JP2019128173A - Data processing device, displacement observation system, data processing method and data processing program - Google Patents

Abstract

To correctly extract a cyclic noise included in time series data.SOLUTION: A data processing device 10 comprises a cyclic noise extraction unit 11 and a cyclic noise smoothing unit 12. The cyclic noise extraction unit 11 extracts a cyclic noise included in each time series data. The cyclic noise smoothing unit 12 smooths each of a plurality of noises separated at a cycle of cyclic noises, which arranged in time series, out of the cyclic noises.SELECTED DRAWING: Figure 1

Description

  The present invention relates to the extraction of periodic noise included in a plurality of time-series analysis data.

  Positioning using a GNSS (Global Navigation Satellite Systems) positioning signal such as GPS (Global Positioning System) has been put into practical use. As a highly accurate positioning method, there is positioning (interference positioning) using the carrier wave phase of the positioning signal.

  For example, Patent Document 1 describes static positioning, which is a type of positioning using a carrier phase. Patent Document 1 is a system for observing the displacement of a slope, and performs smoothing processing on time-series analysis data (for example, displacement data) obtained by static positioning.

JP 2004-144623 A

  However, the analysis data may include periodic noise. Periodic noise is noise that fluctuates at a predetermined period. For example, when the above-mentioned positioning signal is used, periodic noise is noise due to multipath or diffraction.

  Further, such periodic noise further includes another noise such as an observation error, and it is not easy to accurately extract the periodic noise.

  Therefore, an object of the present invention is to provide a technique for accurately extracting periodic noise included in time-series data.

  A data processing apparatus according to the present invention includes a periodic noise extraction unit and a periodic noise smoothing unit. The periodic noise extraction unit extracts periodic noise included in each of the time-series data. The periodic noise smoothing unit performs smoothing processing for each of a plurality of periodic noises that are separated by the periodic noise period among the periodic noises arranged in time series.

  In this configuration, since smoothing is performed using only a plurality of data separated by the period of the periodic noise, only high frequency noise is removed while the form of the periodic noise is maintained.

  According to the present invention, it is possible to accurately extract periodic noise included in time-series data.

It is a block diagram of a data processor concerning an embodiment of the present invention. It is a block diagram which shows the 1st aspect of the period noise smoothing part which concerns on embodiment of this invention. It is a figure for demonstrating the concept of the process of the periodic noise smoothing part which concerns on embodiment of this invention. It is a block diagram of another aspect of the data processor which concerns on embodiment of this invention. It is a block diagram showing composition of a periodic noise suppression part concerning an embodiment of the present invention. It is a block diagram which shows the 2nd aspect of the period noise smoothing part which concerns on embodiment of this invention. It is a flowchart which shows the main processes of the data processing method which concerns on embodiment of this invention. It is a flowchart which shows the 1st aspect of the smoothing method of periodic noise. It is a flowchart which shows the 2nd aspect of the smoothing method of periodic noise. It is a block diagram which shows the 3rd aspect of the period noise smoothing part which concerns on embodiment of this invention. It is a functional block diagram of a displacement observation system including a data processing device according to the present embodiment.

  A data processing apparatus, a data processing method, and a data processing program according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram of a data processing apparatus according to an embodiment of the present invention.

  As shown in FIG. 1, the data processing device 10 includes a periodic noise extraction unit 11 and a periodic noise smoothing unit 12. The periodic noise extraction unit 11 and the periodic noise smoothing unit 12 are realized by, for example, hardware such as an IC and a program executed in the hardware.

  The time-series analysis data D is input to the periodic noise extraction unit 11. The time-series analysis data D is sequentially obtained based on the observation period. This time series analysis data D corresponds to "time series data" of the present invention.

  The time-series analysis data D is, for example, a displacement amount of a baseline vector calculated from the carrier wave phase (corresponding to the observation value) of the positioning signal. In this case, the period of the time-series analysis data D is based on the time of the positioning system included in the positioning signal.

  The time-series analysis data D is not limited to the displacement amount of the baseline vector, but may be a value calculated from observation values sequentially acquired at predetermined time intervals. Furthermore, observation values themselves may be used as analysis data.

  The periodic noise extraction unit 11 is realized by a filter in which the frequency band of periodic noise is in the pass band and the frequency band of displacement is in the attenuation band. For example, the periodic noise extraction unit 11 is realized by a high pass filter (HPF). Specifically, assuming that periodic noise is noise due to multipath or diffraction, the periodic noise extraction unit 11 passes frequency components having a period of one stellar day and attenuates frequency components of a period longer than one stellar day Realized by a high-pass filter.

