Detection of Rail Defects with Deep Learning Controlled Autonomous UAV

With the widespread use of high-speed trains in recent years, rail transport has become a more comfortable and safe means of transportation. The safety and maintenance of railways are critical to safe travel. Conventional railway inspection systems are carried out by measuring train as well as manual control along the way. Human-based inspection systems are slow and measuring train-based inspection systems are expensive and occupy the line being inspected. In this study, a method is proposed for the control of the rail track with an autonomous unmanned aerial vehicle (UAV). The proposed method uses the deep Hough transform method for autonomously moving on the rail. Unlike normal image processing-based techniques, this method does not need any preprocessing and parameter adjustment. After removing the rails from the obtained rail images, the rail defects are detected by semantic segmentation. The developed method was compared with those in the literature, and it was seen that better results were obtained.