Big data curve retrieval method and device based on curve fingerprints

Abstract

The invention provides a big data curve retrieval method and device based on curve fingerprints, and the method comprises the steps: determining a data fingerprint of a to-be-detected curve; performing matching check on the data fingerprint of the curve to be detected and all data fingerprints representing data in a pre-constructed index database, and determining the matching size; determining representative data with the highest matching performance with the curve to be detected, and reading a data block associated with the representative data with the highest matching performance; determining the matching between all the curve data fingerprints in the associated data blocks and the data fingerprint of the curve to be detected; sorting all the curve data fingerprints in the associated index database based on the matching degree; according to the method, the sorting result is output to the display terminal, data block matching is performed in the index database based on the data fingerprints, and then data fingerprint matching is further performed, so that matching retrieval of the to-be-detected curve can be quickly realized, the data retrieval efficiency is greatly improved, and the method has relatively high practicability.

Description

Technical Field

[0001] The present invention relates to the technical field of spacecraft environmental testing, and in particular to a big data curve retrieval method and device based on curve fingerprint. Background Art

[0002] Spacecraft environmental testing is an important test project in the process of spacecraft development. It verifies the design of the spacecraft so that the spacecraft can withstand various space environments and work normally throughout its life cycle. It also conducts environmental inspections on the manufacturing quality of the spacecraft to discover potential defects in materials, components, manufacturing processes, etc., thereby ensuring the reliability of the spacecraft's on-orbit operation. During the life cycle of the spacecraft, experiments and tests generate a large amount of data. At present, test data is still managed and applied separately according to professions and types. Various databases have high requirements for users, and the operation is cumbersome, and they cannot be obtained and queried efficiently and easily. The rapid response requirements for spacecraft design and testing are getting higher and higher. More than 80% of the design of new spacecraft needs to inherit successful experience, and historical data is the support. At the same time, there are inextricable connections between different professions and different models.

[0003] Therefore, how to quickly complete curve matching retrieval in a large amount of spacecraft environmental test data has become a technical problem that technical personnel in this field urgently need to solve. Summary of the invention

[0004] The present invention provides a big data curve retrieval method and device based on curve fingerprint, which is used to solve the defect in the prior art that it is impossible to quickly complete curve matching retrieval in a large amount of spacecraft environmental test data.

[0005] In a first aspect, the present invention provides a big data curve retrieval method based on curve fingerprint, comprising:

[0006] Determine the data fingerprint of the curve to be detected;

[0007] Performing a matching check on the data fingerprint of the curve to be detected and the data fingerprints of all representative data in the pre-built index library to determine the matching degree;

[0008] Determine the representative data with the highest matching degree with the curve to be detected, and read in the data block associated with the representative data with the highest matching degree;

[0009] Determining the matching of all the curve data fingerprints in the associated data block with the data fingerprint of the curve to be detected;

[0010] Based on the matching degree, all the curve data fingerprints in the associated index library are sorted;

[0011] The sorting result is output to a display terminal.

[0012] According to a big data curve retrieval method based on curve fingerprint provided by the present invention, before determining the data fingerprint of the curve to be detected, it also includes:

[0013] Preprocess the data in all spacecraft environmental test databases;

[0014] In the preprocessed data, a first curve is randomly selected to extract a data fingerprint as first representative data, and is stored in a first data block;

[0015] Selecting a preset number of curves having the smallest difference from the data fingerprint of the first representative data from the database, and saving them in the first data block;

[0016] Among the remaining curves in the database, select the curve having the largest difference from the data fingerprint of the first representative data as the second representative data;

[0017] Selecting a preset number of curves having the smallest difference from the data fingerprint of the second representative data from the database, and saving them in a second data block;

[0018] Until the data fingerprints of all the curves in the database are stored by at least one data block, an index library is obtained.

