Content:
The intrinsic structure significantly influences the mechanical properties of leather. In consequence, knowledge of leather’s hierarchical structure is essential in order to find the most suited leather for specific
application. Leather structure based parameters are of major importance for both manufacturing and leather processing industries. In this respect, intensive structure investigations have been subjected in
continuous research work.
Quantitative image analysis combined with stochastic micro-structure modelling and numerical simulation of macroscopic properties is a promising approach to gain a deeper understanding of complex relations between material’s micro-structure geometry and macroscopic properties. Key ingredient is a reliable geometric description provided by the quantitative analysis of 3D images of the material micro-structures. For leather, both imaging and image analysis are particularly challenging, due to the multi-scale nature of the leather’s micro-structure. Scales in leather are not well separated. Previously, high resolution computed tomography allowed 3D imaging of purely vegetable tanned leather samples at micro- and submicro- scale. Segmentation of leather structure as well as of typical structural elements in resulting image data is however hampered by a strong heterogeneity caused by lower scale structural information.
The first method for automatic segmentation of typical structural elements at varying scales combined morphological smoothing with defining and iteratively coarsening regions using the waterfall algorithm on
local orientations. It yields a hierarchical segmentation of the leather into coarse and fine structural elements that can be used to analyze and compare the structure of leather samples. Size and shape of the
structural elements as well as their sub-structure yield information, e. g. on undulation, branching, thickness, cross-sectional shape, and preferred directions.
In order to compare the micro-structure of leather samples from various body parts or even species, the segmentation has to be applicable without extensive pre-processing and parameter tuning. Robustness can be gained by applying smoothing methods that are adapted to the goal of defining image regions by similar local orientation. The challenge is that the space of fiber orientations in 3D is not equipped with an order. Motivated by a recent approach for nevertheless defining erosion and dilation on the sphere, we suggest new definitions for these morphological base transformations on the space of directions in 3D. We present segmentation results for 3D images of leather samples derived by these new morphological smoothing methods.
Take-Away:
The intrinsic structure significantly influences the mechanical properties of leather.
Leather’s hierarchical structure can be analyzed by quantitative 3D image analysis combined with stochastic micro-structure modelling.
Segmentation results for 3D images of leather samples derived by new morphological smoothing methods.
| Identifer | oai:union.ndltd.org:DRESDEN/oai:qucosa:de:qucosa:34193 |
| Date | 28 June 2019 |
| Creators | Dietrich, Sascha, Schulz, H., Hauch, K., Schladitz, K., Godehardt, M., Orlik, J., Neusius, D. |
| Contributors | International Union of Leather Technologists and Chemists Societies |
| Publisher | Verein für Gerberei-Chemie und -Technik e. V., Forschungsinstitut für Leder und Kunststoffbahnen (FILK) gGmbH |
| Source Sets | Hochschulschriftenserver (HSSS) der SLUB Dresden |
| Language | English |
| Detected Language | English |
| Type | info:eu-repo/semantics/publishedVersion, doc-type:conferenceObject, info:eu-repo/semantics/conferenceObject, doc-type:Text |
| Rights | info:eu-repo/semantics/openAccess |
| Relation | urn:nbn:de:bsz:14-qucosa2-340872, qucosa:34087 |
Page generated in 0.0027 seconds