PIG Tracking Utilizing Fiber Optic Distributed Vibration Sensor and YOLO

This paper proposed a strategy of tracking pipe PIG utilizing Fiber Optic Distributed Vibration Sensor (FODVS) and YOLO object detection algorithm. Phase sensitive Optical Time Domain Reflectometer (phi-OTDR) is employed as the distributed sensor to collect the real field vibration signal generated as a result of the collision between leather bowl and weld joint. The space-time graphs of phi-OTDR are prepared to construct data sets to train a typical YOLOv3 net model. The trained model is proved to be able to accurately capture the invert-V signature in the space-time graph thus revealing the real time position of a PIG. This paper provides a concept-proof preliminary demonstration on the promising combination of FODVS and object detection scheme for pursuing better performance of events recogntion.

Automated summary

This paper proposed a strategy of tracking pipe PIG utilizing Fiber Optic Distributed Vibration Sensor (FODVS) and YOLO object detection algorithm. The trained model is proved to be able to accurately capture the invert-V signature in the space-time graph thus revealing the real time position of a PIG. This paper provides a concept-proof preliminary demonstration on the promising combination of FODVS and object detection scheme for pursuing better performance of events recognition.