[0019] According to a big data curve retrieval method based on curve fingerprint provided by the present invention, the step of determining the data fingerprint of the curve to be detected includes:

[0020] Determine the data type of the curve to be detected, wherein the data type includes static data and dynamic data;

[0021] Based on the feature point fitting method of Douglas-Peucker algorithm, the static data is processed to obtain the corresponding data fingerprint;

[0022] Based on the secondary feature extraction method after frequency domain transformation, the dynamic data is processed to obtain the corresponding data fingerprint.

[0023] According to a big data curve retrieval method based on curve fingerprint provided by the present invention, the feature point fitting method based on the Douglas-Peucker algorithm processes the static data to obtain the corresponding data fingerprint, including:

[0024] Determine the original line elements of the static data;

[0025] Connect the starting point and the end point of the original line element to obtain a fitted straight line;

[0026] Determine the node farthest from the fitting straight line, and if the farthest distance is less than the tolerance, retain the node farthest from the fitting straight line;

[0027] The node with the longest distance is used as the starting point to obtain a new fitting straight line, until an expression with accuracy satisfying the tolerance limit is obtained as the data fingerprint.

[0028] According to a big data curve retrieval method based on curve fingerprint provided by the present invention, the dynamic data is processed based on the secondary feature extraction method after frequency domain transformation to obtain the corresponding data fingerprint, including:

[0029] Determine the data fingerprint length;

[0030] Dividing the dynamic data into equal segments corresponding to the length of the data fingerprint in the logarithmic domain;

[0031] The maximum value of the dynamic data in each of the equally divided segments is determined as the amplitude of the equally divided segments to form a data fingerprint of the dynamic data.

[0032] According to a big data curve retrieval method based on curve fingerprint provided by the present invention, before determining the data fingerprint of the curve to be detected, it also includes:

[0033] Perform amplitude normalization processing on the curve to be detected to obtain normalized data;

[0034] The normalized data is aligned by interpolating the logarithmic spacing to achieve data preprocessing.

[0035] According to a big data curve retrieval method based on curve fingerprint provided by the present invention, the amplitude normalization processing is performed on the curve to be detected to obtain normalized data, including:

[0036] Determine the acceleration response spectrum of the curve to be tested and the mechanical test input reference spectrum;

[0037] Inputting the acceleration response spectrum and the mechanical test input reference spectrum into a preset formula to obtain normalized data;

[0038] The preset formula is:

[0039]

[0040] Where Y(ω) is the acceleration response spectrum, G(ω) is the normalized data, and U(ω) is the mechanical test input reference spectrum.

[0041] According to a big data curve retrieval method based on curve fingerprint provided by the present invention, the matching check and determining the matching size include:

[0042] Determine the sequence length of the data fingerprint of the curve to be detected and the sequence length of the data fingerprint representing the data;

[0043] Based on the sequence length, a matrix grid is constructed, and the Euclidean distance between any two points in the matrix grid is determined as the matching size.

[0044] A big data curve retrieval method based on curve fingerprint provided by the present invention also includes:

[0045] The regularization function corresponding to the Euclidean distance satisfies the preset constraints;

[0046] The preset constraint conditions include: boundary conditions, continuity and monotonicity.

[0047] In a second aspect, the present invention further provides a big data curve retrieval device based on curve fingerprint, comprising:

[0048] A first determination module, used to determine the data fingerprint of the curve to be detected;

[0049] A first matching module is used to check the matching of the data fingerprint of the curve to be detected with the data fingerprints of all representative data in the pre-built index library to determine the matching degree;

[0050] A second determination module is used to determine the representative data with the highest matching degree with the curve to be detected, and read in the data block associated with the representative data with the highest matching degree;

[0051] A second matching module, used to determine the matching of all the curve data fingerprints in the associated data block with the data fingerprint of the curve to be detected;

[0052] A sorting module, used to sort all the curve data fingerprints in the associated index library based on the matching degree;

[0053] The output module is used to output the sorting result to a display terminal.

[0054] In a third aspect, the present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a big data curve retrieval method based on curve fingerprint as described above is implemented.

