| Popis: |
In a pipeline monitoring system, the engines of machines passing by can lead to vibration excitation for the optical fiber. The signal generated by the vehicle driving is similar to the excavation disturbance signal. Therefore, it may generate incorrect classification for one dimensional signal recognition by the pre-warning system. An image - based signal processing method was built to solve these problems. Firstly, completing differential operation for all detection points in the whole detection range, and dividing the differential data basing on the threshold value. Then extracting the event areas based on the method of the image introduced in this paper. Finally, utilizing the Hough transform to extract the existing line segments, and analyzing the slopes to determine the type of events. This method can effectively eliminate noise interference and reduce misinformation and omission alarm at the same time. This research will ultimately provide the basis for judging the output of alarm events in distributed optical fiber vibration sensing system. |
