Predicting Severity of Cracks in Concrete using Fuzzy Logic

Detection of cracks in concrete is highly recommended in early stages of construction to avoid potential hazards later. Need for automatic detection through image capturing and processing is rising. Till date only classifiers to detect cracks based on image processing have been designed. This paper presents an automatic system to detect severity of cracks. Every captured image of a crack is segmented through clustering. A fuzzy inference system is designed to output a risk prediction score based on the segments. This score can be helpful to decide any further actions to be taken once the crack is detected.