[0055] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described big data curve retrieval methods based on curve fingerprints.

[0056] In a fifth aspect, the present invention further provides a computer program product, comprising a computer program, which, when executed by a processor, implements any of the above-mentioned big data curve retrieval methods based on curve fingerprints.

[0057] The present invention provides a big data curve retrieval method and device based on curve fingerprint, comprising: determining the data fingerprint of the curve to be detected; performing a matching check on the data fingerprint of the curve to be detected and the data fingerprints of all representative data in a pre-built index library to determine the matching degree; determining the representative data with the highest matching degree with the curve to be detected, and reading in the data block associated with the representative data with the highest matching degree; determining the matching degree of all curve data fingerprints in the associated data block with the data fingerprint of the curve to be detected; sorting all curve data fingerprints in the associated index library based on the matching degree; outputting the sorting result to a display terminal, performing data block matching in the index library based on the data fingerprint, and further performing data fingerprint matching, so as to quickly realize the matching retrieval of the curve to be detected, greatly improve the data retrieval efficiency, and have strong practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0059] Figure 1 is a flow chart of a big data curve retrieval method based on curve fingerprint provided in this embodiment;

[0060] Figure 2 is a schematic diagram of the principle of the Douglas-Peucker algorithm provided in this embodiment;

[0061] Figure 3 This is a schematic diagram of the index library construction process provided in this embodiment.

[0062] Figure 4 is a schematic diagram of the structure of a big data curve retrieval device based on curve fingerprint provided in this embodiment;

[0063] Figure 5 It is a schematic diagram of the structure of the electronic device provided in this embodiment. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] Figure 1 is a flow chart of a big data curve retrieval method based on curve fingerprint provided in this embodiment, Figure 2 is a schematic diagram of the principle of the Douglas-Peucker algorithm provided in this embodiment, Figure 3 This is a schematic diagram of the index library construction process provided in this embodiment.

[0066] like Figure 1 As shown, the big data curve retrieval method based on curve fingerprint provided by the embodiment of the present invention mainly includes the following steps:

[0067] 101. Determine the data fingerprint of the curve to be detected.

[0068] In a specific implementation process, the curve to be tested is defined as the curve to be tested, that is, data matching the curve to be tested is searched in a large amount of historical test data. After the curve to be tested is determined, the fingerprint of the curve to be tested must be determined. The curve fingerprint refers to some curve characteristics that can characterize the curve as unique, specifically:

[0069] Determine the data type of the curve to be detected, which includes static data and dynamic data. Data such as temperature, humidity, and slowly changing pressure have a slow changing trend and belong to static data. For static data, the feature point fitting method based on the Douglas-Peucker (DP) algorithm is used to process it and obtain the corresponding data fingerprint.

[0070] like Figure 2 As shown in , the specific method of the characteristic point fitting method of the Douglas-Peucker (DP) algorithm is as follows: determine the original line elements of the static data, such as Figure 2 0 in , connect the starting point and the end point in the original line feature to get the fitted line a; determine the node c that is farthest from the fitted line a, and the distance from the node c to the line a is b. If the farthest distance b is less than the limit error D, then keep the farthest node, such as Figure 2 Then take the farthest node as the starting point to get a new fitting line, and repeat the steps as shown in Figure 2 3 in the equation until the accuracy meets the tolerance limit. Figure 2 4 in , as the data fingerprint.

[0071] Vibration, noise, impact and other test data have high sampling rates and large data volumes, and are dynamic data. For dynamic data, a secondary feature extraction method based on frequency domain transformation is used to process the dynamic data and obtain the corresponding data fingerprint.

[0072] Secondary feature extraction is performed on frequency domain data such as sine swept spectrum and random PSD spectrum. The calculation process is as follows: determine the data fingerprint length, that is, set a data fingerprint length L; divide the dynamic data into equal segments corresponding to the data fingerprint length in the logarithmic domain, that is, divide the current signal X(N) into L segments in the logarithmic domain, and the length of each segment is Determine the maximum value of the dynamic data X(N) in each equally divided segment as the amplitude of the equally divided segment to form a data fingerprint Pool of the dynamic data X (L).

