Determination of in-plane and off-plane surface displacements with grids virtually applied to digital images

Digital image devices have experienced an enormous increment in their capabilities, associated with a significant reduction in the economic effort. In particular, the increasing number of pixel made available for each picture allows developing software that is able to perform precise surface characterizations. In the present paper the interest is oriented into two directions. The firs one concerns the detecting the geometric features of surfaces through digital image comparison. The method does not require stereo image processing but it is based on a single camera vision. The base of this first part of the work regards the displacements of a grid virtually applied on the surface. To this goal the real printed grid case is firstly discussed. The grid virtually attached to the pictures identifies a finite element mesh associated to the comparing images. The second part aims to evaluate surface strains experienced on the specimen surface. The algorithm performs the analysis of the two comparing images, before and after the application of loads. The method is based on differential calculation of the Jacobian matrix and the solution of an over-determined system of equations. The differential calculation is truncated at the first order and iteratively applied. Two different strategies are proposed: a partial grouping of pixels by equation averaging; the use of Hu’s invariants applied to sub-images. These two approaches are discussed and compared.