GNSS coordinate displacement detection method and system combined with long and short time window segmentation

Abstract

The invention relates to the technical field of displacement detection, and particularly provides a GNSS coordinate displacement detection method and system combined with long and short time window segmentation, and the method comprises the steps: collecting original GNSS coordinate data, and carrying out the sampling of the original GNSS coordinate data based on a preset time interval, and obtaining a continuous time sequence; dividing the time sequence based on a long-time detection window length and a short-time detection window length to obtain a long-time window and a short-time window; segmenting the long-time window and the short-time window, performing multiple hypothesis testing based on segmentation points, and respectively determining possible displacement points of the short-time window and the long-time window according to a testing result and a preset threshold value; and performing displacement detection on the time window where the possible displacement point exists based on Bayesian reasoning to obtain an accurate displacement point. Compared with the prior art, the displacement of different feature types can be detected more comprehensively and accurately.

Description

Technical Field

[0001] The present invention relates to the field of displacement detection technology, and in particular to a GNSS coordinate displacement detection method and system combining long and short time window segmentation. Background Art

[0002] Displacement measurement is an important technology for monitoring the spatial position change of objects or ground points. It has a wide range of applications, including building deformation, earthquake crustal movement, landslides and ground subsidence. The Global Navigation Satellite System (GNSS) is a key technical means for displacement measurement. It accurately calculates the position and motion trajectory of ground receivers through satellite signals. The advantage of GNSS lies in its global coverage, all-weather, and long-term continuous monitoring capabilities, which is particularly suitable for large-scale, long-term displacement monitoring. Through precise GNSS positioning technology, millimeter-level or even sub-millimeter-level displacement accuracy can be obtained, which is widely used in earthquake monitoring, bridge and tunnel monitoring, dam safety monitoring and other fields, and has become an indispensable tool in displacement measurement.

[0003] At present, the existing displacement detection based on time series segmentation usually only targets displacements with obvious and stable mutations in a short period of time, and may ignore displacements that change slowly in a short period of time or that do not change obviously due to the influence of GNSS observation conditions, resulting in missed detections. The use of short-time window detection may also be affected by short-time abnormal jumps in the time series, resulting in false detections. In this case, the review based on the statistics of the number of possible change points is also difficult to achieve the purpose of verification and screening. Summary of the invention

[0004] In order to overcome the defects of the prior art described above that only short-term detection is performed, resulting in missed detection or false detection, the present invention provides a GNSS coordinate displacement detection method and system combining long and short time window segmentation.

[0005] In order to achieve the above technical effects, the technical solution of the present invention is as follows:

[0006] A GNSS coordinate displacement detection method combining long and short time window segmentation includes the following steps:

[0007] Collecting original GNSS coordinate data, and sampling the original GNSS coordinate data based on a preset time interval to obtain a continuous time series;

[0008] A long time window and a short time window are set, and the time series is divided based on the long time window and the short time window respectively, and the time series data in any window is segmented, and multiple hypothesis tests are performed based on the segmentation points, and the possible displacement points in the window are determined according to the test results and the preset threshold.

[0009] Based on Bayesian reasoning, displacement detection is performed on the time window where possible displacement points exist to obtain accurate displacement points;

[0010] The displacement estimation value is obtained by calculation based on the possible displacement points and the accurate displacement points.

[0011] The present invention also proposes a GNSS coordinate displacement detection system combined with long and short time window segmentation, the system comprising:

[0012] Data acquisition module: used to collect original GNSS coordinate data, and sample the original GNSS coordinate data based on a preset time interval to obtain a continuous time series;

[0013] Displacement point inference module: used to divide the time series based on the long time window and the short time window respectively, segment the time series data in any window, perform multiple hypothesis tests based on the segmentation points, and determine the possible displacement points in the window according to the test results and the preset threshold;

[0014] Displacement point review module: used to detect the time window where possible displacement points exist based on Bayesian reasoning to obtain accurate displacement points;

[0015] Displacement value calculation module: used to calculate the displacement estimation value based on the possible displacement points and the accurate displacement points.

