PipesDefectDetection: A novel deep learning framework for real-time gas pipeline safety monitoring and defect recognition
Eginov A.A. Eginova S.A.
September 2025Elsevier B.V.
SoftwareX
2025#31
PipesDefectDetection is an open-source software framework for real-time visual detection of gas pipeline defects using deep learning. Built upon the YOLOv11 architecture and deployed via a lightweight Streamlit interface, the system enables engineers to identify surface anomalies – such as corrosion, dents, and deformation – directly from inspection images. The model is trained on field-collected datasets and supports continuous improvement via Roboflow-based annotation pipelines. With real-time inference, exportable reports, and platform independence, the software facilitates condition-based maintenance and risk-informed decision-making. This paper describes the architecture, functionalities, impact, and validation of the platform in operational environments.
Computer vision , Corrosion analysis , Deep learning , Defect detection , Pipeline integrity , Real-time monitoring
Text of the article Перейти на текст статьи
Technical department at “Tolagai-2050” LLP, Almaty, Kazakhstan
University of California, Berkeley, United States
Technical department at “Tolagai-2050” LLP
University of California
10 лет помогаем публиковать статьи Международный издатель
Книга Публикация научной статьи Волощук 2026 Book Publication of a scientific article 2026