Deep Learning-based Severity Classification of Concrete Cracks using YOLOv8 for Structural Health Analysis
Year: 2025
Type: article
Abstract: Crack detection and severity classification are essential tasks in structural health monitoring, especially for critical civil infrastructure such as roads and bridges. Traditional methods rely heavily on manual inspection, which is time-consuming, costly, and prone to human error. This study introduces a computer vision approach using the YOLOv8n model to automatically classify concrete surface conditions into six categories, ranging from "No Crack" to "Very Large Crack." This approach addresses the critical gap between binary detection and actionable severity interpretation in civil engineering inspections. After augmenting and preprocessing the dataset, the model was trained over 10 epochs and achieved high classification accuracy ranging from Top-1: 97.1% to Top-5: 99.9% on a multi-class dataset of over 11,501 annotated images.
Authors: Carson Bowling, Luke Pierini, Wisam Bukaita. Ph.D
Institution: Lawrence Technological University
Language: English
Subfield: Civil and Structural Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4414026994