[0016] The present invention also proposes 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 computer program, the GNSS coordinate displacement detection method combining long and short time window segmentation as described in the present invention is implemented.

[0017] The present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the GNSS coordinate displacement detection method combining long and short time window segmentation as described in the present invention is implemented.

[0018] Compared with the prior art, the present invention has the following beneficial effects:

[0019] The present invention introduces a long-term detection window into the displacement detection based on time series segmentation, and combines the long-term detection window with the short-term window segmentation, so as to more comprehensively and effectively detect the displacements of different characteristic types, and realize the accurate detection of displacements while ensuring the original short-term mutation displacement detection; in addition, the displacement detection method based on Bayesian reasoning is used as a review instead of the statistics of the number of possible change points, which improves the accuracy of displacement detection to a certain extent and avoids misjudgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 The flowchart is a GNSS coordinate displacement detection method combining long and short time window segmentation.

[0021] Figure 2 This is an architecture diagram of a GNSS coordinate displacement detection system that combines long and short time window segmentation. DETAILED DESCRIPTION

[0022] The accompanying drawings are only used for illustrative purposes and are not to be construed as limiting the present invention;

[0023] It is understandable to those skilled in the art that some well-known descriptions may be omitted in the drawings.

[0024] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0025] Example 1

[0026] This embodiment proposes a GNSS coordinate displacement detection method combining long and short time window segmentation, such as Figure 1 , which is a flow chart of the GNSS coordinate displacement detection method combining long and short time window segmentation according to the present embodiment.

[0027] The GNSS coordinate displacement detection method combined with long and short time window segmentation proposed in this embodiment includes the following steps:

[0028] Collecting original GNSS coordinate data, and sampling the original GNSS coordinate data based on a preset time interval to obtain a continuous time series;

[0029] Setting a long time window and a short time window, dividing the time series based on the long time window and the short time window respectively, segmenting the time series data in any window, performing multiple hypothesis tests based on the segmentation points, and determining possible displacement points in the window according to the test results and a preset threshold;

[0030] Based on Bayesian reasoning, displacement detection is performed on the time window where possible displacement points exist to obtain accurate displacement points;

[0031] The displacement estimation value is obtained by calculation based on the possible displacement points and the accurate displacement points.

[0032] In this embodiment, by setting long and short time windows, detection is performed separately in the long and short time windows, and displacement points are found independently. By introducing the synergistic mechanism of long and short time detection windows, the limitations of the traditional single time scale are broken, and comprehensive and accurate detection of displacements of different characteristic types is achieved; specifically, the long time window can capture the trend of slow and continuous displacement changes, for example, the long time window monitors slow displacement every 24 hours to monitor the possible impact of the rainy season on the stability of the mountain; the short time window retains the sensitivity to sudden displacement, for example, the short time window monitors sudden displacement every 5 minutes to deal with sudden displacement caused by heavy rain; the two work together to ensure the accuracy of the original short-term sudden displacement detection, and at the same time significantly expand the scope of application of displacement detection. Secondly, the displacement detection method based on Bayesian reasoning is used as a verification method, which effectively improves the accuracy of displacement detection and greatly reduces the risk of misjudgment.

[0033] In an optional embodiment, the step of sampling the original GNSS coordinate data based on a preset sampling interval to obtain a continuous time series further includes: performing data removal on the sampled time series based on a quartile method and a z-score method; the expression is as follows:

[0034]

[0035] Among them, z X is the z-score statistic, if z X If it is greater than the threshold, it is an outlier and will be removed, otherwise it will be retained; X represents the coordinate data value, μ represents the mean of the time series statistics, σ represents the standard deviation of the time series statistics, and k1 is the coefficient term of the z-score threshold; Q1 and Q3 are the lower quartile and upper quartile of the time series, IQR is the interquartile range, and k2 is the threshold coefficient of the quartile method. If X is greater than Q3+k2·IQR or less than Q1-k2·IQR, it is an outlier and will be removed, otherwise it will be retained.