  The periodic noise extraction unit 11 filters the time-series analysis data D to generate a time-series periodic noise Dn. The periodic noise extraction unit 11 outputs time-series periodic noise Dn to the periodic noise smoothing unit 12. This periodic noise Dn includes noise having a shorter period (higher frequency) than the period of Dn of the periodic noise.

  The periodic noise smoothing unit 12 is realized by the following configuration. FIG. 2 is a block diagram showing a first aspect of the periodic noise smoothing unit according to the embodiment of the present invention.

  As shown in FIG. 2, the periodic noise smoothing unit 12 includes a memory 121, a setting unit 122, and a low pass filter (LPF) 123.

  The periodic noise Dn is sequentially stored in the memory 121.

  The setting unit 122 sets a plurality of periodic noises Dn to be output to the low pass filter 123 among the time-series periodic noises Dn stored in the memory 121. The setting unit 122 includes, for example, an elapsed time counter, a remainder calculation unit, and an output setting unit.

  The elapsed time counter counts the elapsed time from the activation start time of the data processing device 10. This count is based on the time of the positioning system when the above positioning signal is used. The remainder calculation unit performs remainder calculation based on the period for the count value. The output setting unit determines a plurality of periodic noises Dn to be output out of the time-series periodic noises Dn stored in the memory 121, using the result of the remainder calculation.

  Specifically, the setting unit 122 executes the following process. FIG. 3 is a diagram for explaining the concept of processing of the periodic noise smoothing unit according to the embodiment of the present invention.

  The multi-pass and diffraction of the fixed point, for example, are unique when fixed to a slope or the like and measured over a long period of time, and depend on topography and structures. And since it depends on the topography, it occurs repeatedly with a one-star daily cycle.

  When the latest periodic noise Dn (n) is stored in the memory 121, the setting unit 122 corresponds to a predetermined period K (for example, one stellar day) in the past starting from the latest periodic noise Dn (n). A plurality of periodic noises Dn that are separated every period K) are set. That is, as illustrated in FIGS. 2 and 3, the setting unit 122 outputs periodic noise Dn (n), periodic noise Dn (n−K), and periodic noise Dn (n−2K) as the plurality of periodic noises Dn to be output. , And periodic noise Dn (n-3K), ..., and periodic noise Dn (n-mK) are set. Here, m is a positive integer. The number of the plurality of periodic noises Dn to be output can be set as appropriate. That is, for example, in the case of FIG. 2, the number of the plurality of periodic noises Dn to be output is five or more m + 1, but it may be plural.

  Among the time-series periodic noises Dn stored in the memory 121, a plurality of periodic noises Dn whose output is set by the setting unit 122 are input to the low-pass filter 123. The setting unit 122 executes the setting process of the plurality of periodic noises Dn output from the memory 121 every time the latest periodic noise Dn (n) stored in the memory 121 is updated.

  The low pass filter 123 performs filter processing (smoothing processing) on a plurality of input periodic noises Dn, and outputs smoothed periodic noise Dns. The low pass filter 123 passes frequency components of the periodic noise Dn and attenuates frequency components of a cycle (high frequency) shorter than the periodic noise Dn.

  The low-pass filter 123 performs a filter process every time a plurality of input periodic noises Dn are updated. Thereby, the periodic noise Dns after smoothing consisting of the same time series as the analysis data D can be obtained.

  By using such a configuration and processing, as shown in FIG. 3, periodic noise Dns after smoothing (see thick solid line in FIG. 3) is included in periodic noise Dn (see thin solid line in FIG. 3) Higher frequency noise is suppressed.

  Thereby, the time-series periodic noise Dn included in the time-series analysis data D can be accurately extracted.

  That is, the extracted periodic noise includes random noise. When a low-pass filter is applied to this periodic noise in time-sequential order according to a general method, data near an arbitrary time is used, so that data near the time is smoothed. Therefore, as the degree of smoothing is increased, the extracted periodic noise is gradually flattened, and the periodic noise removal effect is reduced.

  However, by using the configuration and processing of the present embodiment, instead of using data around an arbitrary time, smoothing is performed using only a plurality of data separated by a period of periodic noise. Is maintained, only high frequency noise is removed. Therefore, even if the smoothing period is long, the periodic noise is not flattened, and the shape of the periodic noise can be made more accurate. This makes it possible to increase the periodic noise removal effect in the subsequent processing.

  The periodic noise Dn thus extracted is used as follows. FIG. 4 is a block diagram of another aspect of a data processing apparatus according to an embodiment of the present invention.