[0073] Among them, before obtaining the data fingerprint, it is necessary to preprocess the curve to be detected, perform amplitude normalization processing on the curve to be detected, and obtain normalized data; then, the normalized data is aligned according to the logarithmic spacing interpolation method to realize data preprocessing.

[0074] Amplitude normalization, specifically, determining the acceleration response spectrum and mechanical test input reference spectrum of the curve to be tested; inputting the acceleration response spectrum and mechanical test input reference spectrum into a preset formula to obtain normalized data;

[0075] The default formula is (1):

[0076]

[0077] Where Y(ω) is the acceleration response spectrum, G(ω) is the normalized data, and U(ω) is the mechanical test input reference spectrum.

[0078] The data alignment method is to interpolate the data according to logarithmic spacing, and the interpolation method is conformal piecewise cubic Hermite interpolation:

[0079] a. In each subinterval x k ≤x≤x k+1 In the above, the polynomial P(x) is a cubic Hermite interpolation polynomial for a given data point, with a specified derivative (slope) at the interpolation point;

[0080] b. P(x) performs y interpolation, that is, P(x i )=y i , and the first-order derivative is continuous, but the second-order derivative may not be continuous;

[0081] c. Cubic interpolation P(x) has a shape-preserving effect. i The slope at is chosen in a specific way so that P(x) preserves the shape and corresponding monotonicity of the data, thus ensuring that the interpolated signal has exactly the same monotonicity and extreme values ​​as the original signal.

[0082] Since the frequency range may be different during each test, data alignment facilitates data retrieval and comparison.

[0083] 102. Check the matching of the data fingerprint of the curve to be detected with the data fingerprints of all representative data in the pre-built index library to determine the matching degree.

[0084] After determining the data fingerprint of the curve to be detected, it is necessary to search in the pre-built index library. Therefore, the index library needs to be built first.

[0085] like Figure 3 As shown, all the data in the spacecraft environmental test database are preprocessed; in the preprocessed data, the first curve S1 is randomly selected to extract the data fingerprint, which is used as the first representative data of the first data block N1 and saved in the first data block N1; a preset number (N) of curves with the smallest difference in data fingerprint from the first representative data of S1 are selected from the database and saved in the first data block N1 (N value is the number of curves in the data block set artificially).

[0086] Among the remaining curves in the database, the curve with the largest difference in data fingerprint from the first representative data is selected as the second representative data S2; a preset number (N) of curves with the smallest difference in data fingerprint from the second representative data S2 are selected in the database and saved in the second data block N2.

[0087] Repeat the process of selecting the second representative data, that is, reselect a curve from the remaining curves as the new second representative data, until the data fingerprints of all curves in the database are saved by at least one data block, and an index library is obtained. In the process of repeatedly selecting the second representative data, two principles are followed: i) the data fingerprints in different fast index libraries are not exclusive, that is, the same curve may be saved in different data blocks at the same time; ii) the representative data of the next data block is found from the database that is not contained in all existing data blocks. The second representative data is continuously updated during the cycle and is not the only data.

[0088] After the index library is built, the curve data matching the curve data fingerprint to be detected is searched in the index library. First, the representative data is searched. Since the index library is composed of multiple data blocks, each data block includes a representative data, the data fingerprint of the data to be detected can be compared with each representative data first, and then the data block with the highest matching to the data to be detected can be determined.

[0089] 103. Determine representative data that has the highest matching performance with the curve to be detected, and read in a data block associated with the representative data that has the highest matching performance.

[0090] After determining the representative data with the highest matching degree, since the differences between different representative data are relatively large, the N curve data included in the data block corresponding to the representative data with the highest matching degree has the highest matching degree with the curve to be detected. i Read in.

