An Automated Measuring System of Cracks in Concrete which Uses Image Processing Techniques and Artificial Intelligence Theory

The measurement of cracks in concrete surfaces is a procedure undertaken to diagnose deterioration and the prospective life of concrete structures. The authors have been studying a measuring system which uses computer image processing to reduce labour requirements and increase reliability. In this study crack data were first defined and then a data processing method which combines general-purpose image processing techniques with artificial intelligence theory was developed in order to extract crack patterns from concrete surface data and measure crack length, direction and width. It was found that crack measurements can be processed automatically or with the aid of simple manual operations.