Surface Crack

Surface crack detection and characterization are crucial for ensuring structural integrity and safety across various applications, from civil infrastructure to additive manufacturing. Current research heavily utilizes deep learning, employing convolutional neural networks (CNNs), transformer-based architectures, and graph neural networks to achieve accurate segmentation and classification of cracks from images and point clouds, often incorporating techniques like transfer learning and semi-supervised learning to improve efficiency and generalization. These advancements enable automated, efficient, and potentially more precise crack detection and analysis compared to traditional manual methods, leading to improved structural health monitoring and maintenance strategies.