Crosscorrelation image processing for surface shape reconstruction using fiducial markers

Abstract

Abstract Optical methods for deformations diagnostic and surface shape measurement are widely used in scientific research and industry. Most of these methods are based on triangulating a set of two-dimensional points in the images appropriate to the same three-dimensional points of the object in space. Various algorithms to search such points are applied. The possibility of using cross-correlation processing of digital images to search these points is considered in the work. Algorithms based on the correlation function calculation are widely employed in such a popular flow diagnostic method as PIV. The cameras of a stereo system for surface shape measurement can be widely spaced, and the tilt angles relative to the surface can differ significantly. This leads to the fact that the images taken from the cameras cannot be directly processed by the correlation function because it is not invariant to rotation. To solve this problem, fiducial markers are used to find an initial estimate of displacement of the images relative to each other. This approach makes it possible to successfully apply correlation processing for stereo system images with a large stereo base.