Seite - 781 - in Book of Full Papers - Symposium Hydro Engineering

3.3. MONITORING VISUAL CHANGES The detection of visual changes in images is a well-researched topic in the field of remote sensing [12]. Consequently, a wide range of approaches were developed for automatic unsupervised change detection analysis in the past years. However, it turns out that these algorithms are well suited not only for large-scale satellite images but also for high-resolution ortho images of concrete defects. One approach is to compare the pixel intensities of two images by e.g. differencing, rationing or by image regression [12]. It is simple, easy to implement but is sensitive to the acquisition conditions. Therefore, it is essential to pre- process the data, i.e. to match the images in terms of radiometry and spatial coverage. We adopted the analysis of [12, 13] and focus on gray value differences in the ortho images. We validated the algorithm by placing shaving foam near a concrete crack to simulate an increase of efflorescence. The approach proved suitable to detect the changes and even to compute the area of change. Nevertheless, changing lighting conditions pose a problem to the algorithm, requiring more investigations for establishing a fully automatic change detection analysis. However, it assists the dam engineer and provides quantitates to assess the dam safety objectively. REFERENCES [1] ICOLD. Dam Surveillance Guide. Technical Report 158, International Commission on Large Dams. 109p, 2013. [2] HENRIQUES M. J., ROQUE D. Unmanned aerial vehicles (UAV) as a support to visual inspections of concrete dams. Dam World. Portugal, Lisbon, April 21–24, 2015, 12p, 2015. [3] BOAVIDA J., OLIVEIRA A., BERBERAN A. Dam monitoring using combined terrestrial imaging systems, 13th FIG Symposium on Deformation Measurement and Analysis, LNEC, Lisboa, 2008.+ [4] CAMP G., CARREAUD P., LANÇON H. Large Structures: Which Solutions for Health Monitoring? International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. XL-5(W2), pp. 137–141, 2013. [5] BERBERAN A., PORTELA E.A., BOAVIDA J. Assisted visual inspection of dams as a tool for structural safety control. A case study. Hydro 2006 – maximizing the benefits of hydropower, 8p, 2006. [6] Leica MS60/TS60, User Manual, Version 1.0. Leica Geosystems, Heerbrugg, Switzerland. 90p, 2016. [7] EHRHART M., LIENHART W. Image-based dynamic deformation monitoring of civil engineering structures from long ranges. Image Processing: Machine Vision Applications VIII. SPIE 9405. 14p, 2015. 781