively before and after cell test. This made it possible to identify, so-called hot pressing, as the step in which the membrane lost most of its thickness. After cell operation, the use of X-ray CT at large field of view and a recently developed layer detection algorithm made it possible to visualize compression of the membrane in a pattern identical to that of the flow plate channels. This compression pattern would have been difficult to determine with conventional electron microscopy or X-ray CT without semi-automated layer detection.

A unidirectional (UD) glass fibre reinforced polymer (GFRP) composite was scanned at varying resolutions in the micro-scale with several imaging modalities. All six scans capture the same region of the sample, containing well-aligned fibres inside a UD load-carrying bundle. Two scans of the cross-sectional surface of the bundle were acquired at a high resolution, by means of scanning electron microscopy (SEM) and optical microscopy (OM), and four volumetric scans were acquired through X-ray computed tomography (CT) at different resolutions. Individual fibres can be resolved from these scans to investigate the micro-structure of the UD bundle. The data is hosted at https://doi.org/10.5281/zenodo.1195879 and it was used in Emerson et al. (2018) [1] to demonstrate that precise and representative characterisations of fibre geometry are possible with relatively low X-ray CT resolutions if the analysis method is robust to image quality. Keywords: Geometrical characterisation, Polymer-matrix composites (PMCs), Volumetric fibre segmentation, Automated fibre tracking, X-ray imaging, Microscopy, Non-destructive testing

Imaging with X-ray computed tomography (CT) enables non-destructive 3D characterisations of the micro-structure inside fibre composites. In this paper we validate the use of X-ray CT coupled with image analysis for characterising unidirectional (UD) fibre composites. We compare X-ray CT at different resolutions to optical microscopy (OM) and scanning electron microscopy (SEM), where we characterise fibres by their diameters and positions. In addition to comparing individual fibre diameters, we also model their spatial distribution, and compare the obtained model parameters. Our study shows that X-ray CT is a high precision technique for characterising fibre composites and, with our suggested image analysis method for fibre detection, high precision is also obtained at low resolutions. This has great potential, since it allows larger fields of view to be analysed. Besides analysing representative volumes with high precision, we demonstrate that based on our methodology for individual fibre segmentation it is now possible to study complete bundles at the fibre scale and reveal inhomogeneities in the physical sample.

Important properties of fiber-reinforced composites, such as stiffness, compression strength, and fatigue resistance, are sensitive to fiber alignment. In this paper, we use structure tensor analysis on CT images to characterize the fiber orientations in three samples of unidirectional fiber-reinforced composites: pultruded carbon, pre-preg carbon, and non-crimp glass fiber fabric. Our results show that the fibers in the pultruded sample are more aligned than fibers in the two other samples. Through local quantitative analysis, we show that misalignment of the individual pre-preg layers contributes to the overall fiber misalignment in the material. For the non-crimp composite, we show that both the stitching of the unidirectional bundles and the backing bundles affect the fiber alignment in unidirectional bundles. Quantification of the misalignment caused by these effects allows manufacturers to tune production parameters, such as stitching thread tension, to minimize the misalignment of the fibers. All our notebooks, code, and data are available online.

The mechanical properties of composite fiber materials are highly dependent on the orientation of the fibers. Micro-CT enables acquisition of high-resolution 3D images, where individual fibers are visible. However, manually extracting orientation information from the samples is impractical due to the size of the 3D images. In this paper, we use a Structure Tensor to extract orientation information from a large 3D image of non-crimp glass fiber fabric. We go through the process of segmenting the image and extracting the orientation distribution step-by-step using structure tensor and show the results of the analysis of the studied non-crimp fabric. The Jupyter notebooks and Python code used for the data-analysis are publicly available, detailing the process and allowing the reader to use the method on their own data. The results show that structure tensor analysis works well for determining fiber orientations, which has many useful applications.

