Review of Crack Detection System for Industrial Pipe

Industrial pipes play a critical role in various sectors, and their structural integrity is essential for safe and efficient operation. Over time, cracks may develop due to stress, corrosion, fatigue, or environmental factors, potentially leading to failures if not detected early. This paper presents a Python-based real-time crack detection system for industrial pipes, utilizing advanced image processing and machine learning techniques. The system supports both static image analysis and live webcam feed inspection, enabling flexible and continuous monitoring. Key functionalities include (1) crack detection using Canny edge detection and adaptive thresholding, (2) support for TensorFlow-based machine learning models for complex pattern recognition, (3) adjustable detection parameters for sensitivity tuning, and (4) statistical analysis of crack dimensions such as area. In addition to its detection capabilities, the system offers advanced visualization features, including split-screen comparisons, crack overlays, and severity-based highlighting. Built using Python with libraries such as OpenCV, TensorFlow, and CustomTkinter, the framework provides a modern, user-friendly interface and is scalable for integration into industrial monitoring setups. By enabling real-time condition assessment and facilitating early maintenance actions, the system significantly enhances the safety, reliability, and longevity of industrial piping infrastructure.