ARCHIVES

The work describes a serpentine locomotion robot for the detection of cracks in gas pipelines. A camera has been added to the robot to simulate a camera that would be used for inspecting a pipeline. All tests were conducted in a controlled environment using a wired connection between the camera and the laptop. The tests did not take place in the actual pipeline. While testing, images of cracks were presented manually to the camera. The live feed from the camera was sent to the laptop for image processing. The YOLOv8 deep learning model was used to detect cracks in real time. When a crack was detected, a bounding box was drawn and an alert message was generated. The frames in which cracks were detected were saved automatically with a time stamp in a separate file. This project demonstrates that AI-assisted crack detection can be an effective tool for inspection. This robot is a low-cost, easy-to-use robotic crack-detection and will be enhanced in the future to perform autonomous pipeline inspections.