Published Paper Details:

It uses the Unity Neural Encoding Tool (U-Net) and LeNet-5 to present a novel approach to crack-level classification using Artificial Intelligence techniques. For safety reasons, the primary goal of this study is to locate cracks in a variety of structures, including walls, bridges, and infrastructure. A substantial dataset of 10,000 images was used for this project, of which 8,000 were designated for training and 2,000 for testing. With the help of state- of-the-art deep learning and image processing techniques, the model can classify six distinct types of cracks. They are Normal crack, Deep crack, Gap crack, Riss crack, Wall care crack, and No crack. Using cutting-edge architectures makes it easier to improve upon current practices. In addition to improving safety and maintenance procedures in the civil engineering and construction industries, this research advances the field of quality monitoring by providing a trustworthy method for locating and categorizing cracks.