基于下变频和IQ解调的外差Φ-OTDR模式识别方法
Pattern Recognition Method for Heterodyne Φ-OTDR Based on Down Vonversion and IQ Demodulation
本文基于模拟下变频器、数字IQ解调和BP神经网络，采用FPGA+DSP数据采集和处理架构，提出了一种全嵌入式高信噪比，高分辨率和低成本的外差型Φ-OTDR模式识别方法。论文针对外差型Φ-OTDR，通过DSP、FPGA及其外围硬件电路替代原有的GZH级高速采集卡和信号发生器，减低了系统的体积和成本。在此基础上，本文设计了基于时空域二维图提取形态学特征的方法，并采用BP神经网络进行分类识别，相对于传统的针对一维信号进行模式识别的方法误报率更低，识别率更高。经过试验证明，本文设计的基于FPGA+DSP全嵌入式并行信号处理架构，满足了实时监测的要求，信噪比高达12.43db，事件识别准确率达到97.78％。
his paper proposed one high SNR, high recognition rate and low cost Heterodyne Φ-OTDR recognition system based on analog down conversion, digital IQ demodulation and bp-neturon network. In order to reduce cost and size, this system use FPGA and peripheral circuit to replace GHZ DAQ and signal generator. Compared to traditional modern recognition based on one dimension's features, this paper extracted features from two dimensions (time and space) based on morphological image processing. Finally, bp-neturon network was used to multi-class recognition. In order to achieve real-time monitoring requirements, this system proposed FPGA+DSP dual-core architecture of parallel signal processing. Experiments shows, this system's SNR was 12.43db and recognition rate was 97.78％.
曾周末、沈隆翔、沙洲、封皓
无线电设备、电信设备电子技术应用通信
相位敏感光时域反射仪，模式识别，外差探测，数字IQ解调，形态学特征提取，全嵌入式。
:phase-sensitive optical time-domain reflectometer(Φ-OTDR)pattern recognition heterodyne detection digital IQ demodulationand feature extraction in Morphology Embedded
曾周末,沈隆翔,沙洲,封皓.基于下变频和IQ解调的外差Φ-OTDR模式识别方法[EB/OL].(2017-04-14)[2025-08-11].http://www.paper.edu.cn/releasepaper/content/201704-149.点此复制
评论