Survey on Crack Detection on Crack Types Using Machine Learning Techniques

Abstract

In real time scenario, cracks are very common in building, bridge, road, pavement, railway track, automobile, tunnel and aircraft. The presence of crack diminishes the value of the civil infrastructure and hence it is necessary to estimate the severity of crack. Crack detection and classification techniques with quantitative analysis play a major role in finding the severity of crack. The various quantitative metrics are length, width and area. Due to the rapid development in technology, number of images acquired for analysis is growing enormously. Therefore, automatic crack detection and classification techniques for civil infrastructure are essential. This paper focuses on three objectives: (i) Analysis of various crack detection and classification techniques based on crack types (ii) Implementation of Otsu’s based thresholding method for crack detection (iii) Design of proposed system.