[0036] In this embodiment, in GNSS data, abnormal data may be generated due to environmental changes, ionosphere, troposphere effects, and multipath effects, which will affect subsequent statistical analysis, thereby affecting positioning accuracy and the accuracy of analysis results. By using the quartile method and z-score method to eliminate abnormal values, the interference of errors on the data can be reduced, and the stability and availability of the data can be improved.

[0037] In an optional embodiment, the step of dividing the time series based on the long time window and the short time window respectively further includes: decomposing and reconstructing the original coordinate time series based on Fourier transform of the time series obtained by dividing the long time window; the expression of the Fourier transform is as follows:

[0038]

[0039] Among them, x(j) is the time series, j is the sequence number of the data in the time series; FFT(k) spectrum, k is the corresponding spectrum sequence number; L is the length of the time series.

[0040] In this embodiment, long-time window data may be affected by observation noise. Performing window sequence denoising first is beneficial to obtaining a relatively smooth sequence. The Fourier transform can effectively decompose the periodic components in the signal, remove high-frequency noise, and restore the real trend or periodic changes, which helps to improve the smoothness and stability of long-time window data.

[0041] In an optional embodiment, the step of performing multiple hypothesis tests based on the segmentation points includes: obtaining subsequences based on the segmentation points within the window, and performing variance test, Z test and standard normal homogeneity test on each segmentation point in turn; and considering segmentation points that satisfy the variance test, Z test and standard normal homogeneity test as possible displacement points, and the expression thereof is as follows:

[0042] VT i =std1 i 2 +std2 i 2 <VT0

[0043]

[0044] Among them, VT i , ZT i SNHT i They represent the statistics of variance test, Z test and standard normal homogeneity test at the split point i respectively; VT0, ZT0 and SNHT0 represent the preset thresholds of the corresponding test statistics; i represents the data sequence number in the time series window corresponding to the split point, L represents the length of the time series in the window, and L min ≤i≤LL min ;ave1 i and ave2 i Respectively represent the mean of the subsequences divided at point i; std1 i and std2 i Respectively represent the standard deviation of the subsequences divided at point i; x j Represents the jth time series data; if the above expression is satisfied, the current segmentation point i is considered to be a possible displacement point.

[0045] In this embodiment, by performing segmentation on long time windows and short time windows, and performing multiple hypothesis tests based on the segmentation points, the accuracy and reliability of displacement point detection are significantly improved. Specifically, by performing variance test, Z test and standard normal homogeneity test on each segmentation point in turn, it is ensured that only points that meet all statistical conditions are determined as possible displacement points, thereby effectively filtering out false displacements caused by noise or abnormal fluctuations. The variance test can identify significant volatility changes in the data, the Z test further confirms whether these fluctuations exceed the normal statistical range, and the standard normal homogeneity test ensures the consistency of data distribution and avoids misjudgment due to non-normal distribution. The multiple hypothesis testing method can accurately identify possible displacement points in complex GNSS data, thereby improving the robustness and accuracy of the displacement detection system.

[0046] In an optional embodiment, the step of performing displacement detection on the time window where the possible displacement points exist based on Bayesian reasoning includes:

[0047] Assuming that there is a displacement point in the time window where the possible displacement point is located, construct a priori probability distribution;

[0048] Based on the time series data, performing a Bayesian update on the prior probability distribution to obtain a posterior probability distribution;

[0049] Based on the posterior probability distribution, the Markov Chain Monte Carlo method is used to perform sampling based on conditional probability to obtain the parameters of the hypothesized displacement point, and the hypothesized displacement point is re-submitted to the multiple hypothesis test. If the test passes, the hypothesized displacement point is used as the accurate displacement point, otherwise it is considered that there is no displacement point.

[0050] In this embodiment, the possible displacement points are rechecked and detected based on Bayesian reasoning. Specifically, the Bayesian update process combines the actual observation information of the time series, and further improves the accuracy of displacement point identification through the posterior probability distribution; secondly, the conditional probability sampling is combined with the Markov chain Monte Carlo method to ensure that the assumptions of the displacement points conform to statistical laws. Finally, multiple hypothesis tests are re-performed to ensure that the displacement points that meet the test conditions are output as accurate displacement points, reducing the risk of misjudgment.