A deeper knowledge of the architecture of the peripheral nerve with three-dimensional (3D) imaging of the nerve tissue at the sub-cellular scale may contribute to unravel the pathophysiology of neuropathy. Here we demonstrate the feasibility of X-ray phase contrast holographic nanotomography to enable 3D imaging of nerves at high resolution, while covering a relatively large tissue volume. We show various subcomponents of human peripheral nerves in biopsies from patients with type 1 and 2 diabetes and in a healthy subject. Together with well-organized, parallel myelinated nerve fibres we show regenerative clusters with twisted nerve fibres, a sprouted axon from a node of Ranvier and other specific details. A novel 3D construction (with movie created) of a node of Ranvier with end segment of a degenerated axon and sprout of a regenerated one is captured. Many of these architectural elements are not described in the literature. Thus, X-ray phase contrast holographic nanotomography enables identifying specific morphological structures in 3D in peripheral nerve biopsies from a healthy subject and from patients with type 1 and 2 diabetes. Funding Agencies|EU InterReg project MAX4ESSFUN [KU-008]; Swedish Research CouncilSwedish Research Council [E0605401, E0605402]; Lund University

We present a novel method for characterizing the microstructure of a material from volumetric datasets such as 3D image data from computed tomography (CT). The method is based on a new statistical model for the distribution of voxel intensities and gradient magnitudes, incorporating prior knowledge about the physical nature of the imaging process. It allows for direct quantification of parameters of the imaged sample like volume fractions, interface areas and material density, and parameters related to the imaging process like image resolution and noise levels. Existing methods for characterization from 3D images often require segmentation of the data, a procedure where each voxel is labeled according to the best guess of which material it represents. Through our approach, the segmentation step is circumvented so that errors and computational costs related to this part of the image processing pipeline are avoided. Instead, the material parameters are quantified through their known relation to parameters of our model which is fitted directly to the raw, unsegmented data. We present an automated model fitting procedure that gives reproducible results without human bias and enables automatic analysis of large sets of tomograms. For more complex structure analysis questions, a segmentation is still beneficial. We show that our model can be used as input to existing probabilistic methods, providing a segmentation that is based on the physics of the imaged sample. Because our model accounts for mixed-material voxels stemming from blurring inherent to the imaging technique, we reduce the errors that other methods can create at interfaces between materials., Comment: Manuscript accepted for publication in Materials Characterization

Optical coherence tomography angiography (OCTA) is a rapidly developing technique which generates angiographic images non-invasively. This study proposes a method to determine the vessel density in OCTA scans in general and especially the local peripapillary vessel density. The method produces vessel density heatmaps that contain information about the local vessel density. We apply the method in a small study to demonstrate its applicability and its potential to investigate the influence of local conditions on the vessel density. In the study, we combine information from enhanced depth imaging optical coherence tomography (EDI-OCT) about the location of optic disc drusen (ODD) with information from OCTA about the vessel density. We see a reduction in local peripapillary vessel density in peripapillary sections with a large volume of ODD.

Morphological analysis of the left atrial appendage is an important tool to assess risk of ischemic stroke. Most deep learning approaches for 3D segmentation is guided by binary labelmaps, which results in voxelized segmentations unsuitable for morphological analysis. We propose to use signed distance fields to guide a deep network towards morphologically consistent 3D models. The proposed strategy is evaluated on a synthetic dataset of simple geometries, as well as a set of cardiac computed tomography images containing the left atrial appendage. The proposed method produces smooth surfaces with a closer resemblance to the true surface in terms of segmentation overlap and surface distance., Comment: MIDL 2019 [arXiv:1907.08612]

The present study investigates the use of close-range photogrammetry for automated air void analysis of hardened concrete without contrast enhancement. First, polished plane sections of concrete with varying air contents were analyzed by photogrammetry. The data in the form of point clouds were processed to identify the air voids. Second, the same plane sections were subjected to traditional air void analysis by contrast enhancement. Results of both methods were compared. The method based on photogrammetry showed potential, but it could not capture the smallest air voids. Better camera resolution and/or methods to reduce noise in the point cloud are needed.

