Detection of Surface Defects Inside Concrete Pipelines Using Trained Model on JetRacer Kit

This paper is the development of a crack pipeline defect detection software from the inside for real-time damage assessment using trained model. Collected data on the number of pixels found in the frame that signals damage using a ready-made camera from the JetRacer Kit. The analysis and display of the results of the detected cracks are performed using visualization by a rectangle around damaged concrete. The possibilities of evaluating detected damage based on data on pixels filling a crack in the pipeline are described. The quality of detection of cracks in pipes with artificial and natural lighting is checked.