Three-dimensional Reconstruction Of Leather Fibers Woven Structure

Leather is a natural polymer biomaterial made from natural animal skins via a series of chemical and physical processes. It has an intricate three-dimensional(3D) network tightly woven by natural collagenous fiber bundles in 3D space, compounded with tanning agent, retanning agent and fatliquoring agent etc. The collagenous fiber woven structure is the structure foundation of the leather physical and mechanical performance, and has a critical guiding significance to the development of leather making technology. However, there are very few reports on the 3D leather structure. As reported previously, collagenous fiber woven structure was deduced from observing and analyzing leather sectional images, which is inaccurate and unscientific and can not reveal the true woven structure completely owing to its structural complex characteristic of 3D form.In this thesis, 3D reconstruction technology which had widely used in the field of biology, medicine and materials was introduced. The technology of collagenous fiber woven structure reconstruction was developed and the 3D digital models of collagenous woven structure were obtained. It will lay a key foundation for further research on structure-performance relations of leather.The main contents include the following aspects:1. Obtaining of collagen fiber cross section images. Metallographic sample preparation method is an important method on phase structure research of alloy materials. A new technique of metallographic sample preparation for obtaining the section images of leather and other soft materials had been developed. The metallographic sample preparation steps was Embedding—Mounting—Grinding—Polishing —Microscopic Examination. The embedding resin type, gel time and temperature, post-cure temperature and time, polishing time and microscope model which affected the image quality of the leather section were investigated. The optimized operation conditions were: E-51 epoxy resin was selected as embedding medium; the leather was saturated by resin; the gelation time was 30 minutes, room temperature; the post-cure treatment was then at 70oC for 3 hours; the polishing time was about 10 seconds(150r/min). light reflection mode and coaxial light model were suitable for the observation.2. 3D reconstruction of collagenous fiber woven structure. In order to meet the demands of 3D-Doctor software, the interlamellar spacing of two adjacent successive serial images must be equal. The grinding time was forecasted by studying the relation between gerinding time and the depth of material to be removed. A micrometer caliper(a precision of 1mm) was used to measure the depth of material removed during grinding at four fixed points near the leather sample. The specimen to be removed between adjacent sequence sections was controlled to be 15 ±1mm by artificial grinding method. 3D reconstruction was completed via the connection of the fiber bundle section boundaries. In order to make the image boundaries more clear, The image preprocessing method includes grayscale transformation, histogram equalization, median filtering were adopted to sharpen the boundaries and remove noises.The applicability and accuracy of different image registration method were investigated for images with different quality. Softwares for automatic, semi-automatic and manual registration were developed by our research team according to the need of different characteristics sequence images registration. The individual fiber bundle can be extracted from the 3D reconstructed model and be measured individually by defining each fiber bundle as a target object with a certain color which was different from the other objects around it in 3D-Doctor software.3. 3D reconstruction of collagenous fibers woven structure based on micro-CT. 3D image can be built by micro CT based on the different absorption to the X-ray of every parts whose density is different. The 3D image of a leather sample was obtained by this method. The instrument scanning parameters and the digital slice way were investigated to obtain the leather section image sequence. The fiber boundaries in micro-CT images were extracted through the threshold boundary extraction method combining with manually adjustment. By comparing with the metallographic method, the resolution of micro-CT images was lower than the metallographic method. But the advantages of the micro-CT method were that more sequence images with small layer spacing were easily obtained and the reconstruction fiber bundles with no ladder, and can better show their form and extension directions. This method is suitable for investigating the woven structure. The fiber bundle morphologies were showed and researched in detail. The bending angles of individual fiber bundle were measured and analyzed, and was found that concentrated in 100-140 degrees4. Research on the accuracy of 3D digital model and the fractal characteristic of collagenous fibers woven structure. The void fraction calculated from the 3D leather digital models with a magnification(objective) of 5× and 10× were 40.3% and 39.6% respectively. The related void fraction of the leather smple measured by mercury intrusion method was 38.4%. The accuracy of the 3D digital model was validated through the conformity of the void fraction calculated from the 3D digital models and measured by mercury intrusion method. Based on the research on the characteristic of the fiber bundle morphologies, woven structure and bending angles, the fractal characteristic of fiber bundles was determined. The fractal dimensions of the fiber bundle sections were calculated by a slit island method( L1/D∝A1/2) combined with their circumferences and areas in sequence images. The fractal dimensions were 1.240 and 1.294 in the images obtained with the magnification(objective) of 5× and 10×respectively. Combined the relations( S1/D ∝ V1/3) between the fractal surface and the space region circled by the fractal surface, the fractal dimensions of the fiber bundle surface were calculated, which were 2.534 and 2.586 in the leather 3D digital models obtained with the magnification(objective) of 5× and 10× respectively.