Study On The Application Of 3D Image Processing And Analysis For Nano-CT

After the discovery of X-ray, the X-ray imaging technique was quickly developed. X-ray imaging is the best tool to observe the internal structures of thick samples with high-resolution and non-destruction benefited by the short wave length and strong penetration of X-ray, so it has been widely used in the industry and medicine. Due to the development of synchrotron radiation source, high-speed and high-sensitivity X-ray detector, high-precision control systems and micro-/nano- fabrication technology in the past three decades, nano-CT microscopy has developed rapidly. National Synchrotron Radiation Laboratory (NSRL) has built a nano-CT beamline based on the capillary condenser and Fresnel zone plate, which has 50 nm spatial resolution and Zernike phase contrast imaging capability. The nano-CT technology has broad potential applications in the microelectronics industry, biomedical, energy and environmental science, materials science and other fields.The main contributions of this thesis are summarized as follows:1. Building and testing of the nano-CT systemThe principle and the main design parameters of optical devices for the nano-CT beamline and workstation were introduced. The works of hardware and software construction and the system test were involved and several important performance results such as spatial resolution, image contrast and CT capability of the nano-CT system were obtained. The results show that the performance of the system is as good as the design targets.2. Development of the imaging experiment methods for large samplesThe rotation stage alignment method according to the feature of existing rotation sample stage was derived, to keep the interesting sample region in the center of view during CT imaging. Based on the alignment method two different imaging experiment methods for large samples, mosaic-CT algorithm and 3D volume data stitching algorithm, were developed. The CT data acquisition and processing modules were also designed using the existing script interface of the TXMcontroller software. The experimental results were carried out to verify and compare these two kinds of methods, and their advantages and shortcomings were also discussed. These results show that the methods for large sample imaging are useful to get the entire 3D structures of some large samples, which are good extension for the existing common CT imaging method.3. Development of 3D image processing and analysis methodsA series of corresponding 3D image processing and analysis methods based on the nano-CT system and the characteristics of various samples were developed.Several important image pre-processing algorithms were proposed to improve the quality of original radiographs. The interface of reading and writing for the original Xradia data file was studied; An iterative denoise algorithm based on the strongest random noise in our system was proposed to effectively eliminate the noise and preserve the details of sample characteristics; A frequency enhancement algorithm was proposed to enhance the sample characteristics of TXM images with better visibilities and contrast as well as the relatively homogeneous flat field background.Several 3D image reconstruction algorithms were introduced. The convolution back-projection algorithm and algebraic reconstruction technique were applied to reconstruct and compare both simulated and experimental data. The results show different influence caused by the different CT parameters. And their advantages and disadvantages were also discussed.Two different 3D image segmentation methods according to the actual samples, adaptive iterative algorithm for optimal threshold and edge detection algorithm were proposed, to segment different 3D volume data with different structural characteristics. The methods were applied and good results were obtained. The results show that based on the sample characteristics, combining different methods is an effective and useful solution for 3D image segmentation.Several methods were proposed to analyze nano-CT results, including 3D structure characterization, measurement and rendering, quantitative calculation for the spatial structure parameters such as volume, surface area and porosity. Especially for the samples of porous structures, the pore connectivity and three-phase boundary (TPB) analysis and characterization methods were proposed to analyze the 3D structures of solid oxide fuel cell electrodes.4. Study of the applications of nano-CT in the field of nano-materials, cell imaging and porous electrode The 3D structures of the snow-like fractal copper sulphide crystal with complex surface morphology were imaged and analyzed. The complicated 3D interior structures of the polycrystalline hollow zinc oxide microspheres were clearly obtained and the parameters such as diameter, volume, porosity, and surface area were calculated by the quantitative analysis of reconstructed data. Especially, three single selected particles with different typical structures were separated and compared. The results reveal that nano-CT is an effective and competent tool for investigating the 3D structures of nanomaterials in the natural environment.Combining 3D volume data stitching algorithm with nano-CT, large filed of view with high spatial resolution has been achieved for non-destructive 3D visualization of clustered yeasts that were too large for a single scan. Different subcellular structures were also quantitative analyzed. It shows high promise for imaging other large samples in future.The 3D microstructures of solid oxide fuel cells (SOFC) electrodes were investigated using proposed porous structure analysis methods. The pore connectivity of cathode and the three-phase connectivity and boundary of anode were quantitative analyzed. The results provide high promise for the further study in SOFC using nano-CT technique.