Papers by Keyword: Multi-Scale

Abstract: A novel multi-dimension and structure element edge-detection based on mathematical morphology is presented to resolve blur problem of classical morphology when detecting an edge to reduce the noise but hard to preserve the details and edge information of the original image effectively. First,pretreatment of the image are completed by close-open operation to eliminate noise; second,do close operation to smooth image,in the end,using the operation of morphological gradient for smooting image,the ideal image edge under the environment of existing noise is obtained,and it is applied to detect the edges of welding pore images. The experimental results show that it is compared with classical Sobel operators,Canny operators and traditional edge detection algorithm, the proposed algorithm has the following distinguished advantages:accuracy of edges detected, a clear outline of the image, and can preserve more image details as well, and insensitive to noise.

Abstract: This study investigates the shape evolution of the pinholes on the bloom surfaces, which are originated from the continuous casting process, during multi-pass hot bar rolling. It is important to track the shape evolution of the pinholes since they can be formed as the sharp surface cracks after hot bar rolling and can initiate the surface bursts in the cold forging process. It is very hard to track the deformation behavior of the pinholes with detection tools during hot rolling, so the numerical simulations can be properly utilized. In general, the size of the pinholes in the bloom surface is order of micrometer although the bloom size is order of millimeter. This size discrepancy between them makes it difficult to discretize the domain including the pinholes for the finite element (FE) simulations. To overcome this limitation of the conventional FE simulation, multi-scale technique coupling the macro and micro models was developed in current study. This technique was implemented into the commercial simulation code, DEFORM-3D. The developed multi-scale simulation technique was capable of simulating the shape evolution of the pinholes through multi-pass hot bar rolling successfully. It is concluded that aspect ratio of the initial pinhole should be larger than 2.0 approximately to prevent it folded.

Abstract: An improved multi-scale segmentation algorithm is proposed in this paper. In order to get segmentation result more efficiently and accurately, watershed transformation is used as an initial segmentation algorithm, and then the objects of regions are merged based on the improved merge rule. The improved regulation for region merging is mainly based on the scale parameter of area-based while the heterogeneity parameter is considered as well. In this way, the failure of considering that some regions with large heterogeneity with their neighborhood are not suitable for merging will be prevented. Experimental results show that the quality and efficiency of remote sensing image segmentation can be greatly improved by the improved multi-scale segmentation algorithm.

Abstract: An edge detection algorithm which is applied to anterior chamber OCT images has been proposed. The algorithm firstly uses multi-structure elements to detect edge on gray level value differences on the same scale, and introduces dynamic adaptive weight to make re-fusion of pixels to gain a multi-structure element morphological edge detection image on the same scale, then confirms weight value and makes multi-scale fusion according to the noise immunity of different scale structure elements to gain the final edge detection image. The simulated results have obvious edge features,it can effectively avoid the occurrence of mutational pixels on the OCT image edge results, compared to traditional edge detection algorithms.

Abstract: Recently intensive research has been carried out on carbon nanotube (CNT) based polymer composites. However, in this work the macro scale IM700 carbon fiber (CF) has been integrated with amino modified multi-wall carbon nanotubes (MWCNT-NH2) within LY-556 epoxy matrix to produce three-phase, multi-scale composites, with the applications in missiles. The fictionalization of CNTs has been carried out for improved and consistent mechanical and physical properties. High frequency probe sonication method was used for homogenous dispersion of CNTs. Mechanical characterization of the multiscale composites fabricated by hand layup process included tensile, flexure and inter-laminar shear stress tests. The addition of small amounts of MWCNTs (upto 1.5 weight %) for the fabrication of multiscale composites resulted in a maximum enhancement in tensile strength by 23% , flexural modulus by 35%, flexural strength by 5% and ILSS by 7%. CF/epoxy composites (without CNTs) have also been characterized for comparison with MWCNT-NH₂ /CF/ epoxy composites. For a more accurate prediction of the Young’s moduli of multi-scale composites several micromechanical models (Voigt-Reuss, Halpin-Tsai, and modified mixture Law) compared with the experimental work. These models have been tested for upto 1.5 weight % of CNTs. The difference of approximate 28% of the values in Young’s modulus has been reported in the classical micromechanical models and experimental results and it comes out 11% for CF/epoxy composites as they do not have reinforcement of CNT’s within it. In all the classical models, the Young’s moduli of the carbon nanotube composites were used as matrix properties.

Abstract: Metal-supported solid oxide fuel cell (SOFC) has a varieties of potential advantages compared to the traditional ceramic supported SOFC. However, degradation issue of metal-supported SOFC is seriously impeding its further development, in particular, the inter-diffusion and interaction of iron, chromium and nickel at substrate/anode interface is known to be a key issue responsible for cell rapid degradation. With respect to the complexity and nonlinearity of degradation mechanism, multi-scale modeling and simulation is regarded as one powerful method to gain a deep insight on degradation mechanism. In present work, multi-scale models were presented to investigate multi-scale physicochemical phenomena happening at interface of anode/substrate, with the attempt to reveal degradation mechanism. The research procedure for the above goal was addressed in detail as well.

Abstract: Water quality–landscape interaction were explored using multiple line regression in watershed, total stream corridor (200-m riparian buffer extending the entire stream length upstream of the site) and the local stream corridor (200-m riparian buffer extending 1000 m upstream of the site).The results showed that the catchment landscape characteristics appeared to have slightly greater influence on water quality than the total stream corridor and local stream corridor. The edge density of construction land and percent cultivated land had the greatest influence on water quality in watershed scale. The edge density of forest land and percent forest had the greatest influence on water quality in total stream corridor scale, but percent forest and PD had the greatest influence on water quality in the local stream corridor scale.

Abstract: The insulators leakage current in transmission line can be used to characterize the filth degree of insulator surface and its accurate measurement is very important. An on-line monitor platform for insulators leakage current is introduced. Then a new multiscale self-correcting method for insulators leakage current is put forward. In this method, the leakage current information is decomposed based on multilayer and multiscale theory to gain each scale detailed signal and smooth the information in the coarsest scale at first. Furthermore, each scale detailed signal is decomposed once again and processed at different thresholds. Afterwards, the signal is reconstructed. The reconstructed detailed signal and the coarsest scale smoothed information are filtered using Kalman filter. In the end, the optimal detailed signal and the smoothed detailed signal are reconstructed and the signal is optimally estimated in the finest scale. Finally, the method was been applied to safety monitoring for a real transmission line. The results show that, compared with traditional Kalman filter method, the precision of new method raises 27%. This method has the characteristics of strong denoising ability and good robustness. It can be applied to the similar complex situation.

Abstract: This paper considers multi-scale modeling strategies for heterogeneous materials while also highlighting the problems of determining experimentally the micro-scale properties and validating such techniques. Multi-scale modeling techniques enable us to capture the influence of (evolving) heterogeneous material microstructures on the overall macroscopic behavior. This paper discusses computational multi-scale modeling techniques for problems both with and without poor scale separation. In developing these powerful multi-scale modeling techniques, the obvious challenge of validating both the material behavior at multiple scales and the associated scale transition methodologies, using advances in material characterization and experimental mechanics, comes into sharp focus and this will be briefly explored here.

Abstract: Traditional viewshed analysis method is usually based on datum plane and local space area, which can not meet the accuracy requirement of the viewshed analysis on 3D Virtual Earth based on datum sphere and large space area. Regarding this, the article discussed a new method of viewshed analysis based on 3D datum sphere, and also realized the real-time visualization of the analysis result based on the 3D Virtual Earth system GeoGlobe. Finally it inspected and verified the validity of the method.