Time domain matched filter, optical fiber sensing system and time domain matched filtering method

Abstract

The invention provides a time domain matched filter, an optical fiber sensing system and a time domain matched filtering method, and the time domain matched filter comprises an accumulation module and a multiplication and addition module connected with the accumulation module. The accumulation module and the multiplication and addition module are used for calculating time domain matching filtering results of input data and filtering parameters, the input data comprise real part data and imaginary part data, and the filtering parameters are quantized filtering parameters. According to the scheme, the accumulation module and the multiplication and addition module are arranged, layered accumulation processing is conducted on the data firstly, then multiplication and addition operation is conducted, the number of times of multiplication operation can be reduced, FPGA hardware resource consumption is reduced, and the resource utilization efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of filtering technology, and in particular to a time-domain matched filter, an optical fiber sensing system, and a time-domain matched filtering method. Background Art

[0002] Matched filtering is a filter generated by processing the swept signal (complex conjugation, where the real parts are equal and the imaginary parts are opposite). It can be used in applications such as signal demodulation and pulse compression. With technological advancements, the original signals used in frequency sweeps are becoming longer to obtain signal details in long-distance, high-bandwidth scenarios. As a result, matched filters are becoming larger as application scenarios expand.

[0003] Currently, signal filtering solutions can be divided into two categories, including frequency domain filtering solutions and time domain filtering solutions. Although the frequency domain solution consumes fewer computing resources, it consumes extremely high storage resources. Conventional FPGAs are difficult to meet the application scenarios that require receiving echo signals with too many single points. The time domain solution requires less cache and does not need to store a large number of intermediate results in the calculation process. Instead, the final result is calculated immediately based on the current signal value in real time and transmitted in a pipeline form without caching a large number of intermediate values, which is suitable for the FPGA computing architecture. However, the time domain matched filtering solution in the existing technology requires large computing resources, and the resource consumption of the multiplier and adder is difficult to meet the requirements. Summary of the Invention

[0004] The purpose of the embodiments of the present application is to provide a time-domain matched filter, an optical fiber sensing system, and a time-domain matched filtering method to solve the above-mentioned problems.

[0005] In a first aspect, an embodiment of the present application provides a time-domain matched filter, comprising: an accumulation module and a multiplication-addition module connected to the accumulation module; the accumulation module and the multiplication-addition module are used to calculate a time-domain matched filtering result of input data and filter parameters, wherein the input data includes real data and imaginary data, and the filter parameters are quantized filter parameters, wherein: The accumulation module includes a first accumulation unit and a second accumulation unit, wherein: The first accumulating unit is configured to respectively accumulate the values ​​in the plurality of target data sets and send the obtained plurality of first accumulated results to the second accumulating unit; wherein each of the target data sets includes real data and / or imaginary data to be multiplied by the same first filtering parameter value to obtain the real part result of the time-domain matched filtering result, and to be multiplied by the same second filtering parameter value to obtain the imaginary part result of the time-domain matched filtering result; the second accumulating unit being configured to accumulate the first accumulating results to be multiplied by the same first filtering parameter value to obtain the real part result among the plurality of first accumulating results to obtain a second accumulating result, and to accumulate the first accumulating results to be multiplied by the same second filtering parameter value to obtain the imaginary part result among the plurality of first accumulating results to obtain a third accumulating result, and to send the second accumulating result and the third accumulating result to the multiplication-addition module; The multiplication and addition module is used to multiply the second accumulated result by the first filtering parameter value and accumulate the multiplied results to obtain the real part result, and multiply the third accumulated result by the second filtering parameter value and accumulate the multiplied results to obtain the imaginary part result.

[0006] In the implementation process of the above solution, by setting up the accumulation module and the multiplication-addition module, the data is first layered accumulated and then the multiplication-addition operation is performed, which can reduce the number of multiplication operations, reduce the consumption of FPGA hardware resources, and improve resource utilization efficiency; on the other hand, by accumulating the input data to be multiplied with the same filter parameter value in advance, the originally repeated multiplication operations are merged, and then the accumulation results are multiplied, which effectively reduces the number of times the multiplier and adder are used, and reduces the consumption of hardware resources.

[0007] In an implementation of the first aspect, the time-domain matched filter further includes: An input data preprocessing module is connected to the accumulation module and is used to perform a decimation operation on the input data so that the data flow rate of the input data matches the processing rate of the time domain matched filter; and store the input data after the decimation operation into a register group.

[0008] In the implementation process of the above scheme, the input data preprocessing module performs extraction operations on the input data to match the data flow rate of the input data with the processing rate of the time-domain matched filter, so that the input data can be efficiently processed by the time-domain matched filter, avoiding data loss or processing delay caused by rate mismatch; on the other hand, extracting operations on large amounts of input data can remove redundant information in the input data, reduce the amount of input data while retaining the key features of the input data. The extraction operation can improve the data quality of the input data, thereby improving the matched filtering effect of the above-mentioned time-domain matched filter.

[0009] In an implementation of the first aspect, the input data preprocessing module, the accumulation module, and the multiplication-addition module calculate the time-domain matched filtering result of the input data and the filtering parameters in a pipeline mode.

[0010] In the implementation process of the above scheme, input data preprocessing, data grouping accumulation, and data multiplication and addition can be divided into four stages for parallel processing. Each stage is independent and closely connected. Data can be transmitted in sequence to form a pipeline structure. There is no need to wait for all data collection to be completed before processing, which is beneficial to improving the data processing efficiency of the above-mentioned time-domain matched filter; on the other hand, unused hardware resources can be allocated to different stages. Compared with the non-pipeline architecture, the above scheme can significantly improve resource utilization; on the other hand, the time-domain matched filter can receive and process input processing in real time and continuously, which is beneficial to improving the real-time and continuity of data processing of the time-domain matched filter and meeting the real-time processing requirements of the monitoring system.

[0011] In an implementation manner of the first aspect, the first accumulating unit and the second accumulating unit perform an accumulation operation by using a merge accumulation method.

[0012] During the implementation of the above scheme, the first accumulation unit and the second accumulation unit can perform accumulation operations in a merge-and-accumulate manner, which can reduce the number of operations and thus optimize the calculation process of the above-mentioned time-domain matched filter, which is beneficial to improving the calculation efficiency of the above-mentioned time-domain matched filter; on the other hand, the merge-and-accumulate can reduce the risk of error accumulation and is beneficial to improving the matched filtering effect of the above-mentioned time-domain matched filter; on the other hand, the merge-and-accumulate groups the data for processing, which is beneficial to improving the utilization rate of parallel computing resources and thus improving the utilization rate of computing resources of the above-mentioned time-domain matched filter.

