Automatic crack monitoring using photogrammetry and image processing J. Valença a,d,⇑ , D. Dias-da-Costa b,e , E. Júlio a,f , H. Araújo c , H. Costa a,d a ICIST, Av. Rovisto Pais, 1049-001 Lisboa, Portugal b INESC, Coimbra, Rua Antero de Quental 199, 3000-033 Coimbra, Portugal c ISR, Department of Electrical and Computer Engineering, University of Coimbra, Rua Luís Reis Santos, 3030-290 Coimbra, Portugal d Department of Civil Engineering, Institute Polytechnic of Coimbra, Rua Pedro Nunes – Quinta da Nora, 3030-199 Coimbra, Portugal e Department of Civil Engineering, University of Coimbra, Rua Luís Reis Santos, 3030-788 Coimbra, Portugal f Department of Civil Engineering, Instituto Superior Técnico, Technical University of Lisbon, Av. Rovisco Pais, 1049-001 Lisbon, Portugal article info Article history: Received 28 July 2011 Received in revised form 27 June 2012 Accepted 27 July 2012 Available online 7 August 2012 Keywords: Laboratorial tests Crack monitoring Crack characterisation Image processing Photogrammetry Monitoring abstract This manuscript presents an integrated approach for automatic crack monitoring combin- ing photogrammetry and image processing. In summary, the strain field obtained from photogrammetric data is used to map the cracked areas where image processing is applied. All processing is completely automatic since only a threshold value, related to the width of the crack, needs to be provided. Direct Shear Tests (DSTs) have been selected for calibration, validation and also as an experimental example. In conclusion, critical areas, the corresponding crack pattern and all related measures (e.g. crack width, length, area or path) could be provided for any stage of loading, until the complete failure of the specimens. Furthermore, all outputs require low computational cost, thus allowing monitoring vast campaigns of laboratorial tests. Ó 2012 Elsevier Ltd. All rights reserved. 1. Introduction The development of automatic procedures for monitor- ing the most relevant parameters during laboratorial tests is of fundamental importance. In the case of concrete members, the evolution of the crack pattern can assist a complete understanding of the overall structural behav- iour up to failure. Most existing approaches to obtain this pattern are still hand-sketched based and the crack open- ing is often evaluated using measuring magnifiers or crack width rulers. Different authors have recently applied image process- ing techniques aiming to characterise cracks [1–6]. These algorithms are based on a sudden variation of intensity of pixels to detect a crack. If other sources of discontinu- ities are present on the surface of the specimen (e.g. voids, stains, shadows), a sudden variation of intensity of pixels is also identified by the algorithm and the detection of a real crack becomes compromised. Therefore, all existing ap- proaches are still applied to small areas under very con- trolled conditions. A new approach has been proposed in [7] to automatically monitor the crack pattern for a com- plete surface. However, the selection of critical regions of analysis still depends on the user, which precludes an automatic procedure. In order to circumvent these drawbacks a new method is herein presented. It was specifically developed to auto- matically monitor crack evolution, on a given surface, by combining photogrammetry and image processing to min- imise user intervention. The strain field, directly computed from photogrammetric data, is used to map the cracked areas where image processing is applied. Furthermore, the spatial resolution is also computed thus allowing mea- suring the width, length or area of any selected crack. This procedure has been calibrated and validated using Direct Shear Tests (DSTs). One of these specimens is herein ana- lysed in detail. 0263-2241/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.measurement.2012.07.019 ⇑ Corresponding author at: Department of Civil Engineering, Institute Polytechnic of Coimbra, Rua Pedro Nunes – Quinta da Nora, 3030-199 Coimbra, Portugal. E-mail address: jonatas@isec.pt (J. Valença). Measurement 46 (2013) 433–441 Contents lists available at SciVerse ScienceDirect Measurement journal homepage: www.elsevier.com/locate/measurement