Optics and Lasers in Engineering 115 (2019) 208–216 Contents lists available at ScienceDirect Optics and Lasers in Engineering journal homepage: www.elsevier.com/locate/optlaseng Digital Image Correlation for discontinuous displacement measurement using subset segmentation Ghulam Mubashar Hassan Department of Computer Science & Software Engineering, University of Western Australia, Australia a r t i c l e i n f o Keywords: Digital image correlation Deformation measurement DIC Subset segmentation Discontinuous displacement measurement Reconstruction of displacement fields a b s t r a c t Deformation measurement is normally achieved by using Digital Image Correlation (DIC) technique when de- formation is not discontinuous. However, the presence of discontinuities makes the deformation process very challenging and DIC fails. An innovative technique is proposed in this study which splits the subset (segment) of an image into multiple parts and use segmented subset of the image for correlation process. The performance of the proposed technique is evaluated using different experiments where different types of discontinuities are introduced in the deformation process at different angles and having different discontinuity opening sizes. The obtained results are compared with the recently proposed Discontinuous Digital Image Correlation (DDIC) tech- nique. The results show that the proposed technique is more reliable and having high accuracy which reaches upto 1/100th of a pixel under favorable circumstances. 1. Introduction Deformation measurement in many fields require high accuracy. For instance, in mining and structural engineering to predict failures of structures, in bio-mechanics to analyze stability of movement of differ- ent body parts, in electronics engineering to measure expansion of inte- grated circuits due to change in temperature, in mechanical engineering to analyze land and air vehicles’ stability under high speed pressure and heavy load, in medical images to measure movement and changes in dif- ferent body parts, and in seismology to understand movement between tectonic plates [1–3]. Deformation measurement involves measuring displacements and strains of each point on the surface of specimen which is under the process of deformation. These measurements subsequently help to re- constructs displacement and strain fields. This study focuses on two di- mensional deformation measurements. There are different ways to measure the deformation. Traditionally, strain gauges are used by connecting them on the surface of specimen at multiple points to measure their displacements. However due to physi- cal nature of the devices, deformation measurement process cannot be completed remotely using strain gauges without being in contact with the surface of specimen. In addition, the physical size of strain gauges limits its measurement to limited number of points on the surface of specimen. Therefore, deformation measurements needs to be interpo- lated at missing points on the surface of specimen. Lastly, another major limitation of the strain gauges is measuring displacement of point in a single direction only. E-mail address: ghulam.hassan@uwa.edu.au Alternate to physical deformation measurement process, a remote method is to use images. This process requires two images: one be- fore and one after deformation called as reference and deformed images respectively. There are many techniques to measure deformation with high accuracy using images. Most of these techniques are based on the principle of correlation of segments of images and are generally catego- rized as variants of Digital Image Correlation (DIC) [4–17]. The process of DIC involves selecting a segment of the reference im- age such that the point of interest is at the center of the selected seg- ment of the image. This segment of the image is known as subset of an image. In DIC, the correlation between subset of the reference image and the subset of the deformed image is maximized by estimating the displacements of the points of reference subset in the deformed image and their gradients. This requires DIC to optimize six variables which include displacements in two directions (horizontal and vertical) and their four gradients. The maximized correlation provides the estimated new displaced position of the point of interest in the deformed image. Image correlation process is repeated for all points of the reference im- age to obtain their new positions in the deformed image. This helps to reconstruct complete displacement fields in both horizontal and vertical directions. The strain fields can be obtained by using the estimated gra- dients of the displacements optimized during subset correlation process. Another way of obtaining strain fields is by calculating the gradients of the obtained displacement fields numerically. The measured displace- ments accuracy by DIC is reported to achieve one-hundredth of a pixel under suitable conditions [18]. The minimum expected accuracy for de- https://doi.org/10.1016/j.optlaseng.2018.12.003 Received 13 August 2018; Received in revised form 5 November 2018; Accepted 4 December 2018 0143-8166/© 2018 Elsevier Ltd. All rights reserved.