Rapid entropy-based detection and properties measurement of concrete spalling with machine vision for post-earthquake safety assessments Stephanie German a,⇑ , Ioannis Brilakis b , Reginald DesRoches a a School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, GA 30332, USA b Laing O’Rourke Centre for Construction Engineering and Technology, University of Cambridge, Cambridge CB2 1PZ, UK article info Article history: Received 4 November 2011 Received in revised form 15 June 2012 Accepted 18 June 2012 Available online 13 July 2012 Keywords: Spalling detection Property retrieval Post-earthquake reconnaissance Machine vision Image processing Reinforced concrete abstract The current procedures in post-earthquake safety and structural assessment are performed manually by a skilled triage team of structural engineers/certified inspectors. These procedures, and particularly the physical measurement of the damage properties, are time-consuming and qualitative in nature. This paper proposes a novel method that automatically detects spalled regions on the surface of reinforced concrete columns and measures their properties in image data. Spalling has been accepted as an impor- tant indicator of significant damage to structural elements during an earthquake. According to this method, the region of spalling is first isolated by way of a local entropy-based thresholding algorithm. Following this, the exposure of longitudinal reinforcement (depth of spalling into the column) and length of spalling along the column are measured using a novel global adaptive thresholding algorithm in con- junction with image processing methods in template matching and morphological operations. The method was tested on a database of damaged RC column images collected after the 2010 Haiti earth- quake, and comparison of the results with manual measurements indicate the validity of the method. Published by Elsevier Ltd. 1. Introduction Currently, the damage incurred due to earthquakes is evaluated manually by a team of certified inspectors and structural engi- neers; they follow guidelines provided by the Federal Emergency Management Agency (FEMA) and the Applied Technology Council (ATC) and assess the impact of visual damage (e.g. concrete spall- ing) on critical structural components to make sure that the dam- aged building remains stable and maintains an adequate level of structural integrity [1,2]. Following the January 12, 2010 earthquake in Haiti, over 100,000 houses were destroyed and nearly 190,000 damaged in the Port-au-Prince and surrounding areas [3]. Following the Febru- ary 22 earthquake in the Canterbury/Christchurch area of New Zealand in 2010–2011 (third in a series of four significant quakes in the area within a year), 10,000 houses were damaged to the point of demolition [4]; however, the most significant damage took place in the Central Business District (CBD), home to 4000 com- mercial buildings, 1000 of which would ultimately be demolished [5]. In each of these cases, the large inventory of damaged build- ings called for an equally large structural assessment task force, and in the aftermath of the earthquakes, several inefficiencies sur- faced in the current procedure for the post-earthquake building assessment. According to a US Senate Report, Haiti had made little progress in rebuilding in the 5 months since its earthquake due to an absence of leadership, disagreements among donors and gen- eral disorganization, all regarding the post-earthquake building assessments [6]. In the Canterbury region, it took 10 days to in- spect just 75% of those buildings in the CBD, and over 1300 com- mercial buildings were given red or yellow placards, restricting access to the entire CBD through September 2011, 7 months after the third earthquake [7]. Research has shown that the existing procedures for response and assessment is not only too time-consuming, but also highly biased, relying on the qualitative judgment of a structural engineer or inspector [8]. The information obtained during the structural evaluations is qualitative in nature, classifying the damage to a building into one of four categories (light, moderate, average, se- vere). The definition of each can vary significantly from evaluator to evaluator [9]. The enormous demand instituted in the event of large natural disasters, such as these earthquakes, is far beyond the current capacity of response teams. Thus, the societal and eco- nomic impact of the inefficiencies associated with assessment pro- cedures is only heightened by the fact that, until the evaluations are completed, those affected are homeless and/or jobless. These inefficiencies can be overcome if the current manual eval- uation practices are fully automated based on the visible structural elements and the visible damage which lays on them. In order to automate the process, both the load-bearing members in a struc- ture and the damage to the surface of the structural member need to be detected and assessed. Machine vision based methods have 1474-0346/$ - see front matter Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.aei.2012.06.005 ⇑ Corresponding author. Tel.: +1 8322640784. E-mail address: s.german@gatech.edu (S. German). Advanced Engineering Informatics 26 (2012) 846–858 Contents lists available at SciVerse ScienceDirect Advanced Engineering Informatics journal homepage: www.elsevier.com/locate/aei