Detection and Length Estimation of Cracks in Enhanced Underwater Concrete Structure Images using Image Processing

Abstract

Underwater concrete structures crack detection and computing the length of the crack formed on those structures is arduous. The development of underwater concrete structures is increasing rapidly. A well-organized, systematic inspection of the cracks is required to ensure the safety of the people, strength and life of the concrete structures. Cracks occurring on concrete structures are a result of heavy load on the structures, low-quality materials used during construction, temperature changes, haze and underwater environment conditions. Inspecting deep inside the water is a challenging issue. Hence, digital cameras often equipped in submarines are opted to visualize, observe and capture images of the concrete structures on inspection. The images acquired from the underwater environment consist of noise, suspended particles, uneven, non-uniform, low contrast, and low-light conditions. This paper aims to enhance the low-light underwater images using dual illumination estimation, reduce the noise from the images, remove the suspended particles, detect the crack formed on the concrete structure and calculate the length of the crack. The enhanced underwater images of the cracks present on concrete structures produced by the proposed algorithm can be easily identified by the observer. The length of the crack determines a specific treatment method to treat the crack. The proposed model targets to detect the crack and compute the total length of the crack formed on the underwater concrete structures.