FPGA Design And Implementation Of Seismic Data Compression Based On Compressed Sensing

With the continuous development of geophysical exploration technology,seismic exploration instruments are advancing in the direction of multi-parameter,multi-dimensional and multi-component.A large number of seismographs deployed at the seismic exploration site will generate massive seismic exploration data.In the face of massive real-time recovery of seismic exploration data,there will be network congestion or even network disconnection in the transmission network,so a more high-speed and reliable transmission network is needed to ensure the real-time and reliability of data transmission.How to solve the network congestion caused by the large amount of data,and at the same time ensure the real-time performance of data transmission is the core problem to be solved.Data compression method can reduce the amount of data from the source of data transmission,improve transmission efficiency and ensure real-time data transmission without upgrading the transmission network.The traditional data compression method is to compress the seismic data stored after the acquisition.The compression process is generally transform,quantization and encode,and the decompression process is the inverse of the compression process.This compression method needs to compress the complete seismic data,which takes a long time and consumes a lot of resources,and the real-time performance of data compression and transmission is poor.Aiming at the problems faced by traditional data compression methods,this paper proposes a compression data-based seismic data compression method based on Compressed Sensing(CS)theory.The method can directly compress the collected seismic data,and does not need to operate the stored complete seismic data,and the data compression has better real-time performance.Aiming at the seismic data compression and reconstruction method under the theory of compressed sensing,this article has carried out the following three aspects of research: the selection of sparse representation method,the construction of measurement matrix and the design of reconstruction algorithm.Firstly,the wavelet transform is used to sparsely represent the seismic data to obtain sparse representation coefficients.Then,a chaotic Bernoulli measurement matrix is constructed by using a chaotic sequence,and the seismic data sparse representation coefficient is compressed and measured by the measurement matrix to obtain a measured value.Finally,the measured value,that is,the compressed data,is transmitted to the data center for storage in real time.When the original seismic data is needed,the measured value is reconstructed and restored by the orthogonal matching pursuit(OMP)algorithm.In this paper,XILINX's XC6SLX45 FPGA chip is used to improve the original seismograph and build peripheral circuits.The overall structure of seismic data compression based on FPGA is proposed.In addition,the sub-modules such as seismic data loading,data calculation and sparse representation are further designed,and the effectiveness of each module is verified by simulation experiments,so as to ensure the sparse representation and compressed measurement process of seismic data on the lower computer(seismograph).The compression and reconstruction experiments were carried out on 10 sets of different seismic data with a length of 1000 sampling points,and the experimental results were analyzed.The results show that the seismic data compression method proposed in this paper,under different compression ratios,the PSNR of the reconstructed data is above 60 dB,the RMSE is below 0.12,and the hardware compression time is about 1.2ms.The seismic data compression algorithm is implemented on the FPGA hardware platform,and the real-time performance of data compression is guaranteed.In the face of massive seismic exploration data,the data compression method of this paper is used to compress the collected seismic data in the seismograph,reduce the data volume from the source of the data transmission,and improve the transmission efficiency without optimizing the transmission network.Therefore,the data loss problem caused by network congestion or even network disconnection in the seismic exploration data transmission network is solved.At the same time,the original seismic data is reconstructed by using the orthogonal matching pursuit algorithm to ensure the integrity of the seismic data,which is important and practical.significance.