Scholarlink

<p>This study evaluated the efficacy of convolutional neural network-based algorithm, YOLOv8s in distress detection on concrete pavements and assessed its performance with low-volume datasets. The research scope encompassed field surveys, drone-based image collection, customization of YOLOv8s model, image annotations, model training and testing, and accuracy analysis. Seven survey sites were selected, from which a dataset of 881 raw images with annotations was compiled. Subsequently, this dataset was divided into three parts: 75% for training, 5% for validation and 20% for testing purposes. The YOLOv8 algorithm demonstrated a prediction capability with 88% precision and 7.7% recall value. Overall, this study highlights the potential of utilizing advanced algorithms in automated technology as an alternative to the conventional manual procedure for effective distress detection in concrete pavements.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Distress Identification in Concrete Pavement Using Convolution Neural Network: An Application of Low-Cost Condition Survey

  • 期刊论文

  • 发表日期：

    2026年4月22日

查看全文
  • Mrityunjay Midya,
  • Mahabir Panda,
  • Gourab Saha

摘要

This study evaluated the efficacy of convolutional neural network-based algorithm, YOLOv8s in distress detection on concrete pavements and assessed its performance with low-volume datasets. The research scope encompassed field surveys, drone-based image collection, customization of YOLOv8s model, image annotations, model training and testing, and accuracy analysis. Seven survey sites were selected, from which a dataset of 881 raw images with annotations was compiled. Subsequently, this dataset was divided into three parts: 75% for training, 5% for validation and 20% for testing purposes. The YOLOv8 algorithm demonstrated a prediction capability with 88% precision and 7.7% recall value. Overall, this study highlights the potential of utilizing advanced algorithms in automated technology as an alternative to the conventional manual procedure for effective distress detection in concrete pavements.