| In recent years,distributed optical fiber sensing technology has been widely applied in the field of public buildings,bridges,tunnels and other infr-astructure safety monitoring.It plays an important role because of its passive,anti-electromagnetic interference,and long-distance large-scale monitoring characteristics.Distributed optical fiber sensing technology has attracted the extensive attention and research of scholars at home and abroad.However,the distributed optical fiber sensor system will get a huge amount of data in the process of large-scale continuous monitoring of infrastructures.And with the improvement of the performance of the distributed optical fiber sensor system,the amount of data collected per unit time will increase by tens or even hundreds of times,which leads to the problem of data expansion.Data expansion will influence data storage and data process,which makes the distributed optical fiber sensor system unable to give full play to its advantages and role in practical application.Aiming at the above problems,the following studies have been conducted around the echelon reduction and coding storage of Phase-OTDR distributed optical fiber vibration data:(1)Signal-to-noise separation of Phase-OTDR distributed optical fiber vibration data.According to the amplitude difference of signal and noise of Phase-OTDR distributed optical fiber vibration data,an Ostu signal-to-noise separation method based on simulated annealing optimization is proposed.Based on the difference between the amplitude of vibration signal and noise signal,the Phase-OTDR distributed optical fiber vibration data is denoised by using Ostu threshold algorithm optimized by simulated annealing.The denoising effect is compared with the existing empirical mode decomposition method,and the research on signal-to-noise separation of Phase-OTDR distributed optical fiber vibration data is completed.(2)Spatiotemporal compression of Phase-OTDR distributed optical fiber vibration data.According to the two-dimensional structure of Phase-OTDR distributed optical fiber vibration data,a two-dimensional spatiotemporal data compression method based on improved shearlet sparsity representation is proposed.The sparse representation of distributed optical fiber vibration signal is presented by shearlet transform.The spatiotemporal compression is completed by selecting frequency band and coefficients.At the same time,the performance evaluation and effect comparison are conducted between the method of this paper and the compression methods based on two-dimensional wavelet transform by using SNR and compression ratio.The vibration signal is compressed by the method of this paper,the compression ratio can reach up to 71%,which verifies the feasibility of the data compression method,and the research on spatiotemporal compression of Phase-OTDR distributed optical fiber vibration data is completed.(3)Spectrum analysis of Phase-OTDR distributed optical fiber vibration data.In the process of signal-to-noise separation of distributed optical fiber vibration data,it is impossible to completely separate the vibration signal from the noise.At the same time,in the process of spatiotemporal,part of the signal characteristic information will be lost due to the lossy compression.In the evaluation of echelon reduction effect of distributed optical fiber vibration data,the compression ratio and signal-to-noise ratio can only show that the method in this paper has achieved good echelon reduction effect.It cannot verify the retention degree of the feature information of the reduced data.Therefore,the frequency spectrum of the vibration signal before and after compression and reconstruction is analyzed from the aspects of signal energy,signal amplitude,bandwidth and so on.Ten kinds of time-frequency features of the signal before and after compression and reconstruction are extracted,and a random forest classification model is established to recognize and classify the signal.The recognition rate of the model is maintained at more than 95%.In addition,two typical classification models of SVM and KNN are established,and both of them maintain high classification accuracy for vibration signals,which further verifies the effectiveness of the spatiotemporal compression method in this paper.(4)Coding storage of Phase-OTDR distributed optical fiber vibration data.According to the double precision float storage format of compressed shearlet coefficients,an improved Huffman coding compression storage method is proposed.On the basis of echelon reduction,the Huffman coding method is studied.In order to improve the coding efficiency,the construction of Huffman tree in Huffman coding process is changed.After further coding,the storage space of the Phase-OTDR distributed optical fiber vibration data is reduced by nearly 30%and achieves the maximum compression.The research on coding storage of Phase-OTDR distributed optical fiber vibration data is completed.In this paper,the echelon reduction and coding storage of Phase-OTDR distributed optical fiber vibration data can compress the monitoring data fully,thus reduce the storage space of data and improve the efficiency of data transmission.It has great application value in the field of distributed optical fiber sensing. |
