Optical Image Analysis Based Concrete Damage Detection

摘要

The deteriorating condition of concrete infrastructure and the prohibitive costs required for upgrading them require the development of innovative and effective decision support tools. Such tools will enable predictions about the condition and future performance of concrete structures and the allocation of limited funds for optimized maintenance that yields improved reliability and minimum life cycle costs. Furthermore, improved control on the variability between predicted and actual repair quantities would tend to reduce the long-term unit price of repair work, as contractors adapt to less instances of work shortfalls. There is a growing need for precise and reliable methods that use non-destructive testing (NDT) techniques to assess the deterioration in concrete structures and the extent of associated damage. Current use and development of monitoring methods and imaging techniques employing NDT is limited due to the unavailability of enough experimental data and field applications. This research aims to enhance image processing methods for application to imagery of concrete blocks that were exposed outdoors to the elements, and concrete slabs that were kept in the laboratory in controlled environments, in order to extract quantitative deterioration information from the concrete images obtained using various NDT imaging methods. Evaluations are done on the efficiency of the various types of NDT imaging methods of colour and greyscale imagery, in detecting different types of concrete deterioration and correlation between damage quantities obtained from processing of the concrete imagery and test measurements recorded for the laboratory specimens is established.