Application of Random Forest in Signal Recognition of Distributed Optical Fiber Mining

• 1、Automation School,Beijing University of Posts and Telecommunications,Beijing,100876
• 2、Beijing Municipal Institute of Labour Protection,Beijing,100000

Abstract：Based on the fully distributed optical fiber detection signal, according to the characteristics of the random forest algorithm in machine learning generally have a good effect on the classification problem, this paper proposes a random forest algorithm based on distributed optical fiber signal recognition method to mining, mining range accuracy signal recognition and accurate recognition can be improved. The specific method is: the collected signal mining and normal signal preprocessing, noise removal by filtering method, the signal in the sliding window frame, sub frame signal for feature extraction, including time domain and frequency domain characteristics of mining, according to the signal and the normal signal characteristic value, constructs a classifier using random forest algorithm the classification and recognition of the signal. The test signal is used to test the recognition method. The experimental results show that the proposed method has a good recognition effect.

Keywords： distributed optical fiber mining signal recognition random forest smoothing filter feature extraction