Real-Time Concrete Surface Crack Detection Using Computer Vision Model—YOLO_v8

For the assessment of the safety and remaining life of any structure, accurate health monitoring is very important. The process of structural health monitoring can be assisted using the new advancements in machine learning. The paper proposes the use of the latest version 8 of the YOLO (You Only Look Once) algorithm which is a fast image classification and detection algorithm developed by Ultralytics famous for high speed and accuracy and can be used for live object detection which makes it a good fit for Surface Crack detection. The YOLO model will be used to propose the important critical areas in a structure based on the number of cracks for further assessment. YOLO_v8 is the latest instalment in the YOLO family of algorithms. The proposed method can be used with any device capable of recording video and sending a live feed to the computer system like a UAV or a mobile phone making it accessible and a very functional method.