[0013] In an implementation of the first aspect, the first accumulating unit is further configured to: The first accumulation results of the accumulation operations completed within the same cycle are combined and sent to the second accumulation unit; wherein the completion cycle of the accumulation operation is obtained based on the amount of data to be accumulated.

[0014] In the implementation process of the above scheme, by adopting the time-sharing alignment strategy, the transmission and processing of data can be reasonably arranged according to the actual accumulation time of each target data set, avoiding excessive allocation of registers for waiting for synchronization, thereby greatly reducing the register consumption of the matched filtering process, which is beneficial to improving the resource utilization of the above-mentioned time-domain matched filter; on the other hand, the time-sharing alignment strategy can enable the second accumulation unit to further process the calculation results of the different target data sets that have completed the accumulation calculation in the first accumulation unit in the same cycle, reducing the idle waiting time of the second accumulation unit, which is beneficial to improving the computational efficiency of the above-mentioned time-domain matched filter; on the other hand, because the time-sharing alignment strategy avoids the large amount of register consumption caused by direct beat synchronization, it reduces the timing problems caused by unreasonable register distribution or too long data transmission path, which is beneficial to improving the overall performance of the above-mentioned time-domain matched filter.

[0015] In an implementation manner of the first aspect, the second accumulating unit is further configured to: The second accumulation result and / or the third accumulation result that complete the accumulation operation in the same cycle are combined and sent to the multiplication and addition module; wherein the completion cycle of the accumulation operation is obtained based on the amount of data to be accumulated.

[0016] In the implementation process of the above scheme, the time-sharing alignment strategy is adopted, which can reasonably arrange the transmission and processing of data according to the actual accumulation time of each target data set, avoid excessive allocation of registers for waiting for synchronization, thereby greatly reducing the register consumption of the matched filtering process, which is beneficial to improving the resource utilization of the above-mentioned time-domain matched filter; on the other hand, the time-sharing alignment strategy can enable the multiplication and addition module to further process the calculation results of the different target data sets in the second accumulation unit that have completed the accumulation calculation in the same cycle, reducing the idle waiting time of the multiplication and addition module, which is beneficial to improving the computational efficiency of the above-mentioned time-domain matched filter; on the other hand, because the time-sharing alignment strategy avoids the large amount of register consumption caused by direct beat synchronization, it reduces the timing problems caused by unreasonable register distribution or too long data transmission path, which is beneficial to improving the overall performance of the above-mentioned time-domain matched filter.

[0017] In an implementation of the first aspect, the target data set is determined by a relationship graph; The relationship graph includes a first node and a second node, the first node is used to represent the first filtering parameter value, the second node is used to represent the second filtering parameter value, and the edge connecting the first node and the second node is used to represent real data and / or imaginary data to be multiplied with the same first filtering parameter value to obtain a real result in the time-domain matched filtering result, and to be multiplied with the same second filtering parameter value to obtain an imaginary result in the time-domain matched filtering result; The target data set includes a set of edges connecting the same first node and the same second node.

[0018] In the implementation process of the above scheme, the target data set is obtained based on the graph theory method, and the set to which it belongs can be directly determined by analyzing the edges between the nodes, which is conducive to improving the grouping efficiency; on the other hand, the number of sets that need to be multiplied is reduced, thereby reducing the demand for multiplier resources, which is conducive to improving the resource utilization of the above-mentioned time-domain matched filter; on the other hand, the set partitioning method based on graph theory can flexibly construct the corresponding graph structure and group it according to the coefficient distribution and quantization characteristics of different filters, which is conducive to improving the flexibility of the above-mentioned time-domain matched filter.

[0019] In an implementation of the first aspect, the relationship graph includes at least one of a relationship graph represented by an adjacency matrix, a relationship graph represented by an adjacency list, a relationship graph represented by a cross linked list, and a relationship graph represented by an adjacency multi-list.

[0020] In the implementation process of the above scheme, different forms of relationship diagrams can be selected to determine the target data set. Different relationship diagram representation methods provide a variety of data organization and processing methods, which makes it convenient to select appropriate methods according to actual needs and data characteristics, enhances the flexibility of data processing, and enables the above-mentioned time-domain matched filter to adapt to a variety of application scenarios.

[0021] In a second aspect, an embodiment of the present application provides a time-domain matched filtering method, the method comprising: obtaining input data and filtering parameters; wherein the input data comprises real data and imaginary data; the filtering parameters are quantized filtering parameters; respectively accumulating the values ​​in a plurality of target data sets to obtain a first accumulation result; wherein each of the target data sets comprises real data and / or imaginary data to be multiplied with the same first filtering parameter value to obtain the real result in the time-domain matched filtering result, and to be multiplied with the same second filtering parameter value to obtain the imaginary result in the time-domain matched filtering result. ; Accumulating the first accumulated results to be multiplied with the same first filtering parameter value to obtain the real part result among the multiple first accumulated results to obtain a second accumulated result; Accumulating the first accumulated results to be multiplied with the same second filtering parameter value to obtain the imaginary part result among the multiple first accumulated results to obtain a third accumulated result; Multiplying the second accumulated result with the first filtering parameter value and accumulating the multiplied results to obtain the real part result; Multiplying the third accumulated result with the second filtering parameter value and accumulating the multiplied results to obtain the imaginary part result.

[0022] In a third aspect, an embodiment of the present application provides a fiber optic sensing system, comprising: The time-domain matched filter provided by the first aspect or any possible implementation manner of the first aspect.

[0023] Other features and advantages of the present application will be described in the following description and, in part, will become apparent from the description or be understood by practicing the embodiments of the present application. The objectives and other advantages of the present application can be achieved and obtained through the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without creative work.

