Modelling and Analysis of Advanced Dynamic Learning Techniques and Anisotropic Diffusion Filtering for Automated Crack Detection in Civil Structures

This study introduces a cutting-edge automated system that leverages dynamic learning methods for detecting cracks in civil structures. By combining pioneering preprocessing techniques with leading-edge machine learning algorithms, the proposed system aims to surpass the constraints of conventional manual inspection approaches. The methodology incorporates Anisotropic Diffusion Non-Local Mean Filtering (ADNLMF) for noise reduction and YOLO v3 for precise object detection. The system's effectiveness is substantiated using an extensive dataset, displaying substantial enhancements in both accuracy and efficiency.