Detection of Crack on Concrete Surface Using Semi-Automated Digital Image Processing Technique
Cracks detection is an important responsibility for monitoring and maintaining architectural safety. The manual cracks detection method takes a long time and suffers from the personal decision of the controller. In this research work we develop intelligent models according to image processing techniques for semi-automated cracks identification and analysis. The new model takes advantage of integration between different image processing approaches to improve the crack detection process .It is able to identify crack objects and analyze their properties including orientation, length, and weight. The main goal of this developed model is to improve the precision of crack detection results. Moreover, this paper employs and develops image processing model for crack detection to study the impact of highest total aggregate size, Fiber Reinforced Polymer (FRP) sheet bond length and FRB sheet bond width on the percentage of crack weight between concrete and FRP sheet. Experimental results show that the newly-built image processing model can detect crack defects in digital images completely. Accuracy indicators achieved have reached about 91% completeness, 97% correctness and 85% quality for the selected data. Moreover, the results pointed out that the percentage of crack weight between FRP sheet and concrete increases as the highest total aggregate size used decreases up to the best highest total aggregate size of 12.5mm, in addition the percentage of crack weight increases with FRP bond length and width decreased. Hence, the designed model can be a valuable tool for construction engineers in the maintenance and rehabilitation of concrete structures.
Keywords: Cracks, Image processing, Fiber Reinforced Polymer (FRP), Concrete.