|M.Sc Student||Samak Dan|
|Subject||Three Dimensional Reconstruction and Visualization of|
|Department||Department of Mechanical Engineering||Supervisors||Professor Anath Fischer|
|Professor Daniel Rittel|
Fracture phenomena have been studied for centuries by many scientists. The way material is breaking, fracture behavior and its direction are resulting among other things from the environment influence and the material properties. During the centuries, techniques for measuring fractures have much evolved. Microscope discovery helped these techniques; it progressed from eye observation, to image analysis. The analysis of fracture surface imaging progressed considerably with the advent of the microscopes. Nowadays, high magnification microscopes enable us to visualize fracture surfaces at high level of details. Extraction of mechanical properties from two dimensional (2D) analysis of fracture surface microscope images are already in progress. But until now, very few researches rely on three dimensional (3D) images of fracture surface.
The objective of the research is to reconstruct the 3D topography of fracture surfaces for visualization and analysis. A method was developed that reconstructs the fracture surfaces as 3D with texture. The method is based on a stereo-photogrammetry technique. The system sets two images of a given object taken from a Scanning Electron Microscope (SEM) device. These images are oriented one with respect the other with a small tilt angle (of 5-10 degrees).
Using a stereo reconstruction algorithm which is based on the simple geometry of the image positions, the 3D points are computed. After applying a triangulation method on these points, a 3D mesh is reconstructed. Then the texture is mapped on the surface using one of the input images. The resulting texture map exhibits a very realistic 3D surface. This fracture surface provides a variety of details about the failure process, for example, the roughness of the surface following a direction. Superimposition of two surfaces composing the fracture was also developed. It enables a great 3D realistic visualization of the fracture and permits the fracture analyst to understand the behavior of the fracture. The performance of the proposed method have been analyzed and demonstrated on geometric validation cases.