Design Of Distributed Optical Fiber Sensor Signal Processing System

The measurement distance of high-performance distributed fiber optic sensors can reach several hundred kilometers,and their detected signals are relatively weak.Therefore,signal processing systems are required to have a high signal-to-noise ratio.The performance of sensors is usually improved through cumulative noise reduction.Distributed fiber optic sensors require accumulation times of thousands or even millions of times,making it difficult to achieve such a large amount of high-speed real-time data stored in memory or transmitted to the CPU for real-time processing.In this paper,based on theories such as distributed fiber optic sensing and FPGA,a signal processing system is designed and implemented from top to bottom.The paper proposes a scheme for real-time accumulation using DDR,focusing on the optimal strategy for achieving the accumulation processing function at maximum sampling rate under the influence of factors such as DDR read/write switching time and refresh time.The system can accumulate 2-channel 14bit sampling data with a sampling rate below 320MSPS,accumulated frame count below 65535,and sampling points below 200k and transmit the processing results to the CPU.The signal transmission bandwidth is approximately 1553MB/s.The paper places high-speed and real-time accumulation operations on the front-end,reducing the pressure of back-end transmission,storage,and processing,while minimizing the utilization of hardware resources such as BRAM,leaving sufficient space for the design and implementation of other functions of distributed fiber optic sensors.