[0025] Figure 1 A schematic diagram of the structure of the time-domain matched filter provided in an embodiment of the present application; Figure 2 A schematic diagram of the distribution of pre-quantization filtering parameters provided in an embodiment of the present application; Figure 3 A schematic diagram of the distribution of post-quantization filtering parameters provided in an embodiment of the present application; Figure 4 A schematic diagram of the calculation process of the time-domain matched filter provided in an embodiment of the present application; Figure 5 A schematic diagram of the placement of real data in the adjacency matrix provided in an embodiment of the present application; Figure 6 A schematic diagram of the placement of imaginary data in the adjacency matrix provided in an embodiment of the present application; Figure 7 A schematic diagram of a process for merging and accumulating data within a group provided in an embodiment of the present application; Figure 8 A schematic diagram of the merge-accumulate time-sharing alignment strategy provided in an embodiment of the present application; Figure 9 A schematic diagram of the data processing pipeline of the second accumulator unit provided in an embodiment of the present application; Figure 10 A schematic diagram of a solution for updating the position of a value of a register group according to a decimation rate provided in an embodiment of the present application; Figure 11 A schematic diagram of a flow chart of an input data preprocessing module, a first accumulator unit, a second accumulator unit, and a multiplication-addition module in a pipeline mode in a certain scenario provided by an embodiment of the present application; Figure 12 Provided for this application Figure 11 The working principle diagram of the pipeline architecture shown; Figure 13 A schematic diagram of a flow chart of a time-domain matched filtering method provided in an embodiment of the present application; Figure 14 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0026] The following will describe the technical solutions in the embodiments of the present application in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present application and are therefore only examples and cannot be used to limit the scope of protection of the present application.

[0027] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned figure descriptions are intended to cover non-exclusive inclusions.

[0028] In the description of the embodiments of this application, the technical terms "first" and "second" are used only to distinguish different objects and should not be understood to indicate or imply relative importance or implicitly specify the quantity, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, the meaning of "plurality" is more than two, unless otherwise clearly and specifically defined.

[0029] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0030] In the description of the embodiments of this application, the term "and / or" is simply a description of the association relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent the following three situations: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this document generally indicates that the associated objects are in an "or" relationship.

[0031] Currently, some related technologies use the Winogard algorithm to optimize time-domain matched filtering calculation schemes. By performing certain combinations, the final overall resource consumption is reduced. However, this algorithm is generally applicable to optimization calculations after data is fully acquired and cannot be applied to matched filtering schemes in the case of real-time uninterrupted data flow. In addition, some related technologies mainly optimize calculation schemes for multiple groups of filters simultaneously. The optimization strategy is to optimize the common parts between different groups. However, this algorithm can only be applied to cases where the number of sampling points of the swept frequency signal is small, and cannot be applied to cases where the swept frequency signal has an extremely long number of points (for example, tens of thousands of points).

[0032] Based on this, an embodiment of the present application provides an FPGA implementation solution for a time-domain matched filter. The filter sets an accumulation module and a multiplication-addition module to first perform layered accumulation processing on the data, and then perform multiplication-addition operations, which can reduce the number of multiplication operations, reduce FPGA hardware resource consumption, and improve resource utilization efficiency; on the other hand, by accumulating the input data to be multiplied with the same filter parameter value in advance, the originally repeated multiplication operations are merged, and then the accumulated results are multiplied, which effectively reduces the number of times the multiplier and adder are used, and reduces the consumption of hardware resources.

[0033] Before introducing the above-mentioned time-domain matched filter, the filtering scheme of the time-domain matched filter is introduced first: Time-domain matched filtering is a signal processing technique used for signal detection, recognition, and feature extraction. It multiplies the received signal with a known reference signal point by point in the time domain and accumulates the result. The result is used to measure the similarity between the received signal and the reference signal.

[0034] This solution is aimed at the application of matched filtering in complex scenarios, so the input data of the above time domain matched filter can be expressed as , the filter parameters can be expressed as ,in, Indicates the sampling point number, is the number of sampling points. The time domain matched filtering solutions in the prior art include: 1. Calculate the product of each sampling point in the input signal and its corresponding filter parameter, that is:

[0035] This step mainly calculates 、 、 and ; Second, the accumulation is then performed to obtain the real and imaginary results of the matched filter, namely:

[0036]

[0037] The above-mentioned solution of performing multiplication operation first and then addition operation needs to consume a large number of multipliers, and the existing computing architecture is difficult to meet its multiplier requirements.

[0038] The time domain matched filter is introduced as follows: See Figure 1 An embodiment of the present application provides a time-domain matched filter 100 , comprising: an accumulation module 110 and a multiplication-addition module 120 connected to the accumulation module 110 .

[0039] The accumulation module 110 and the multiplication and addition module 120 are used to calculate the time domain matched filtering results of the input data and the filter parameters, where the input data includes real data and imaginary data, and the filter parameters are quantized filter parameters, where: The accumulation module 110 includes a first accumulation unit 111 and a second accumulation unit 112, wherein: The first accumulating unit 111 is configured to respectively accumulate the values ​​in the plurality of target data sets and send the obtained plurality of first accumulated results to the second accumulating unit 112; wherein each target data set includes real data and / or imaginary data to be multiplied by the same first filter parameter value to obtain a real part result in the time-domain matched filtering result, and to be multiplied by the same second filter parameter value to obtain an imaginary part result in the time-domain matched filtering result; wherein the first filter parameter value is a filter parameter value used to calculate the real part result, and the second filter parameter value is a filter parameter value used to calculate the imaginary part result; a second accumulating unit 112 configured to accumulate first accumulated results to be multiplied by the same first filtering parameter value to obtain a real part result among the plurality of first accumulated results to obtain a second accumulated result, accumulate first accumulated results to be multiplied by the same second filtering parameter value to obtain an imaginary part result among the plurality of first accumulated results to obtain a third accumulated result, and send the second accumulated result and the third accumulated result to the multiplication-addition module 120; The multiplication and addition module 120 is used to multiply the second accumulated result by the first filtering parameter value and accumulate the multiplied results to obtain a real part result, and multiply the third accumulated result by the second filtering parameter value and accumulate the multiplied results to obtain an imaginary part result.

