Digital image correlation (DIC) methods enable accurate 2D and 3D measurements of changes in images.1 These methods use tracking and image registration techniques to make full-field non-contact measurements of displacements and strains in a wide variety of engineering applications. DIC techniques can also be used to measure changes in art objects. However, most of these applications require long-term monitoring.
We have designed a new method that automatically merges temporally distributed DIC data.2, 3 The software and hardware of the basic DIC method have been enhanced. In addition, the results are combined with thermal camera data, so that the new procedure is suited to power plant and chemical industry process control applications, e.g., boiler drum annealing supervision, district heating pipeline diagnostic tests, and pipeline displacement measurements.4–6 The technique can also be used to monitor the health of art exhibits.
In the DIC method we have designed, we use a flat calibration artifact (CA) as a reference for merging data acquired by two cameras placed at different positions. The CA remains permanently at the measurement site and is kept at a fixed position during consecutive measurements. For every data set, we transform the images obtained from the pair of cameras to the coordinate system of the reference CA. The transformation matrix is estimated by analyzing the 3D positions of markers, which is achieved through triangulation of the corresponding 2D positions in the stereo camera images. Standard 3D DIC analysis is subsequently performed on the images we obtain through our automatic data merging procedure (see Figure 1).
Figure 1. Flow chart illustrating the automatic merging of temporally distributed data in the 3D digital image correlation (DIC) method. CAref3D: Calibration artifact in reference data set. CAimg3D: Calibration artifact in data set to be merged. T3D: 3D transformation matrix.
Figure 2. (a) Schematic of a power plant pipeline system. (b) Photograph of the 3D DIC system set up to monitor the suspension at point A9. FOV: Field of view.
We used our DIC method to monitor the strength of a pipeline structure in a power plant (see Figure 2). During cycles in the pipeline's movement (related to the temperature and pressure of steam), components of its suspensions and supports can become damaged, and the pipeline may move in non-permissible directions. This can, in turn, cause deflections and propagation of cracks due to internal stresses. We used the 3D DIC method to measure the displacement of a pipe cross and its suspension—see Figure 2(b)—during a night shutdown and morning start-up of a power plant's steam generators. We used these measurements to assess the technical condition of the selected components. Example displacement maps are shown in Figure 3, and graphs of the 3D displacement of a selected point are shown in Figure 4. Our 3D DIC monitoring system provides a history of the pipeline use. The system can be used to assess the technical condition of the pipeline and to allow better planning for system overhauls.
Maps of displacement (U, V, W), in three dimensions, of a pipe cross and suspension at point A9 of the pipeline system shown in Figure 2
. These measurements were taken when the steam pressure was at its minimum. Displacement measurements were made specifically at points Q1, Q2, and Q3 (see Figure 4
Displacements at point Q1 in the U, V, and W directions (see Figure 3
) as a function of steam (a) pressure (P) and (b) temperature (T) in the left (L) and right (R) threads of the power station pipeline.
We also applied our 3D DIC method to monitor localized deformation of a canvas painting subjected to environmental stresses caused by rapid changes in relative humidity. During our experiment, DIC data was collected every 20s together with relative humidity and temperature values. We correlated the processed displacement maps with the humidity data, and combined the maps to make video animations for subsequent inspection. An example map, from the 55th minute of the experiment, when the relative humidity reached its maximum value (73%), is shown in Figure 5(a). The local defects caused by material discontinuities are clearly visible in the displacement maps. A separate data presentation method, however, is required to make quantitative analyses of the displacements.
Figure 5. Results of monitoring a repaired canvas painting that was subjected to rapid humidity changes. (a) Relative displacement maps in the U, V, and W directions at 73% relative humidity. (b) Displacement changes in a cross section through the repaired canvas—indicated by the white lines in (a)—as a function of time. The color scale for each map is different.
We monitored a repaired canvas painting by measuring the displacements at specific points along a line at the center of a vertical cut that is butt-joined to synthetic adhesive. The results from these measurements are plotted as a function of time in Figure 5(b). The images show the development of a characteristic deformation pattern in the region of the repair. Our method is capable of detecting the location, range, shape, and temporal development of such deformations. The deformation is automatically registered, and the measurements can be made with high frequency over long periods. We can quantitatively record deformations over a whole canvas, as well as more small-scale strains that are caused by repairs.
In summary, we have designed a new 3D DIC technique that automatically merges temporally distributed data. So far, our work has focused on its application for the structural conservation of paintings, but it can also be used to make quick assessments of other artifacts such as tapestries or wooden sculptures.7 Our future development of the 3D DIC data merging procedure will focus on the implementation of improved interpolation algorithms. We also plan to offer the technology as a service8 to the power, building, and chemical industries, as well as in the form of compact digital cameras with DIC functionality for museums.
Financial support for this work came from Poland's National Centre for Research and Development within the Opt4Blach project, and the European Union Structural Funds project MONIT: Health Monitoring Lifetime Assessment of Structures.
Małgorzata Kujawińska, Marcin Malesa, Krzysztof Malowany
Warsaw University of Technology
Małgorzata Kujawińska is a professor of applied optics and the head of the Photonics Engineering Division in the Faculty of Mechatronics. She is an international expert in coherent and non-coherent methods of full-field optical metrology and the recipient of the 2013 SPIE Chandra S. Vikram Award in optical metrology.
Marcin Malesa is a PhD candidate and a co-founder of KSM Vision Ltd.
Krzysztof Malowany is a PhD candidate and a co-founder of KSM Vision Ltd.
1. M. Sutton, J. J. Orteu, H. Schreier, Image Correlation for Shape, Motion, and Deformation Measurements , Springer, New York, 2009.
2. M. Malesa, M. Kujawińska, Modified two-dimensional digital image correlation method with capability of merging of data distributed in time, Appl. Opt. 51, p. 8641-8655, 2012.
3. M. Malesa, M. Kujawińska, Deformation measurements by digital image correlation with automatic merging of data distributed in time, Appl. Opt. 52, p. 4681-4692, 2013.
4. M. Kujawińska, M. Malesa, K. Malowany, P. M. Błaszczyk, Application of image based methods for monitoring and measurements of structures in power stations, Key Eng. Mater. 518, p. 24-36, 2012.
5. M. Malesa, M. Kujawińska, K. Malowany, B. Siwek, Diagnostic of structures in heat and power generating industries with utilization of 3D digital image correlation, Proc. SPIE 8788, p. 87881X, 2013. doi:10.1117/12.2021580
6. M. Malesa, K. Malowany, U. Tomczak, B. Siwek, M. Kujawińska, A. Siemińska-Lewandowska, Application of 3D Digital Image Correlation in maintenance and process control in industry, Comp. Indust.
, 2013. doi:0.1016/j.compind.2013.03.012
7. M. Malesa, K. Malowany, L. Tymińska-Widmer, E. A. Kwiatkowska, M. Kujawińska, B. J. Rouba, P. Targowski, Application of digital image correlation (DIC) for tracking deformations of paintings on canvas, Proc. SPIE 8084, p. 80840L, 2011. doi:10.1117/12.889452