[0091] 104. Determine the matching of all the curve data fingerprints in the associated data block with the data fingerprint of the curve to be detected.

[0092] Get the corresponding data block N i After that, at this time, it is necessary to compare the data fingerprint of each curve data in the data block with the data fingerprint of the curve to be detected one by one, so as to i The N curve data in are compared to obtain N comparison results, that is, the matching data of N data fingerprints are obtained.

[0093] 105. Based on the matching degree, all the curve data fingerprints in the associated index library are sorted.

[0094] Get data block N i After comparing the matching degree between the N curve data in the image and the curve to be detected, they are sorted in descending order of matching degree.

[0095] 106. Output the sorting result to the display terminal.

[0096] The final results after sorting are output to a display terminal, for example, to a software display area. By displaying the search and matching results in the order of matching, it is easier for users to fully understand more similarities with the curve to be detected.

[0097] Therein, the process of matching the data to be detected with the corresponding curve data in the data block, or with the representative data, can be performed in two ways, one is a correlation coefficient matching algorithm, and the other is a dynamic time planning matching algorithm.

[0098] The correlation coefficient matching algorithm is as follows: Pearson linear correlation coefficient is used as an indicator to evaluate the similarity of curves. The calculation formula for the correlation coefficient of two curves x and y is (2):

[0099]

[0100] In the formula, rho(x, y) represents the similarity, x i represents the i-th variable of array x, Represents the average value of array x, y i represents the i-th variable of array y, Represents the mean of array y.

[0101] The dynamic time planning matching algorithm is specifically: using a time (frequency) regularization function W that meets certain conditions to describe the time correspondence between the curve to be detected and the curve in the index library, and solving the regularization function corresponding to the minimum cumulative distance when two straight lines are matched.

[0102] Determine the sequence length n of the data fingerprint of the curve to be detected Q and the sequence length m of the data fingerprint representing the data C, as expressed in (3) and (4):

[0103] Q=q1,q2,q3,...,q n (3)

[0104] C=c1,c2,c3,...,c m (4)

[0105] In order to align the two sequences, an n×m matrix grid is constructed based on the sequence length, and the Euclidean distance between any two points in the matrix grid is determined as the matching size. That is, the matrix element (i, j) represents q i and c j The distance between two points d(q i , c j ), as shown in formula (5):

[0106] d(q i , c j )=(q i -c j ) 2 (5)

[0107] Among them, the distance d(q i , c j ) is the Euclidean distance. The smaller the distance, the higher the similarity. Each matrix element (i, j) of this matrix represents the point q i and c j Alignment is essentially about finding a path W in the matrix so that the distance between two sequences is minimized when they are aligned along this path.

[0108] The regularization function W corresponding to the Euclidean distance satisfies the preset constraints, which include: boundary conditions: the starting point and the ending point of the sequence must be directly aligned; continuity: it is impossible to match across a point, and it can only be aligned with its adjacent points. This ensures that every coordinate in Q and C appears in W; monotonicity: the subscripts in Q and C corresponding to W must be monotonically increasing. The minimum distance path W that satisfies the above constraints is the optimal path.

[0109] By adopting the big data curve retrieval method based on data fingerprint of the present invention, the data fingerprint is matched with data blocks in the index library, and then the data fingerprint is further matched, so as to realize the requirement of searching for specific shape curves in massive data, greatly improve the data retrieval efficiency, and have strong practicality.

[0110] Based on the same general inventive concept, the present invention also protects a big data curve retrieval device based on curve fingerprint. The big data curve retrieval device based on curve fingerprint described below and the big data curve retrieval method based on curve fingerprint described above can refer to each other.

[0111] Figure 4 It is a structural diagram of a big data curve retrieval device based on curve fingerprint provided in this embodiment.

