A method to measure small local strains in concrete surfaces using its natural texture and image cross‐correlation

摘要

This paper proposes a method to measure the displacement of every point on a concrete surface using its original surface texture and digital image cross-correlation (DICC). The main advantage is that it does not need to add or attach any particular pattern or target to the surface, contrary to the current methods based in DICC. The method is tested in a concrete probe during loading-unloading cycles under its yield limit, with no cracks appearing in the surface. The total surface of the probe was tessellated in optimal local areas, and DICC with peak refinement was performed on each local area. In order to find the optimal size for that areas, we propose here a method based on the image local entropy. We show that, with natural nonoptimized textures, the proposed technique allows to locally measure the concrete deformation due to small strains (0.1%) with an error below 50 mu epsilon. Although results are more inaccurate than those obtained by a strain gauge, our method is able to noninvasively determine the strains in a concrete surface under small loads with reasonable accuracy, thus allowing its use in unreachable or historical structures where no attachment or target painting is possible.