  The data processing device 10A shown in FIG. 4 is different from the data processing device 10 shown in FIG. 1 in that a periodic noise suppression unit 13 is provided. The other configuration of the data processing device 10A is the same as that of the data processing device 10, and the description of the same parts is omitted.

  The periodic noise suppression unit 13 receives time-series analysis data D and the periodic noise Dns after smoothing.

  FIG. 5 is a block diagram showing the configuration of a periodic noise suppression unit according to an embodiment of the present invention. As shown in FIG. 5, the periodic noise suppression unit 13 includes a subtractor 131 and a smoothing filter 132. The smoothing filter 132 can be omitted.

  The subtracter 131 subtracts the smoothed periodic noise Dns from the analysis data D. Thereby, the periodic noise contained in the analysis data D is suppressed.

  The smoothing filter 132 smoothes the analysis data after subtraction and outputs analysis data Do after suppression of periodic noise. For example, the smoothing filter 132 is realized by a low-pass filter, for example. Thereby, analysis data in which further noise is suppressed can be obtained. For example, as indicated by the broken line in FIG. 3, analysis data having a displacement amount of substantially 0 can be obtained for an observation target having a displacement amount of 0. This makes it possible to obtain observation data more appropriately indicating the actual state of the observation target.

  The periodic noise smoothing unit described above can also be realized by the configuration shown below. FIG. 6 is a block diagram showing a second mode of the periodic noise smoothing unit according to the embodiment of the present invention.

  As shown in FIG. 6, the periodic noise smoothing unit 12A includes a setting unit 122, a plurality of low pass filters 124, a switch 125, and a switch 126.

  The setting unit 122 has a configuration similar to that of the setting unit 122 shown in FIG. 2 described above, and executes control of the switch 125 and the switch 126 according to the result of the remainder calculation. The switch 125 and the switch 126 are controlled to select the same low pass filter 124 among the plurality of low pass filters 124.

  The plurality of low pass filters 124 have the same configuration, and are realized, for example, by a general filter circuit such as a first-order IIR filter.

  With such a configuration, similar to the above-described configuration and processing, smoothing is performed using only a plurality of pieces of data that are separated by the period of each periodic noise. Therefore, high-frequency noise is maintained while maintaining the form of periodic noise. Only removed. Therefore, even if the smoothing period is extended, the periodic noise is not flattened and the form of the periodic noise can be made more accurate. This makes it possible to increase the periodic noise removal effect in the subsequent processing.

  Also, the plurality of low pass filters 124 are prepared by the number that divides the cycle K of the periodic noise Dns at predetermined intervals. The switch 125 and the switch 126 are switched by the setting unit 122 at each switching timing according to the number of division of the cycle K at predetermined intervals. Thereby, the periodic noise Dns after the smoothing which consists of a time series is obtained.

  As described above, even if the configuration of the periodic noise smoothing unit 12A is used, it is possible to accurately extract the periodic noise Dns after the smoothing, which is the time series, from the analysis data D that is the time series.

  In the above description, a mode in which the data processing apparatus is realized by a plurality of functional units has been described. However, the information processing apparatus such as a personal computer, a program executed by the information processing apparatus, a storage medium that stores the program, and the like It can also be realized. In this case, the data processing apparatus may execute the following data processing method. FIG. 7 is a flowchart showing main processing of the data processing method according to the embodiment of the present invention. In addition, the specific process of each process shown in FIG. 7 is mentioned above, and description is abbreviate | omitted here.

  As shown in FIG. 7, the data processing apparatus extracts time-series periodic noise Dn from time-series analysis data D (S11). The data processor smoothes the time-series periodic noise Dn (S12). The data processing apparatus subtracts periodic noise Dns after time series smoothing from time series analysis data D, and suppresses periodic noise included in the analysis data D (S13). The data processing device smoothes the analysis data D after the suppression of the periodic noise (S14).

  FIG. 8 is a flowchart showing a first aspect of the periodic noise smoothing method. In addition, the specific process of each process shown in FIG. 8 is mentioned above, and description is abbreviate | omitted here.

  As shown in FIG. 8, the data processing apparatus sequentially stores the periodic noise Dn (S21). The data processing apparatus selects the periodic noise Dn from the stored periodic noise Dn in time series at a period K (for example, a period of one star day unit) (S22). The data processor smoothes the selected periodic noise Dn (S23). These processes are performed every time the latest periodic noise Dn is input.

  FIG. 9 is a flowchart showing a second aspect of the periodic noise smoothing method. In addition, the specific process of each process shown in FIG. 9 is mentioned above, and description is abbreviate | omitted here.