[0112] like Figure 4 As shown, this embodiment provides a big data curve retrieval device based on curve fingerprint, including:

[0113] A first determination module 401 is used to determine the data fingerprint of the curve to be detected;

[0114] The first matching module 402 is used to check the matching of the data fingerprint of the curve to be detected with the data fingerprints of all representative data in the pre-built index library to determine the matching degree;

[0115] The second determination module 403 is used to determine the representative data with the highest matching degree with the curve to be detected, and read in the data block associated with the representative data with the highest matching degree;

[0116] A second matching module 404 is used to determine the matching of all the curve data fingerprints in the associated data block with the data fingerprint of the curve to be detected;

[0117] A sorting module 405 is used to sort all the curve data fingerprints in the associated index library based on the matching degree;

[0118] The output module 406 is used to output the sorting result to a display terminal.

[0119] Figure 5 It is a schematic diagram of the structure of the electronic device provided in this embodiment.

[0120] like Figure 5As shown, the electronic device may include: a processor 510, a communication interface 520, a memory 530 and a communication bus 540, wherein the processor 510, the communication interface 520 and the memory 530 communicate with each other through the communication bus 540. The processor 510 may call the logic instructions in the memory 530 to execute a large data curve retrieval method based on curve fingerprints, the method comprising: determining the data fingerprint of the curve to be detected; performing a matching check on the data fingerprint of the curve to be detected and the data fingerprints of all representative data in the pre-built index library to determine the matching degree; determining the representative data with the highest matching degree with the curve to be detected, and reading the data block associated with the representative data with the highest matching degree; determining the matching degree of all curve data fingerprints in the associated data block with the data fingerprint of the curve to be detected; sorting all curve data fingerprints in the associated index library based on the matching degree; and outputting the sorting result to the display terminal.

[0121] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0122] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the big data curve retrieval method based on curve fingerprints provided by the above methods, and the method includes: determining the data fingerprint of the curve to be detected; performing a matching check on the data fingerprint of the curve to be detected with the data fingerprints of all representative data in a pre-built index library to determine the degree of matching; determining the representative data with the highest matching degree with the curve to be detected, and reading in the data block associated with the representative data with the highest matching degree; determining the matching degree of all curve data fingerprints in the associated data block with the data fingerprint of the curve to be detected; based on the matching degree, sorting all curve data fingerprints in the associated index library; and outputting the sorting result to a display terminal.

[0123] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the big data curve retrieval method based on curve fingerprints provided by the above-mentioned methods, the method comprising: determining the data fingerprint of the curve to be detected; performing a matching check on the data fingerprint of the curve to be detected with the data fingerprints of all representative data in a pre-built index library to determine the degree of matching; determining the representative data with the highest matching degree with the curve to be detected, and reading in the data block associated with the representative data with the highest matching degree; determining the matching degree of all curve data fingerprints in the associated data block with the data fingerprint of the curve to be detected; sorting all curve data fingerprints in the associated index library based on the matching degree; and outputting the sorting result to a display terminal.

[0124] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0125] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0126] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A big data curve retrieval method based on curve fingerprint, characterized in that: include: Determine the data fingerprint of the curve to be detected; Performing a matching check on the data fingerprint of the curve to be detected and the data fingerprints of all representative data in the pre-built index library to determine the matching degree; Determine the representative data with the highest matching degree with the curve to be detected, and read in the data block associated with the representative data with the highest matching degree; Determining the matching of all the curve data fingerprints in the associated data block with the data fingerprint of the curve to be detected; Based on the matching degree, all the curve data fingerprints in the associated index library are sorted; The sorting result is output to a display terminal.

2. The big data curve retrieval method based on curve fingerprint according to claim 1 is characterized in that: Before determining the data fingerprint of the curve to be detected, the method further includes: Preprocess the data in all spacecraft environmental test databases; In the preprocessed data, a first curve is randomly selected to extract a data fingerprint as first representative data, and is stored in a first data block; Selecting a preset number of curves having the smallest difference from the data fingerprint of the first representative data from the database, and saving them in the first data block; Among the remaining curves in the database, select the curve having the largest difference from the data fingerprint of the first representative data as the second representative data; Selecting a preset number of curves having the smallest difference from the data fingerprint of the second representative data from the database, and saving them in a second data block; Until the data fingerprints of all the curves in the database are stored by at least one data block, an index library is obtained.