[0040] The above-mentioned quantization refers to the process of converting the continuous or discrete signal amplitude into a finite number of discrete values, and quantizing the filter parameters refers to converting the filter parameters from the original data into fixed-point numbers of the target bit width. During the quantization process, several filter parameters can be quantized into one filter parameter, and the real data or imaginary data that need to be multiplied by the filter parameter can also be accumulated in advance, thereby reducing the number of multipliers used. The above-mentioned filter parameters may include real parameters and imaginary parameters. Generally, the real parameters and imaginary parameters will use the same general quantization method, for example, both use 9-bit quantization. The above-mentioned first filter parameter value and the second filter parameter value both refer to the filter parameter values ​​after quantization, and the first and the second are only to distinguish the filter parameter values ​​used to calculate the real part result and the filter parameter values ​​used to calculate the imaginary part result. Taking a matched filter with 25,000 filter parameters as an example, the distribution of the single component of the filter coefficient before quantization is as follows: Figure 2 As shown ( Figure 2The horizontal axis represents the filter coefficient number, and the vertical axis represents the filter parameter value). After the filter parameters are quantized to nine bits, the filter parameters are allocated to 512 fixed-point numbers. The distribution of the filter parameters on each fixed-point number after quantization is shown in the figure below. Figure 3 As shown ( Figure 3 The horizontal axis represents the quantized filter parameter, and the vertical axis represents the number of sampling points that match the same quantized filter parameter). Figure 3 Taking the leftmost filter parameter as an example, over 300 pre-quantized filter parameters are assigned to the quantized filter parameter. In subsequent calculations, all data that needs to be multiplied by these 300 pre-quantized filter parameters can be accumulated in advance before being multiplied by the quantized filter parameter. This example shows that parameter quantization combined with the accumulation-then-multiplication approach in the above solution can save a large number of multipliers.

[0041] In addition, it can be understood that the distribution of the numerical points of the real and imaginary parameters quantized by the same quantization method is also the same, and the filter parameter values ​​are symmetrically distributed compared to the zero point. Based on the above characteristics of the filter parameter values, it can be seen that the filter parameter and filter parameters There are only positive and negative signs in the values, but based on the symmetrical distribution of the filter parameter values ​​compared to the zero point, the filter parameters Negative filter parameters is still the quantized filter parameter value. Therefore, for the above calculation formula , when the first accumulating unit 111 performs calculation, the imaginary data is grouped to be filtered with the first parameter value Multiply to obtain the real part of the result, and wait for the second filter parameter value Multiply to obtain the imaginary part of the target data set. That is, the calculation formula The negative sign is applied to the filter parameter value part, so the multiplication and addition module 120 only needs to perform multiplication and addition operations, and no subtraction operation is required.

[0042] The overall working principle of the above-mentioned time domain matched filter 100 is introduced below: Different from the related art method of performing multiplication operation first and then accumulation operation, the above-mentioned time domain matched filter 100 adopts the method of performing accumulation operation first and then multiplication operation. Specifically: First, using the first accumulating unit 111, the real data and / or the imaginary data that need to be multiplied with the same first filtering parameter value to obtain the real part result of the time-domain matched filtering result and are to be multiplied with the same second filtering parameter value to obtain the imaginary part result of the time-domain matched filtering result are accumulated; Then, the second accumulating unit 112 accumulates the first accumulated results to be multiplied by the same first filtering parameter value to obtain the real part result among the plurality of first accumulated results to obtain a second accumulated result, and accumulates the first accumulated results to be multiplied by the same second filtering parameter value to obtain the imaginary part result among the plurality of first accumulated results to obtain a third accumulated result; Finally, the multiplication and addition module 120 multiplies the second accumulated result by the first filtering parameter value and accumulates the multiplied results to obtain a real part result, and multiplies the third accumulated result by the second filtering parameter value and accumulates the multiplied results to obtain an imaginary part result. It can be seen from the above working principle that compared with the related art operation method of performing multiplication operation first and then accumulation operation, the time domain matched filter 100 provided in the embodiment of the present application adopts the method of performing accumulation operation first and then multiplication operation, which can save a lot of multiplier resources.

[0043] The working principle of the first accumulating unit 111 is described below: See Figure 4 , the calculation method of the real part of the time domain matched filtering result is:

[0044] The imaginary part of the result is calculated as:

[0045] Depend on Figure 4 As can be seen from the above calculation formula, before multiplying with the corresponding filter parameter value, there are overlapping parts in the calculation of the real part result, the calculation of the imaginary part result, and the calculation between the two, including: (1) Need to multiply by the same real part parameter at the same time And need to multiply by the same imaginary parameter at the same time The accumulated result of the real part data , when calculating the real and imaginary results, you can use the real part parameters and imaginary parameters Multiplying them, we can get the real part of the result The term and the imaginary part of the result item; (2) Need to multiply the negative of the same imaginary parameter at the same time And need to multiply by the same real part parameter at the same time The cumulative result of the imaginary part data When calculating the real and imaginary results, you can use the negative of the imaginary parameter and real part parameters Multiplying them, we can get the real part of the result The term and the imaginary part of the result item; (3) As mentioned above, the real and imaginary parameters of the filter parameters will be quantized in a common way. Then the parameter values ​​of the real and imaginary parameters after quantization are also consistent, that is, the real parameters mentioned above are and imaginary parameters The parameter value points are the same, then for the real part result, if the real part parameter with the negative of the imaginary parameter When the same value is taken, wait for the real part parameter Multiply and wait for the negative of the imaginary parameter Real data to be multiplied and imaginary data Accumulation is also possible.

[0046] Therefore, the real data and the imaginary data can be divided into multiple target data sets, and the target data sets include real data and / or imaginary data to be multiplied with the same first filter parameter value to obtain the real result in the time domain matched filtering result, and to be multiplied with the same second filter parameter value to obtain the imaginary result in the time domain matched filtering result.

[0047] The working principle of the first accumulating unit 112 is described below: Since the purpose of the accumulation module 110 is to accumulate the real data and / or imaginary data that need to be multiplied with the same filter parameter value, based on the first accumulation result obtained by the first accumulation unit 112, the second accumulation unit can accumulate all the first accumulation results in the first accumulation result that need to be multiplied with the same first filter parameter value to obtain the real part result, and accumulate all the first accumulation results in the first accumulation result that need to be multiplied with the same second filter parameter value to obtain the imaginary part result.

[0048] Optionally, the target data set is determined by a relationship graph; the relationship graph includes a first node and a second node, the first node is used to represent a first filter parameter value, the second node is used to represent a second filter parameter value, and an edge connecting the first node and the second node is used to represent real data and / or imaginary data to be multiplied with the same first filter parameter value to obtain a real part result in a time-domain matched filtering result, and to be multiplied with the same second filter parameter value to obtain an imaginary part result in a time-domain matched filtering result; For example, the above relationship diagram refers to a relationship diagram in graph theory. The following describes the method for obtaining the above target data set using an adjacency matrix as an example: See Figure 5 and Figure 6 , understandably, Figure 5 and Figure 6The same adjacency matrix is ​​shown. Since the real and imaginary parameters use the same quantization method, the horizontal and vertical coordinate values ​​of the adjacency matrix are both quantized parameter values. The difference between the two is the way the real and imaginary data are put into the adjacency matrix. Specifically: Figure 5 As shown, the real data Real part parameter As the horizontal axis, the imaginary parameter For the ordinate, put the adjacency matrix; Figure 6 As shown, the imaginary data Negative number of imaginary parameter is the horizontal axis, with real part parameter Put the adjacency matrix in for the ordinate.