[0051] In an optional embodiment, the step of constructing a priori probability distribution for a time window where possible displacement points exist assuming that there are displacement points comprises:

[0052] Assuming that the time series data in the time window follows a normal distribution and there is only one displacement point, the displacement observation likelihood function is expressed as:

[0053]

[0054] Among them, x represents time series data, xi represents the i-th time series data in the time window; θ = (μ0, μ1, τ, σ) represents the assumed displacement point; τ represents the displacement time of the assumed displacement point; μ0 and μ1 represent the means of the two subsequences after the time window is divided at time τ; σ represents the standard deviation of the time series data following the normal distribution;

[0055] Assuming that the parameters within θ are independent of each other, the prior distribution of the displacement points is assumed to be:

[0056]

[0057] Among them, P(θ) is the prior probability of the assumed displacement point; P(μ0) and P(μ1) represent the prior probabilities of the mean μ0 and μ1 in the assumed displacement point, respectively; P(τ) represents the prior probability of the displacement point at the displacement time τ; P(σ) represents the prior probability of the standard deviation σ of the assumed displacement point sequence; U(μ min ,μ max ) indicates that the mean follows a continuous uniform distribution, μ min and μ max Indicates the upper and lower limits of uniform distribution; DiscreteU(t1,t L ) indicates that the displacement time follows a discrete uniform distribution, t1 and t L Indicates the upper and lower limits of the displacement moment; N(μ σ ,σ σ ) indicates that the standard deviation follows a Gaussian normal distribution, μ σ and σ σ represents the mean and standard deviation of the standard deviation parameter.

[0058] In the embodiment itself, the changing characteristics of the data are captured by probability modeling based on normal distribution, thereby realizing the description of the displacement point position and distribution characteristics; secondly, the calculation complexity is simplified by assuming independent parameters, thereby improving the prediction efficiency and adaptability of the model; in addition, through the integration of prior information, the robustness of the model to uncertainty and noise is enhanced, while ensuring the rationality and interpretability of the prediction results.

[0059] In an optional embodiment, the method further includes: calculating the displacement amount based on the possible displacement point and the accurate displacement point, and the expression thereof is as follows:

[0060]

[0061] Among them, the subscript τ1 represents the possible displacement point, the subscript τ2 represents the accurate displacement point, and Δx represents the displacement amount.

[0062] In this embodiment, accurate displacement can be obtained through possible displacement points and accurate displacement points. Combining the two displacement points can correct errors, optimize result accuracy, reduce the impact of noise on displacement calculation, and improve the robustness of the model.

[0063] Example 2

[0064] This embodiment proposes a GNSS coordinate displacement detection system combining long and short time window segmentation, and applies the GNSS coordinate displacement detection method combining long and short time window segmentation proposed in Embodiment 1. Figure 2 , which is an architecture diagram of a GNSS coordinate displacement detection system combining long and short time window segmentation according to the present embodiment.

[0065] This embodiment proposes a GNSS coordinate displacement detection system combining long and short time window segmentation, including:

[0066] Data acquisition module: used to collect original GNSS coordinate data, and sample the original GNSS coordinate data based on a preset time interval to obtain a continuous time series;

[0067] Displacement point inference module: used to divide the time series based on the long time window and the short time window respectively, segment the time series data in any window, perform multiple hypothesis tests based on the segmentation points, and determine the possible displacement points in the window according to the test results and the preset threshold;

[0068] Displacement point review module: used to detect the time window where possible displacement points exist based on Bayesian reasoning to obtain accurate displacement points;

[0069] Displacement value calculation module: used to calculate the displacement estimation value based on the possible displacement points and the accurate displacement points.

[0070] It can be understood that the system of this embodiment corresponds to the method of the above-mentioned embodiment 1, and the options in the above-mentioned embodiment 1 are also applicable to this embodiment, so they will not be described repeatedly here.