Background Synchrotron X-ray computed tomography (SXCT) allows for three-dimensional imaging of objects at a very high resolution and in large field-of-view. Purpose The aim of this study was to use SXCT imaging for morphological analysis of muscle tissue, in order to investigate whether the analysis reveals complementary information to two-dimensional microscopy. Methods Three-dimensional SXCT images of muscle biopsies were taken from participants with cerebral palsy and from healthy controls. We designed morphological measures from the two-dimensional slices and three-dimensional volumes of the images and measured the muscle fibre organization, which we term orientation consistency. Results The muscle fibre cross-sectional areas were significantly larger in healthy participants than in participants with cerebral palsy when carrying out the analysis in three dimensions. However, a similar analysis carried out in two dimensions revealed no patient group difference. The present study also showed that three-dimensional orientation consistency was significantly larger for healthy participants than for participants with cerebral palsy. Conclusion Individuals with CP have smaller muscle fibres than healthy control individuals. We argue that morphometric measures of muscle fibres in two dimensions are generally trustworthy only if the fibres extend perpendicularly to the slice plane, and otherwise three-dimensional aspects should be considered. In addition, the muscle tissue of individuals with CP showed a decreased level of orientation consistency when compared to healthy control tissue. We suggest that the observed disorganization of the tissue may be induced by atrophy caused by physical inactivity and insufficient neural activation.

The sudden compressive failure of unidirectional (UD) fibre reinforced composites at loads well below their tensile strengths is a cause of practical concern. In this respect and more generally, analytical and numerical models that describe composite behaviour have been hard to verify due to a lack of experimental observation, particularly in 3D. The aim of this paper is to combine fast in-situ X-ray computed tomography (CT) with advanced image analysis to capture the changes in fibre orientation in 3D during uninterrupted progressive loading in compression of a UD glass fibre reinforced polymer (GFRP). By analysing and establishing correspondence between a sequence of time-lapse X-ray CT images of the composite, we are able for the first time to follow each fibre and quantify the progressive deflection that takes place during axial compression in the steps leading up to fibre micro-buckling and kinking. Even at just 25 % of the failure load, fibres have started to tilt in approximately the direction of the ultimate kink band. The rate of tilting increases as the composite approaches the collapse load. More generally, our approach can be applied to investigate the behaviour of a wide range of fibrous materials under changing loading conditions.

This paper proposes the cache-mesh, a dynamic mesh data structure in 3D that allows modifications of stored topological relations effortlessly. The cache-mesh can adapt to arbitrary problems and provide fast retrieval to the most-referred-to topological relations. This adaptation requires trivial extra effort in implementation with the cache-mesh, whereas it may require tremendous effort using traditional meshes. The cache-mesh also gives a further boost to the performance with parallel mesh processing by caching the partition of the mesh into independent sets. This is an additional advantage of the cache-mesh, and the extra work for caching is also trivial. Though it appears that it takes effort for initial implementation, building the cache-mesh is comparable to a traditional mesh in terms of implementation.

PURPOSE. Optic disc drusen (ODD) are found in up to 2.4% of the population and are known to cause visual field defects. The purpose of the current study was to investigate how quantitatively estimated volume and anatomic location of ODD influence optic nerve function. METHODS. Anatomic location, volume of ODD, and peripapillary retinal nerve fiber layer and macular ganglion cell layer thickness were assessed in 37 ODD patients using enhanced depth imaging optical coherence tomography. Volume of ODD was calculated by manual segmentation of ODD in 97 B-scans per eye. Anatomic characteristics were compared with optic nerve function using automated perimetric mean deviation (MD) and multifocal visual evoked potentials. RESULTS. Increased age (P = 0.015); larger ODD volume (P = 0.002); and more superficial anatomic ODD location (P = 0.007) were found in patients with ODD visible by ophthalmoscopy compared to patients with buried ODD. In a multivariate analysis, a worsening of MD was significantly associated with larger ODD volume (P < 0.0001). No association was found between MD and weighted anatomic location, age, and visibility by ophthalmoscopy. Decreased ganglion cell layer thickness was significantly associated with worse MD (P = 0.025) and had a higher effect on MD when compared to retinal nerve fiber layer thickness. CONCLUSIONS. Large ODD volume is associated with optic nerve dysfunction. The worse visual field defects associated with visible ODD should only be ascribed to larger ODD volume and not to a more superficial anatomic ODD location.

We describe an efficient algorithm that computes a segmented reconstruction directly from x-ray projection data. Our algorithm uses a parametric curve to define the segmentation. Unlike similar approaches which are based on level-sets, our method avoids a pixel or voxel grid; hence the number of unknowns is reduced to the set of points that define the curve, and attenuation coefficients of the segments. Our current implementation uses a simple closed curve and is capable of separating one object from the background. However, our basic algorithm can be applied to an arbitrary topology and multiple objects corresponding to different attenuation coefficients in the reconstruction. Through systematic tests we demonstrate a high robustness to the noise, and an excellent performance under a small number of projections.