3. The big data curve retrieval method based on curve fingerprint according to claim 1 is characterized in that: The step of determining the data fingerprint of the curve to be detected includes: Determine the data type of the curve to be detected, wherein the data type includes static data and dynamic data; Based on the feature point fitting method of Douglas-Peucker algorithm, the static data is processed to obtain the corresponding data fingerprint; Based on the secondary feature extraction method after frequency domain transformation, the dynamic data is processed to obtain the corresponding data fingerprint.

4. The big data curve retrieval method based on curve fingerprint according to claim 3 is characterized in that: The feature point fitting method based on the Douglas-Peucker algorithm processes the static data to obtain the corresponding data fingerprint, including: Determine the original line elements of the static data; Connect the starting point and the end point of the original line element to obtain a fitted straight line; Determine the node farthest from the fitting straight line, and if the farthest distance is less than the tolerance, retain the node farthest from the fitting straight line; The node with the longest distance is used as the starting point to obtain a new fitting straight line, until an expression with accuracy satisfying the tolerance limit is obtained as the data fingerprint.

5. The big data curve retrieval method based on curve fingerprint according to claim 3 is characterized in that: The method of extracting the dynamic data based on the secondary feature after frequency domain transformation to obtain the corresponding data fingerprint includes: Determine the data fingerprint length; Dividing the dynamic data into equal segments corresponding to the length of the data fingerprint in the logarithmic domain; The maximum value of the dynamic data in each of the equally divided segments is determined as the amplitude of the equally divided segments to form a data fingerprint of the dynamic data.

6. The big data curve retrieval method based on curve fingerprint according to claim 1 is characterized in that: Before determining the data fingerprint of the curve to be detected, the method further includes: Perform amplitude normalization processing on the curve to be detected to obtain normalized data; The normalized data is aligned by interpolating the logarithmic spacing to achieve data preprocessing.

7. The big data curve retrieval method based on curve fingerprint according to claim 6 is characterized in that: The amplitude normalization processing is performed on the curve to be detected to obtain normalized data, including: Determine the acceleration response spectrum of the curve to be tested and the mechanical test input reference spectrum; Inputting the acceleration response spectrum and the mechanical test input reference spectrum into a preset formula to obtain normalized data; The preset formula is: Where Y(ω) is the acceleration response spectrum, G(ω) is the normalized data, and U(ω) is the mechanical test input reference spectrum.

8. The big data curve retrieval method based on curve fingerprint according to claim 1 is characterized in that: The matching check and determining the matching degree include: Determine the sequence length of the data fingerprint of the curve to be detected and the sequence length of the data fingerprint representing the data; Based on the sequence length, a matrix grid is constructed, and the Euclidean distance between any two points in the matrix grid is determined as the matching size.

9. The big data curve retrieval method based on curve fingerprint according to claim 8 is characterized in that: Also includes: The regularization function corresponding to the Euclidean distance satisfies the preset constraints; The preset constraint conditions include: boundary conditions, continuity and monotonicity.

10. A big data curve retrieval device based on curve fingerprint, characterized in that: include: A first determination module, used to determine the data fingerprint of the curve to be detected; A first matching module is used to check the matching of the data fingerprint of the curve to be detected with the data fingerprints of all representative data in the pre-built index library to determine the matching degree; A second determination module is used to determine the representative data with the highest matching degree with the curve to be detected, and read in the data block associated with the representative data with the highest matching degree; A second matching module, used to determine the matching of all the curve data fingerprints in the associated data block with the data fingerprint of the curve to be detected; A sorting module, used to sort all the curve data fingerprints in the associated index library based on the matching degree; The output module is used to output the sorting result to a display terminal.

Images