Multidimensional characterization of particle morphology and mineralogical composition using CT data and R-vine copulas
Authors:
Orkun Furat,
Tom Kirstein,
Thomas Leißner,
Kai Bachmann,
Jens Gutzmer,
Urs A. Peuker,
Volker Schmidt
Abstract:
Computed tomography (CT) can capture volumes large enough to measure a statistically meaningful number of micron-sized particles with a sufficiently good resolution to allow for the analysis of individual particles. However, the development of methods to efficiently investigate such image data and interpretably model the observed particle features is still an active field of research. When image d…
▽ More
Computed tomography (CT) can capture volumes large enough to measure a statistically meaningful number of micron-sized particles with a sufficiently good resolution to allow for the analysis of individual particles. However, the development of methods to efficiently investigate such image data and interpretably model the observed particle features is still an active field of research. When image data of particles exhibiting a wide range of shapes and sizes is considered, traditional image segmentation methods, such as the classic watershed algorithm, struggle to recognize particles with satisfying accuracy. Thus, more advanced methods of machine learning must be utilized for image segmentation to improve the validity of subsequent analyzes. Moreover, CT data does not include information about the mineralogical composition of particles and, therefore, additional SEM-EDS image data has to be acquired. In this paper, micro-CT image data of a particle system mostly consisting of zinnwaldite-quartz composites is considered. First, an image segmentation method is applied which uses deep convolutional neural networks, in particular an adaptation of the U-net architecture. This has the advantage of requiring less hand-labeling than other machine learning methods, while also being more flexible with the possibility of transfer learning. Then, fully parameterized models based on vine copulas are designed to determine multivariate probability distributions of descriptor vectors for the size, shape, texture and composition of particles -- allowing for the estimation and interpretable characterization of interdependencies between particle descriptors. For model fitting, the segmented three-dimensional CT data and co-registered two-dimensional SEM-EDS data are used.
△ Less
Submitted 18 January, 2023;
originally announced January 2023.
Description of ore particles from XMT images, supported by SEM-based image analysis
Authors:
Orkun Furat,
Thomas Leißner,
Ralf Ditscherlein,
Ondřej Šedivý,
Matthias Weber,
Kai Bachmann,
Jens Gutzmer,
Urs Peuker,
Volker Schmidt
Abstract:
In this paper, 3D image data of ore particle systems is investigated. By combining X-ray micro tomography (XMT) with scanning electron microscope (SEM) based image analysis additional information about the mineralogical composition from certain planar sections can be gained. For the analysis of tomographic images of particle systems the extraction of single particles is essential. This is performe…
▽ More
In this paper, 3D image data of ore particle systems is investigated. By combining X-ray micro tomography (XMT) with scanning electron microscope (SEM) based image analysis additional information about the mineralogical composition from certain planar sections can be gained. For the analysis of tomographic images of particle systems the extraction of single particles is essential. This is performed with a marker-based watershed algorithm and a post-processing step utilizing a neural network to reduce oversegmentation. The results are validated by comparing the 3D particle-wise segmentation empirically with 2D SEM images which have been obtained with a different imaging process and segmentation algorithm. Finally, a stereological application is shown, in which planar SEM images are embedded into the tomographic 3D image. This allows the estimation of local X-ray attenuation coefficients, which are material-specific quantities, in the entire tomographic image.
△ Less
Submitted 7 June, 2018;
originally announced June 2018.