[0071] Example 3

[0072] This embodiment proposes a computer device, including a memory and a processor, wherein the memory stores computer-readable instructions, wherein when the computer-readable instructions are executed by the processor, the processor executes the steps of the GNSS coordinate displacement detection method combined with long and short time window segmentation proposed in Example 1.

[0073] Example 4

[0074] This embodiment proposes a storage medium having computer-readable instructions stored thereon, wherein the computer-readable instructions, when executed by a processor, implement the steps of the GNSS coordinate displacement detection method combining long and short time window segmentation proposed in Embodiment 1.

[0075] Exemplarily, the storage medium includes, but is not limited to, a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0076] Exemplarily, the instructions, programs, code sets or instruction sets may be implemented using conventional programming languages.

[0077] Exemplarily, the processor includes but is not limited to a smart phone, a personal computer, a server, a network device, etc., and is used to execute all or part of the steps of the GNSS coordinate displacement detection method based on staged learning combined with long and short time window segmentation described in Example 1.

[0078] The terms used in the drawings are for illustrative purposes only and should not be construed as limiting this patent;

[0079] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A GNSS coordinate displacement detection method combining long and short time window segmentation, characterized in that: The following steps are involved: Collecting original GNSS coordinate data, and sampling the original GNSS coordinate data based on a preset time interval to obtain a continuous time series; Setting a long time window and a short time window, dividing the time series based on the long time window and the short time window respectively, segmenting the time series data in any window, performing multiple hypothesis tests based on the segmentation points, and determining possible displacement points in the window according to the test results and a preset threshold; Based on Bayesian reasoning, displacement detection is performed on the time window where possible displacement points exist to obtain accurate displacement points; The displacement estimation value is obtained by calculation based on the possible displacement points and the accurate displacement points.

2. The GNSS coordinate displacement detection method combined with long and short time window segmentation according to claim 1 is characterized in that: The step of sampling the original GNSS coordinate data based on a preset time interval to obtain a continuous time series also includes: removing data from the sampled time series based on the quartile method and the z-score method; the expression is as follows: Among them, z X is the z-score statistic, if z X If it is greater than the threshold, it is an outlier and will be removed, otherwise it will be retained; X represents the coordinate data value, μ represents the mean of the time series statistics, σ represents the standard deviation of the time series statistics, and k1 is the coefficient term of the z-score threshold; Q1 and Q3 are the lower quartile and upper quartile of the time series, IQR is the interquartile range, and k2 is the threshold coefficient of the quartile method. If X is greater than Q3+k2·IQR or less than Q1-k2·IQR, it is an outlier and will be removed, otherwise it will be retained.

3. The GNSS coordinate displacement detection method combined with long and short time window segmentation according to claim 1 is characterized in that: The step of dividing the time series based on the long time window and the short time window respectively further includes: decomposing and reconstructing the original coordinate time series based on Fourier transform of the time series obtained by dividing the long time window; the expression of the Fourier transform is as follows: Among them, x(j) is the time series, j is the sequence number of the data in the time series; FFT(k) spectrum, k is the corresponding spectrum sequence number; L is the length of the time series.

4. The GNSS coordinate displacement detection method combined with long and short time window segmentation according to claim 1 is characterized in that: The step of performing multiple hypothesis tests based on the split points includes: obtaining subsequences based on the split points within the window, and performing variance test, Z test and standard normal homogeneity test on each split point in turn; and considering the split points that satisfy the variance test, Z test and standard normal homogeneity test as possible displacement points, and the expression thereof is as follows: VT i =std1 i 2 +std2 i 2 <VT0 Among them, VT i , ZT i SNHT i They represent the statistics of variance test, Z test and standard normal homogeneity test at the split point i respectively; VT0, ZT0 and SNHT0 represent the preset thresholds of the corresponding test statistics; L represents the length of the time series in the window, i represents the data sequence number of the time series corresponding to the split point, and L min ≤i≤LL min ;ave1 i and ave2 i Respectively represent the mean of the subsequences divided at point i; std1 i and std2 i Respectively represent the standard deviation of the subsequences divided at point i; x j Represents the jth time series data; if the above expression is satisfied, the current segmentation point i is considered to be a possible displacement point.

