Monitoring Data Processing of Distributed Optical Fiber Temperature Sensor Based on Multi-wavelet

Abstract

The data acquired from distributed optical fiber temperature sensor systems inevitably had noise. In order to effectively detect and eliminate the noise, and to acquire more reliable data, an improved matrix pre-processing method and a new soft threshold method were proposed and applied to eliminate the noise of temperature data collected from a distributed optical fiber temperature sensor system (DTS), based on the principle and characteristics of multi-wavelet transforms and combined with the traditional matrix pre-processing method and line repetition pre-processing method. The results show that the improved matrix pre-processing method has much higher signal-to-noise ratio (SNR) and better de-noising effect than the traditional matrix pre-processing method and DB4 wavelet method. Therefore, the method has better de-noising application potential.