[0049] Each position in the adjacency matrix corresponds to a target data set, which corresponds to the set of edges connecting two identical nodes in the relationship graph.

[0050] For example, the relationship between the real data or imaginary data and the filter parameter value to be multiplied can be considered as a mapping relationship, and this mapping relationship can be determined by the relationship between the input data and the filter parameter value to be multiplied by the input data. When the filter parameters are quantized, the above mapping relationship also needs to be updated. When constructing the target data set, the target data set to which the real data or imaginary data belongs can be directly determined using the updated mapping relationship.

[0051] It is understandable that if Figure 5 and Figure 6 As shown, in the first accumulation results obtained by the first accumulation unit 111, the first accumulation results in the same column are the first accumulation results that need to be multiplied with the same first filter parameter value to obtain the real part result, and the first accumulation results in the same row are the second accumulation results that need to be multiplied with the same second filter parameter value to obtain the imaginary part result.

[0052] The above scheme obtains the target data set based on graph theory, and can directly determine the set to which it belongs by analyzing the edges between nodes, which is beneficial to improving the grouping efficiency; on the other hand, it reduces the number of sets that need to be multiplied, thereby reducing the demand for multiplier resources, which is beneficial to improving the resource utilization of the above-mentioned time domain matched filter 100; on the other hand, the set division method based on graph theory can flexibly construct the corresponding graph structure and group it according to the coefficient distribution and quantization characteristics of different filters, which is beneficial to improving the flexibility of the above-mentioned time domain matched filter 100.

[0053] Optionally, the above-mentioned relationship graph includes at least one of a relationship graph represented by an adjacency matrix, a relationship graph represented by an adjacency list, a relationship graph represented by a cross-linked list, and a relationship graph represented by an adjacency multi-list.

[0054] In the above scheme, different forms of relationship diagrams can be used to determine the target data set. Different relationship diagram representation methods provide a variety of data organization and processing methods, which makes it convenient to select appropriate methods according to actual needs and data characteristics, enhance the flexibility of data processing, and enable the above-mentioned time domain matched filter to adapt to a variety of application scenarios.

[0055] Optionally, the time-domain matched filter 100 further includes: The input data preprocessing module 130 is connected to the accumulation module 110 and is used to perform a decimation operation on the input data to match the data flow rate of the input data with the processing rate of the time domain matched filter; and store the decimated input data into the register group.

[0056] The above extraction operation refers to selecting data points at a certain interval from the original input data to reduce the amount of data. The reasons for the extraction operation include: (1) Reduce data rate: The rate of the original data stream may be high, especially when processing long-distance high-bandwidth signals. By extracting data, the amount of data that needs to be processed can be reduced, reducing the demand for hardware resources; (2) Adapting the filter's computational capacity: The computational capacity of the time-domain matched filter 100 is limited. The decimation operation can be used to match the data flow rate of the input data with the filter's processing capacity, thereby enabling the input data to be processed in real time by the time-domain matched filter 100. (3) Reduce storage requirements: On the one hand, the amount of input data itself can be reduced through the extraction operation, thereby reducing the need for register groups required to cache the input data; on the other hand, the data flow rate of the input data can be matched with the processing capability of the filter through the extraction operation. For example, within one cycle, through the data extraction operation, the extracted data is input into the filter, so that the filter can output the time domain matched filtering result within the cycle. Therefore, there is no need to store the data of the next cycle in advance to wait for the filter to process the data of the current cycle.

[0057] For example, the above solution can use a high-frequency clock based on a phase-locked loop in conjunction with a first-in-first-out queue to implement a frequency up-conversion operation. The purpose of the frequency up-conversion operation is: (1) Adapting filter computing power: The upscaling operation can match the data flow rate of the input data with the processing power of the filter, so that the input data can be processed in real time by the time domain matched filter 100.

[0058] (2) By increasing the frequency and aligning the extraction rate, the subsequent computing modules can be transformed from multiple points at low frequencies to one point at high frequencies, thus reducing the number of parallel computing modules.

[0059] The input data preprocessing module 130 in the above scheme performs a decimation operation on the input data so that the data flow rate of the input data matches the processing rate of the time-domain matched filter, thereby enabling the input data to be efficiently processed by the time-domain matched filter 100, avoiding data loss or processing delay due to rate mismatch; on the other hand, performing a decimation operation on a large amount of input data can remove redundant information in the input data, reduce the amount of input data while retaining the key features of the input data. The decimation operation can improve the data quality of the input data, thereby improving the matched filtering effect of the above-mentioned time-domain matched filter 100.

[0060] Optionally, the input data preprocessing module 130 , the accumulation module 110 and the multiplication-addition module 123 calculate the time-domain matched filtering result of the input data and the filtering parameters in a pipeline mode.

[0061] Pipelining is a method of breaking down a complex task into multiple stages and using different hardware or software resources to process different parts in parallel at each stage. Pipelining is widely used in DSP (Digital Signal Processing) and FPGA (Field Programmable Gate Array) design to improve system throughput and efficiency.

[0062] The above scheme divides input data preprocessing, data grouping accumulation, and data multiplication and addition into four stages for parallel processing. Each stage is independent and closely connected, and data can be transmitted in sequence. There is no need to wait for all data collection to be completed before processing, which is beneficial to improving the data processing efficiency of the above-mentioned time-domain matched filter 100; on the other hand, different hardware resources can be allocated to different stages. Compared with the non-pipeline architecture, the above scheme can significantly improve resource utilization; on the other hand, the time-domain matched filter 100 can receive and process input processing in real time and continuously, which is beneficial to improving the real-time and continuity of data processing of the time-domain matched filter.

[0063] Optionally, the first accumulating unit 111 and the second accumulating unit 112 perform the accumulation operation in a merge accumulation manner.

