Stereophotography and spatial surface reconstruction using scanning electron microscopy images

Abstract

Three-dimensional reconstruction technique to fully characterize structural performance of solid materials is suggested. The three-dimensional sample data out of the SEM images taken from different angles were extrapolated, measured and interpreted. In stereometry, the technique of three dimensional SEM imaging is fairly straightforward. Selected specimen area is photographed using SEM imaging tools from two different angles. Tilting is performed using standard SEM manipulation tools. In some cases, the specimen must be rotated to such a position, in which the tilting is done on visual ordinate axis. The resulting images are combined in pairs using any three-dimensional anaglyph software available to produce an anaglyph image, which, in turn, can be analyzed using standard 3D glasses. To achieve finer results, extrapolation of spatial data was done from three or more sample images using visual reconstruction software applications. This technique for recovering spatial data from the SEM pictures (structure-from-motion) is the VisualSFM software, which is an application for spatial reconstruction using structure from sample motion. Using VisualSFM, the images are analyzed for matching points and the camera angle is guessed for each image. Any number of additional viewports can be added to VisualSFM software. Based on this input, a surface is reconstructed where the matching points intersect and a colour value is assigned. The software produces a cloud of points, which has to be processed externally. Freely available software, such as MeshLab can be used to join the point cloud to a mesh and, as a second step of reconstruction, apply surface properties to the mesh polygons. The gold particles were selected as model material for the spatial 3D surface reconstruction.