Content area
Full Text
1. Introduction
Health monitoring of concrete is one of the important tasks for the structural engineers. The durability of concrete structures depends on the quality of the concrete. Cracks in concrete will affect the durability of concrete structures. Cracks in concrete may be due to excessive stress, cyclic loading, over usage, improper construction material and poor quality of construction. The cracks can initiate the corrosion in the structures, decrease the strength and stiffness. Visual inspection of cracks by human is most frequently used method to detect the defects in concrete. It is cost effective. The manual inspections fully relay on the expert’s experience and it may fail in the quantitative analysis. Government and manufacturers have invested a huge amount of money to find the defective areas in concrete structures to ensure safety. The early detection and the continuous monitoring of concrete cracks will ensure the safety of concrete structures and human beings. Early detection will help to make proper plan to rehabilitate the infrastructures.
The computer vision-based methods are very useful for many areas including security surveillance, medical imaging, attention modeling, facial recognition, structural health monitoring (SHM) (Panetta et al., 2016; Andrushia and Thangarajan, 2019; Zhang et al., 2017; Qi et al., 2015). The detection of cracks in concrete is one of the emerging areas of computer vision. The destructive testing and non-destructive testing are the two different methods to find the cracks in civil concrete elements. Stress test, hardness test, crash test are some of the methods to perform destructive testing. These testing procedures are performed by visual examination and it is time consuming, cost effective. Laser testing, infrared and thermal testing, ultrasonic testing (Mohan and Poobal, 2018) are the different methods of non-destructive testing. The automatic detection of cracks can also be carried out by these methods. But the image based automatic crack detection methods are fast. It is one of the substitutes for older human intervention methods. Even though the image-based methods are fast, it has some limitations. The shape of the crack, illumination conditions, noises, shading effects have impact on the image-based crack detection methods. In the recent past, many techniques are developed to detect the cracks in concrete. They are used to locate the cracks automatically in...