[0064] The above-mentioned merging and accumulation is a strategy that reduces the amount of calculation by gradually merging data. It is applicable to scenarios where a large amount of data needs to be accumulated. Figure 7 For example, there are 7 data items in a group that need to be accumulated. The merging and accumulation process includes: (1) Perform the following cumulative operations simultaneously: Accumulate data 1 and data 2 to obtain data 1+2, accumulate data 3 and data 4 to obtain data 3+4, and accumulate data 5 and data 6 to obtain data 5+6; (2) Perform the following cumulative operations simultaneously: Accumulate data 1+2 and data 3+4 to obtain data 1+2+3+4, and accumulate data 5+6 and data 7 to obtain data 5+6+7; (3) Perform the following accumulation operations: The data 1+2+3+4 and the data 5+6+7 are added to obtain the data 1+2+3+4+5+6+7.

[0065] From the above merge-accumulate process, it can be seen that the merge-accumulate method can decompose large-scale data accumulation operations into multiple operations, so that the accumulation operations can be processed in parallel, thereby speeding up the operation.

[0066] The first accumulation unit 111 and the second accumulation unit 112 in the above scheme can perform accumulation operations in a merge-and-accumulate manner, which can reduce the number of operations and thus optimize the calculation process of the above-mentioned time-domain matched filter 100, which is beneficial to improving the calculation efficiency of the above-mentioned time-domain matched filter 100; on the other hand, the merge-and-accumulate can reduce the risk of error accumulation and is beneficial to improving the matched filtering effect of the above-mentioned time-domain matched filter 100; on the other hand, the merge-and-accumulate groups the data for processing, which is beneficial to improving the utilization rate of parallel computing resources and thus improving the utilization rate of computing resources of the above-mentioned time-domain matched filter 100.

[0067] It is understandable that the actual operation is completed in a merge-accumulation manner. The clock cycle required for data is Since the number of data in each set is different, for example, some sets may have more than 300 data that need to be merged and accumulated, while some sets may have only a few data that need to be merged and accumulated, the clock cycle for each set to obtain its accumulation result may also be different. Based on this, the embodiment of the present application provides the following solution: Optionally, the first accumulation unit 111 is further configured to: merge first accumulation results of accumulation operations completed within the same cycle and send the combined results to the second accumulation unit; wherein the completion cycle of the accumulation operation is obtained based on the amount of data to be accumulated.

[0068] For example, the strategy of combining the accumulation results of the accumulation operations completed in the same cycle and sending them to the second accumulation unit 112 in the above solution can be called a time-sharing alignment strategy. The working principle of the time-sharing alignment strategy is described below: See Figure 8 In a certain scenario, the first accumulating unit 111 needs to target data sets ( Figure 8 The target data set is called a group) and the data in it is merged and accumulated. Assume : The amount of data that needs to be merged and accumulated in group 1 is The merged sum of group 1 will be in the clock cycle, that is, in the clock cycles completed; The amount of data that needs to be merged and accumulated in group 2 is 2, and the merging and accumulation of group 2 will be done in the clock cycles, that is, completed in the first clock cycle; The amount of data that needs to be merged and accumulated in group 3 is The merged sum of group 3 will be in the clock cycle, that is, in the clock cycles completed; The amount of data that needs to be merged and accumulated in group 4 is 4, and the merging and accumulation of group 4 will be done in the clock cycle, that is, in the clock cycles completed; The amount of data that needs to be merged and accumulated in group 5 is The merged sum of group 5 will be in the clock cycle, that is, in the clock cycles completed; Then, at the end of the first clock cycle, group 2 completes the merge accumulation and sends the merge accumulation result of group 2 to the second accumulation unit 112. The second accumulation unit 112 starts processing the accumulation result after receiving the accumulation result of the first accumulation unit 111 in the first clock cycle. At the end of the second clock cycle, the merged accumulation of group 3 is completed, so the merged accumulation result of group 3 is first sent to the second accumulation unit 112. After receiving the accumulation result of the first accumulation unit 111 in the second clock cycle, the second accumulation unit 112 continues to process the accumulation result. In the At the end of the first clock cycle, group 1 and group 5 complete the merged accumulation, and the accumulated result of group 1 and the accumulated result of group 5 can be combined and sent to the second accumulation unit 112. The second accumulation unit 112 receives the accumulated result of the first accumulation unit 111 at the first clock cycle. After the accumulated results of the clock cycles are obtained, the accumulated results are processed; In the At the end of the first clock cycle, group 3 completes the merge accumulation, and the accumulation result of group 3 can be sent to the second accumulation unit 112. The second accumulation unit 112 receives the result of the first accumulation unit 111 at the first clock cycle. After the accumulated results of the clock cycles are collected, the accumulated results are processed.

[0069] Also, see Figure 9 The second accumulation unit 112 does not receive all the first accumulation results obtained by the first accumulation unit 111 at one time, but receives the first accumulation results obtained by the first accumulation unit 111 at the end of the current clock cycle in sequence according to the clock cycle, and then the second accumulation unit 112 processes the first accumulation result newly sent by the first accumulation unit 111 in the next clock cycle. When the next clock cycle arrives, the second accumulation unit 112 has half of the data to be processed, and then continues to accumulate all the data after receiving the new data sent by the first accumulation unit 111. In this way, data can be processed in real time and uninterruptedly, thereby improving the data processing efficiency of the above-mentioned time domain matched filter 100.

[0070] Compared with the scheme of direct beat synchronization, the above scheme adopts a time-sharing alignment strategy, which can reasonably arrange data transmission and processing according to the actual accumulation time of each target data set, avoiding excessive allocation of registers for waiting for synchronization, thereby greatly reducing the register consumption of the matched filtering process, which is beneficial to improving the resource utilization of the above-mentioned time domain matched filter 100; on the other hand, the time-sharing alignment strategy can enable the second accumulation unit 112 to further process the calculation results of different target data sets that have completed accumulation calculations in the first accumulation unit 111 in the same cycle, reducing the idle waiting time of the second accumulation unit 112, which is beneficial to improving the computational efficiency of the above-mentioned time domain matched filter 100; on the other hand, because the time-sharing alignment strategy avoids the large amount of register consumption caused by direct beat synchronization, it reduces the timing problems caused by unreasonable register distribution or too long data transmission path, which is beneficial to improving the overall performance of the above-mentioned time domain matched filter 100.

[0071] Optionally, the second accumulation unit 112 is further used to: merge the second accumulation result and / or the third accumulation result completed in the same cycle and send them to the multiplication and addition module 120; wherein the completion cycle of the accumulation operation is obtained based on the amount of data to be accumulated.