We have developed an image analysis pipeline1 that can extract individual fibre tracksfrom low contrast X-ray tomograms of unidirectional composites with high fibre volumefraction. Measuring individual fibre tracks opens up the possibility of modelling thisempirical data in a statistical manner. Thus, allowing to analyse the spatial distributionsof the parameters characterising the orientation and curvature of these individual fibres,which can also provide insights on the interactions amongst the individual fibres.Finite element models (FEMs) can be built from the extracted geometry to simulatethe performance of the scanned fibre structure under realistic conditions. Moreover, as-pects of the fibre architecture that inuence the macroscopic behaviour of the compositecan be quantified. Examples are 2D FEMs to predict the transverse stifiness or the quantification of fibre orientations to estimate the compression strength.1 And last butnot least, already developed analytical and numerical models to describe the composite'sbehaviour can be validated against the observed data.

In life science and material science, it is often desirable to segment a volumetric data set in such a way that multiple materials (phases) are segmented and a tetrahedral mesh representation is obtained for each segment for downstream applications. Unfortunately, obtaining a mesh, typically from CT or MRI scan, is challenging, especially in 3D. This paper proposes an novel approach for volume segmentation using a tetrahedral mesh. Our method employs a deformable model that minimizes the Mumford-Shah energy function. We apply our method to several CT data sets in order to demonstrate its advantages: multi-phase support, robustness to noise, and adaptive resolution outputs. Our method is based on the Deformable Simplicial Complex (DSC) method for tracking deformable interfaces which is designed specifically to deal with topology changes.

We propose a method for retrieving a piecewise smooth surface from noisy data. In data acquired by a scanning process sampled points are almost never on the discontinuities making reconstruction of surfaces with sharp features difficult. Our method is based on a Markov Random Field (MRF) formulation of a surface prior, with the surface represented as a collection of small planar patches, the surfels, associated with each data point. The main advantage of using surfels is that we avoid treating data points as vertices. MRF formulation of the surface prior allows us to separately model the likelihood (related to the mesh formation process) and the local surface properties. We chose to model the smoothness by considering two terms: the parallelism between neighboring surfels, and their overlap. We have demonstrated the feasibility of this approach on both synthetical and scanned data. In both cases sharp features were precisely located and planar regions smoothed.

X-ray computed tomography (CT) is a powerful tool for characterising materials for its ability to reveal their internal structure in a non-destructive manner. The recent advances in Xray imaging have brought high-resolution X-rays to laboratory sources, making this tool available to a broader public. Additionally, thanks to the developments in ultra-fast X-ray imaging at synchrotron beamlines, it is now possible to capture the very fast structural changes inside materials under realistic working conditions, e.g. in operation or under loading. There is a need for advanced image analysis methods that can exploit the information contained in these 3D and 4D data-sets of high spatial and temporal resolution, which often contain image artefacts and noise. We have developed a method to characterise the geometry of materials reinforced with long fibres [1], such as glass and carbon fibre reinforced polymers. The method is based on segmenting individual fibres and the task is specially challenging when the image is noisy and its resolution is limited, because the fibres are densely packed. A limited spatial resolution might arise from the need of performing fast scans, to capture the sudden micro-structural changes that happen when reaching the composite’s collapse load, and will facilitate scanning large fields of view containing many fibres, necessary to ensure representative characterisations of a material’s micro-structure. Due to the robustness of our method to image quality [2], we have been able to characterise fibre orientations and diameter distributions in complete bundles, relevant for investigating the effect of the design and manufacturing processes on the mechanical properties of the materials. Moreover, we have applied our methodology to study the behaviour of a fibre composite under compressive loading. Following the changes in each individual fibre under progressive loading conditions, and correlating these with the initial structure of the material, can reveal the precursors to the very complex damage mechanisms that affect fibre composites.

Insegt Fibre is a tool for volumetric fibre segmentation, that is appropriate for densely packed fibres. The robustness to image quality of the underlying dictionary-based algorithm has opened up new opportunities for micro-structural characterisation of fibre-reinforced polymers. The recent user-friendly version of the robust fibre-detection algorithm makes image-based characterisation of fibrous materials simpler and more accessible.

