Automated Crack Detection in Pipelines Using Advanced Image Processing

The automated detection of pipeline cracks is vital for maintaining the safety and operational lifespan of this crucial infrastructure. Early crack detection prevents costly repairs and minimizes the risk of catastrophic failures and environmental damage. Existing methods primarily rely on manual inspections and traditional image processing techniques, which often lead to inaccuracies and inefficiencies. These methods suffer from high false-negative rates and poor adaptability to different environmental conditions. This study proposes a deep learning framework utilizing YOLOv10 to automate the detection of pipeline cracks with enhanced accuracy. Studies using deep learning report detection accuracies between 80% and 92%, although performance varies across different studies. The range reflects the influence of factors such as dataset and model specifics.