[0072] It can be understood that since the amount of data in each group in the second accumulation unit 112 may also be different, the accumulated data obtained by the second accumulation unit 112 in each clock cycle can also be merged and sent to the multiplication and addition module 120. Its working principle is similar to the working principle of the first accumulation unit 111 merging the data and sending it to the second accumulation unit 112, and will not be repeated in the embodiments of the present application.

[0073] Compared with the scheme of direct beat synchronization, the above scheme adopts a time-sharing alignment strategy, which can reasonably arrange data transmission and processing according to the actual time for each target data set to complete accumulation, avoiding excessive allocation of registers for waiting for synchronization, thereby greatly reducing the register consumption of the matched filtering process, which is beneficial to improving the resource utilization of the above-mentioned time domain matched filter 100; on the other hand, the time-sharing alignment strategy can enable the multiplication and addition module 120 to further process the calculation results of the different target data sets in the second accumulation unit 112 that have completed accumulation calculations in the same cycle, reducing the idle waiting time of the multiplication and addition module 120, which is beneficial to improving the computational efficiency of the above-mentioned time domain matched filter 100; on the other hand, because the time-sharing alignment strategy avoids the large amount of register consumption caused by direct beat synchronization, it reduces the timing problems caused by unreasonable register distribution or too long data transmission path, which is beneficial to improving the overall performance of the above-mentioned time domain matched filter 100.

[0074] The following describes an implementation of a time-domain matched filter 100 using a pipeline mode in a certain scenario. In this scenario, the input data preprocessing module 130, the first accumulator 111 of the accumulator 110, the second accumulator 112 of the accumulator 110, and the multiplication and addition module 120 use a pipeline mode to calculate the time-domain matched filtering result of the input data and the filter parameters. Figure 11 , the four stages of the pipeline include: (1) Phase 1: The input data preprocessing module 130 inputs the sampling result of the analog-to-digital converter ADC, that is, the input data, into the register group, and the register group updates the data position according to the sampling rate. For example, see Figure 10 , using a 2-out-of-1 extraction strategy, then in each clock cycle, each value in the register group is assigned to the register two positions behind it.

[0075] (2) Second stage: The first accumulation unit 111 in the accumulation module 110 sends the real data and the imaginary data to the corresponding round group (i.e., the target data set) according to the register position and the predetermined grouping relationship, and accumulates the data in each group. At the end of the clock cycle, the accumulation result obtained in the clock cycle is transmitted to the second accumulation unit 112.

[0076] (3) The third stage: The second accumulation unit 112 in the accumulation module 110 sends the accumulation result of one round group to the corresponding second round group, and accumulates the data in each group, and the accumulated result is sent to the multiplication and addition module 120; The data included in each two-round group includes: the first accumulated result among multiple first accumulated results to be multiplied with the same first filter parameter value to obtain the real part result, or the first accumulated result among multiple first accumulated results to be multiplied with the same second filter parameter value to obtain the imaginary part result.

[0077] (4) The fourth stage: The multiplication and addition module 120 multiplies the accumulated results of the two rounds by the corresponding filtering parameter values ​​and accumulates them to obtain the final time-domain matched filtering result.

[0078] See Figure 2 The above-mentioned first round group refers to the target data set in the above description, and the second round group refers to the first accumulated result that needs to be multiplied with the same first filter parameter value to obtain the real part result and the first accumulated result that needs to be multiplied with the same second filter parameter value to obtain the imaginary part result.

[0079] It is understood that the four stages described above utilize a pipelined approach for data processing. When the sampling rate exceeds the number of ADC samples in a single cycle, this four-stage pipeline architecture can support continuous input and output of time-domain matched filtering results. If the number of samples in a single cycle exceeds the sampling rate, multiple sets of these pipeline architectures can be used for parallel processing.

[0080] See Figure 13 Based on the same inventive concept, an embodiment of the present application further provides a time-domain matched filtering method, the method comprising: Step S210: obtaining input data and filtering parameters; wherein the input data includes real data and imaginary data; and the filtering parameters are quantized filtering parameters; Step S220: Accumulate the values ​​in the multiple target data sets respectively to obtain a first accumulation result; wherein each target data set includes real data and / or imaginary data to be multiplied with the same first filter parameter value to obtain the real part result in the time domain matched filtering result, and to be multiplied with the same second filter parameter value to obtain the imaginary part result in the time domain matched filtering result; accumulate the first accumulation results to be multiplied with the same first filter parameter value to obtain the real part result in the multiple first accumulation results to obtain the second accumulation result; accumulate the first accumulation results to be multiplied with the same second filter parameter value to obtain the imaginary part result in the multiple first accumulation results to obtain the third accumulation result Step S230: multiplying the second accumulated result by the first filtering parameter value and accumulating the multiplied results to obtain a real part result; multiplying the third accumulated result by the second filtering parameter value and accumulating the multiplied results to obtain an imaginary part result.

[0081] The functions that can be achieved by the above-mentioned time-domain matched filtering method are the same as those of the above-mentioned time-domain matched filter 100. For the implementation schemes of other functions, please refer to the introduction of the time-domain matched filter 100 in the above content, which will not be repeated in the embodiments of this application.

[0082] Based on the same inventive concept, an embodiment of the present application further provides a fiber optic sensing system, including the above-mentioned time domain matched filter 100.

[0083] For example, the fiber optic sensing system can be used for at least one of temperature monitoring, strain monitoring, and disturbance monitoring. Major types of fiber optic sensing systems include distributed fiber optic vibration sensing systems (DVS), distributed fiber optic acoustic wave monitoring systems (DAS), distributed temperature sensing systems (DTS), distributed fiber optic strain and temperature measurement systems (DTSS), and Brillouin distributed sensors.

[0084] Figure 14 This is a schematic diagram of an electronic device provided in an embodiment of the present application. Figure 14 The electronic device 300 includes a processor 310, a memory 320, and a communication interface 330. These components are interconnected and communicate with each other via a communication bus 340 and / or other forms of connection mechanisms (not shown).

[0085] The memory 320 includes one or more (only one is shown in the figure), which may be, but is not limited to, random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), and electrically erasable programmable read-only memory (EEPROM). The processor 310 and other possible components can access the memory 320 and read and / or write data therein.

[0086] The processor 310 includes one or more (only one is shown in the figure), which can be an integrated circuit chip with signal processing capabilities. The above-mentioned processor 310 can be a general-purpose processor, including a central processing unit (CPU), a microcontroller unit (MCU), a network processor (NP), or other conventional processors; it can also be a special-purpose processor, including a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0087] Communication interface 330 includes one or more (only one is shown in the figure) interfaces that can be used to communicate directly or indirectly with other devices to exchange data. For example, communication interface 330 can be an Ethernet interface; a mobile communication network interface, such as a 3G, 4G, or 5G network interface; or other types of interfaces capable of transmitting and receiving data.