The current study investigates the applicability of a dictionary-based fiber-detection algorithm for continuous textured glass fiber yarns in fiber-reinforced polymer composites. Additionally, the study discusses options for improving the fiber-detection algorithm so that it is applicable to continuous textured fiber yarns in fiber-reinforced polymers.

In this paper we propose a novel anisotropic smoothing scheme based on Markov Random Fields (MRF). Our scheme is formulated as two coupled processes. A vertex process is used to smooth the mesh by displacing the vertices according to a MRF smoothness prior, while an independent edge process labels mesh edges according to a feature detecting prior. Since we should not smooth across a sharp feature, we use edge labels to control the vertex process. In a Bayesian framework, MRF priors are combined with the likelihood function related to the mesh formation method. The output of our algorithm is a piecewise smooth mesh with explicit labelling of edges belonging to the sharp features.

Volume segmentation is efficient for reconstructing material structure, which is important for several analyses, e.g. simulation with finite element method, measurement of quantitative information like surface area, surface curvature, volume, etc. We are concerned about the representations of the 3D volumes, which can be categorized into two groups: fixed voxel grids [1] and unstructured meshes [2]. Among these two representations, the voxel grids are more popular since manipulating a fixed grid is easier than an unstructured mesh, but they are less efficient for quantitative measurements. In many cases, the voxel grids are converted to explicit meshes, however the conversion may reduce the accuracy of the segmentations, and the effort for meshing is also not trivial. On the other side, methods using unstructured meshes have difficulty in handling topology changes. To reduce the complexity, previous methods only represent the surfaces, thus they only segment a single region without exterior or interior information (e.g. holes). Finally, yet importantly, previous methods of both representations have issues with multi-material segmentation, where vacuum and overlapping between surfaces occur. This paper proposes a method for volume segmentation using a tetrahedral mesh. The compelling advantages of our method include: natural multi-material support; output is tetrahedral mesh that can be utilized for simulation and analysis directly; and the ability to control the resolution for compact meshes. We are also experimenting to prove our advantages on high accuracy; and the potentiality to accompany shape prior information during segmentation.

We present an efficient image analysis pipeline that enables us to reveal white matter organization in high-resolution 3D non-MRI structural datasets, in cases where a strict image segmentation is not required nor possible. We apply the method to a synchrotron X-ray holographic tomography scan from a healthy mouse sample, and show the organization of axon bundles in a region covering parts of the corpus callosum and the cingulum. The method has a potential to improve our general understanding of white matter organization and our ability to generate realistic phantoms for validation of microstructure modelling from low-resolution diffusion MRI scans.

This is a study of a uni-directional non-crimp fabric reinforced epoxy composite material typically used as the load carrying laminate in wind turbine blades. Based on a 3D xraytomography scan, the bundle and fibre/matrix structure of the composite is segmented. This segmentation is used in a multi-scale finite element model bridging the gap from theindividual fibers organized in bundles to the stitched non-crimp fabric used for building up the load carrying laminates.

We present the first experimental observation and quantification of the evolution of the three-dimensional (3D) deflection of fibres leading up to kink-band formation in notched unidirectional glass fibre-epoxy composites. This has been achieved by in situ synchrotron X-ray computed tomography (CT) quantified by advanced image analysis. The 3D trajectories of individual fibres were extracted to monitor the change in fibre profiles with increasing compressive load. Initially the fibre trajectory varies significantly from fibre to fibre, the waviness of which cannot be represented by a simple idealised profile. However, the change in the fibre shapes during axial compression is similar for all fibres, as the fibres deflect towards the final kinking direction. The precursor to kink-band nucleation has been identified as a local tilting of fibres in the region where the kink bands ultimately form, defining a micro-buckle band. Even at 99.9% of the failure load this micro-buckle band is much narrower (~150 μm (~12 fibre diameters) wide), and the angle of inclination of the band much shallower (~8°), compared to the final 300–400 μm (25–33 fibre diameters) wide, and 25–30° inclined, kink band. It is notable that kink-band boundaries are delineated by fibre fractures, many of which exhibit wedge-shaped multiple fractures.