AI 기반 시멘트 복합체의 균열 검출 및 이미지 프로세싱 최신 기술 동향

Cementitious composites are key materials that determine the durability and safety of structures, and the occurrence and propagation of cracks are major causes of their performance degradation. With recent advances in high-resolution imaging technologies and artificial intelligence (AI), new approaches have emerged for quantitatively evaluating and predicting the crack behavior of cementitious composites. This paper comprehensively reviews the latest trends in image processing and AI-based analytical techniques applied to crack detection and assessment in cementitious composites. Specifically, it systematically analyzes the evolution of methods from traditional image analysis such as binarization, edge detection, and histogram analysis to deep learning-based approaches including convolutional neural networks (CNN), vision transformers (ViT), and generative AI-based restoration techniques. Furthermore, it discusses the potential integration of various sensing technologies such as optical and electron microscopy for microstructural observation, drone/UAV imaging, and infrared or thermal imaging. Finally, this review highlights key challenges related to dataset construction, training pipelines, performance metrics, and field implementation for real-world structural monitoring. The study aims to provide a technological foundation for intelligent maintenance and digital twin realization by proposing research directions for AI-based image interpretation in the evaluation of cracks in cementitious composites.