5. The GNSS coordinate displacement detection method combined with long and short time window segmentation according to claim 4 is characterized in that: The step of calculating the displacement amount based on the possible displacement points and the accurate displacement points includes: calculating the mean of the subsequence based on the accurate displacement points and the possible displacement points, and calculating the displacement estimation value based on the mean, and the expression thereof is as follows: Wherein, subscript τ1 represents the possible displacement point, subscript τ2 represents the accurate displacement point, and Δx represents the displacement estimate.

6. The GNSS coordinate displacement detection method combined with long and short time window segmentation according to claim 1 is characterized in that: The step of performing displacement detection on the time window where the possible displacement points exist based on Bayesian reasoning includes: Assuming that there is a displacement point in the time window where the possible displacement point is located, construct a priori probability distribution; Based on the time series data, performing a Bayesian update on the prior probability distribution to obtain a posterior probability distribution; Based on the posterior probability distribution, the Markov Chain Monte Carlo method is used to perform sampling based on conditional probability to obtain the parameters of the hypothesized displacement point, and the hypothesized displacement point is re-submitted to the multiple hypothesis test. If the test passes, the hypothesized displacement point is used as the accurate displacement point, otherwise it is considered that there is no displacement point.

7. The GNSS coordinate displacement detection method combined with long and short time window segmentation according to claim 6 is characterized in that: The step of constructing a priori probability distribution based on the assumption that a displacement point exists in the time window where the possible displacement point exists includes: Assuming that the time series data in the time window follows a normal distribution and there is only one displacement point, the displacement observation likelihood function is expressed as: Among them, x represents time series data, x i represents the i-th time series data in the time window; θ = (μ0, μ1, τ, σ) represents the assumed displacement point; τ represents the displacement time of the assumed displacement point; μ0 and μ1 represent the means of the two subsequences after the time window is divided at time τ; σ represents the standard deviation of the time series data following the normal distribution; Assuming that the parameters within θ are independent of each other, the prior distribution of the displacement points is assumed to be: Among them, P(θ) is the prior probability of the assumed displacement point; P(μ0) and P(μ1) represent the prior probabilities of the mean μ0 and μ1 in the assumed displacement point, respectively; P(τ) represents the prior probability of the displacement point at the displacement time τ; P(σ) represents the prior probability of the standard deviation σ of the assumed displacement point sequence; U(μ min ,μ max ) indicates that the mean follows a continuous uniform distribution, μ min and μ max Indicates the upper and lower limits of uniform distribution; DiscreteU(t1,t L ) indicates that the displacement time follows a discrete uniform distribution, t1 and t L Indicates the upper and lower limits of the displacement moment; N(μ σ ,σ σ ) indicates that the standard deviation follows a Gaussian normal distribution, μ σ and σ σ represents the mean and standard deviation of the standard deviation parameter.

8. A GNSS coordinate displacement detection system combined with long and short time window segmentation, applied to the GNSS coordinate displacement detection method combined with long and short time window segmentation as claimed in any one of claims 1 to 7, characterized in that: The system comprises: Data acquisition module: used to collect original GNSS coordinate data, and sample the original GNSS coordinate data based on a preset time interval to obtain a continuous time series; Displacement point inference module: used to divide the time series based on the long time window and the short time window respectively, segment the time series data in any window, perform multiple hypothesis tests based on the segmentation points, and determine the possible displacement points in the window according to the test results and the preset threshold; Displacement point review module: used to detect the time window where possible displacement points exist based on Bayesian reasoning to obtain accurate displacement points; Displacement value calculation module: used to calculate the displacement estimation value based on the possible displacement points and the accurate displacement points.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the GNSS coordinate displacement detection method combined with long and short time window segmentation as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the GNSS coordinate displacement detection method combined with long and short time window segmentation as described in any one of claims 1 to 7 is implemented.

Images