[0088] One or more computer program instructions may be stored in the memory 320 , and the processor 310 may read and execute these computer program instructions to implement the time-domain matched filtering method and other desired functions provided in the embodiments of the present application.

[0089] I understand. Figure 14 The structure shown is for illustration only. The electronic device 300 may also include Figure 14 More or fewer components than shown, or with Figure 14 Different configurations shown. Figure 14 Each component shown in the figure can be implemented using hardware, software, or a combination thereof. For example, the electronic device 300 can be a single server (or other device with computing processing capabilities), a combination of multiple servers, a cluster of a large number of servers, etc., and can be both a physical device and a virtual device.

[0090] The present application also provides a computer-readable storage medium having computer program instructions stored thereon. When the computer program instructions are read and executed by a computer processor, the time-domain matched filtering method provided by the present application is executed. For example, the computer-readable storage medium can be implemented as Figure 14 The memory 320 in the electronic device 300.

[0091] In the embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interface, the indirect coupling or communication connection of the device or unit can be electrical, mechanical or other forms.

[0092] In addition, the units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0093] Furthermore, the functional modules in each embodiment of the present application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0094] The above are merely examples of the present application and are not intended to limit the scope of protection of the present application. Those skilled in the art will appreciate that various modifications and variations are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present application shall be included within the scope of protection of the present application.

Claims

1. A time-domain matched filter, characterized in that: include: An accumulation module and a multiplication-addition module connected to the accumulation module; The accumulation module and the multiplication-addition module are used to calculate the time-domain matched filtering results of the input data and the filtering parameters, the input data includes real data and imaginary data, and the filtering parameters are quantized filtering parameters, wherein: The accumulation module includes a first accumulation unit and a second accumulation unit, wherein: The first accumulator is configured to respectively accumulate the values ​​in the plurality of target data sets and send the obtained plurality of first accumulated results to the second accumulator; wherein each of the target data sets includes real data and / or imaginary data to be multiplied by the same first filter parameter value to obtain the real part result in the time-domain matched filtering result, and to be multiplied by the same second filter parameter value to obtain the imaginary part result in the time-domain matched filtering result; wherein the first filter parameter value is the filter parameter value used to calculate the real part result; and the second filter parameter value is the filter parameter value used to calculate the imaginary part result; the second accumulating unit being configured to accumulate the first accumulating results to be multiplied by the same first filtering parameter value to obtain the real part result among the plurality of first accumulating results to obtain a second accumulating result, and to accumulate the first accumulating results to be multiplied by the same second filtering parameter value to obtain the imaginary part result among the plurality of first accumulating results to obtain a third accumulating result, and to send the second accumulating result and the third accumulating result to the multiplication-addition module; The multiplication and addition module is used to multiply the second accumulated result by the first filtering parameter value and accumulate the multiplied results to obtain the real part result, and multiply the third accumulated result by the second filtering parameter value and accumulate the multiplied results to obtain the imaginary part result.

2. The time-domain matched filter according to claim 1, wherein The time domain matched filter also includes: An input data preprocessing module is connected to the accumulation module and is used to perform a decimation operation on the input data so that the data flow rate of the input data matches the processing rate of the time domain matched filter; and store the input data after the decimation operation into a register group.

3. The time-domain matched filter according to claim 2, wherein The input data preprocessing module, the accumulation module and the multiplication and addition module calculate the time domain matched filtering result of the input data and the filtering parameters in a pipeline mode.

4. The time-domain matched filter according to claim 1, wherein The first accumulating unit and the second accumulating unit perform an accumulation operation by adopting a merge accumulation method.

5. The time-domain matched filter according to claim 1, wherein The first accumulating unit is further configured to: The first accumulation results of the accumulation operations completed within the same cycle are combined and sent to the second accumulation unit; wherein the completion cycle of the accumulation operation is obtained based on the amount of data to be accumulated.

6. The time-domain matched filter according to claim 1, wherein The second accumulating unit is further configured to: The second accumulation result and / or the third accumulation result that complete the accumulation operation in the same cycle are combined and sent to the multiplication and addition module; wherein the completion cycle of the accumulation operation is obtained based on the amount of data to be accumulated.

7. The time-domain matched filter according to any one of claims 1 to 6, characterized in that The target data set is determined by a relationship graph; The relationship graph includes a first node and a second node, the first node is used to represent the first filtering parameter value, the second node is used to represent the second filtering parameter value, and the edge connecting the first node and the second node is used to represent real data and / or imaginary data to be multiplied with the same first filtering parameter value to obtain a real result in the time-domain matched filtering result, and to be multiplied with the same second filtering parameter value to obtain an imaginary result in the time-domain matched filtering result; The target data set includes a set of edges connecting the same first node and the same second node.

8. The time-domain matched filter according to claim 7, characterized in that The relationship graph includes at least one of a relationship graph represented by an adjacency matrix, a relationship graph represented by an adjacency list, a relationship graph represented by a cross-linked list, and a relationship graph represented by an adjacency multi-list.

9. A time-domain matched filtering method, characterized in that: The method comprises: Acquire input data and filtering parameters; wherein the input data includes real data and imaginary data; and the filtering parameters are quantized filtering parameters; Accumulating values ​​in a plurality of target data sets respectively to obtain a first accumulation result; wherein each of the target data sets includes real data and / or imaginary data to be multiplied by a same first filter parameter value to obtain a real result in the time-domain matched filtering result, and to be multiplied by a same second filter parameter value to obtain an imaginary result in the time-domain matched filtering result; Accumulating the first accumulated results to be multiplied by the same first filtering parameter value to obtain the real part result among the plurality of first accumulated results to obtain a second accumulated result; Accumulating the first accumulation results to be multiplied by the same second filtering parameter value to obtain the imaginary part result among the plurality of first accumulation results to obtain a third accumulation result; multiplying the second accumulated result by the first filtering parameter value and accumulating the multiplied result to obtain the real part result; The third accumulated result is multiplied by the second filtering parameter value and the multiplied result is accumulated to obtain the imaginary part result.

10. A fiber optic sensing system, characterized in that: include: The time-domain matched filter according to any one of claims 1 to 8.

Images