  As shown in FIG. 9, the data processing apparatus receives an input of periodic noise (S31). The data processing device selects a filter corresponding to the time (S32). The data processing device smoothes the periodic noise Dn using the selected filter (S33). These processes are performed while switching the filter each time the latest periodic noise Dn is input.

  In the above description, the periodic noise smoothing unit 12 selects and filters the periodic noise Dn separated by the period K. However, when the periodic noise Dn is missing at the time position of the period K, the periodic noise smoothing unit uses the periodic noise Dn before and after the time axis to filter the missing periodic noise Dn. Processing may be performed. FIG. 10 is a block diagram showing a third aspect of the periodic noise smoothing unit according to the embodiment of the present invention.

  As shown in FIG. 10, the periodic noise smoothing unit 12B includes a memory 121, a setting unit 122, and a low pass filter 123B. The memory 121 of the periodic noise smoothing unit 12B is the same as the memory 121 described above, and a description thereof is omitted.

  When the latest periodic noise Dn (n) is stored in the memory 121, the setting unit 122 corresponds to a predetermined period K (for example, one stellar day) in the past starting from the latest periodic noise Dn (n). A plurality of periodic noises Dn that are separated every period K) are set. In addition, when the periodic noise Dn is missing at the time position of the period K, the setting unit 122 has a plurality of periodic noises on the time axis with respect to the plurality of periodic noises Dn separated for each period K. Set For example, in the case of FIG. 10, the setting unit 122 sets one cycle noise Dn before and after on the time axis with respect to a plurality of cycle noises Dn spaced apart at every cycle K (on the time axis for the latest cycle noise). The previous (past) periodic noise Dn only) is selected.

  That is, in the case of the example in FIG. 10, if the periodic noise Dn (n−2K) is missing from the periodic noise Dn (n) as the plurality of periodic noises Dn to be output, the setting unit 122 Dn (n-2K + 1) and periodic noise Dn (n-2K-1) are set. Then, the setting unit 122 calculates the periodic noise Dn (n-2K) from the average value of the periodic noise Dn (n-2K + 1) and the periodic noise Dn (n-2K-1) or the like.

  Here, with respect to dropped periodic noise, an aspect is shown in which one periodic noise before and after, that is, two periodic noises before and after in the time series are used. However, the number is not limited to two and may be four or more. In addition, when using many pieces, it is good to interpolate by methods different from smoothing by a low pass filter etc., for example, curve fitting etc.

  In addition, any periodic noise before or after the missing periodic noise may be used, and the missing periodic noise may be replaced with a periodic noise before or after an integer multiple of K in the time series.

  Furthermore, here, the case where periodic noise is missing is shown as an example, but when periodic noises of the past are not uniform at the initial stage, for example, an integer multiple of K in time series with respect to periodic noise to be replaced The replacement may be performed using the periodic noise before or after.

  The low-pass filter 123B executes the filter processing (smoothing processing) of the plurality of periodic noises Dn output from the memory 121 by the setting unit 122, and outputs the periodic noise Dns after the smoothing.

  Even with such a configuration and processing, time-series periodic noise included in time-series analysis data can be accurately extracted, and periodic noise included in the analysis data can be accurately suppressed.

  In the above description, the case where the period of the periodic noise is one stellar day is shown as an example. However, the period of the periodic noise may be appropriately set according to the observation system to be applied and the periodic noise included in the time series data.

  Such an analysis data processing apparatus can be applied to a system using static positioning described below. FIG. 11 is a functional block diagram of an observation system including the data processing device according to the present embodiment. As shown in FIG. 11, the observation system 1 includes an observation station 21, an observation station 22, a reference station 30, a line concentrator 40, a server 50, and an analysis device 60. The observation system 1 corresponds to the “displacement observation system” of the present invention.

  In FIG. 11, there are two observation stations and one reference station, but this is not limitative. The number of observation stations and the number of reference stations are appropriately determined according to the location to be observed. Moreover, although the aspect which uses GPS is shown in FIG. 11, the other positioning system of GNSS can also be used.

  The observation station 21, the observation station 22, and the reference station 30 are connected to the line concentrator 40 by wire or wirelessly. The line aggregator 40, the server 50, and the analysis device 60 are connected via a network 500.

  The observation station 21 includes a GPS antenna 211 and a GPS receiver 212. The GPS antenna 211 receives a positioning signal (GPS signal) transmitted from each of the plurality of positioning satellites SAT1, SAT2, SAT3, and SAT4, and outputs it to the GPS receiver 212. The GPS receiver 212 observes carrier phases of positioning signals from a plurality of positioning satellites SAT1, SAT2, SAT3 and SAT4, respectively. Also, the GPS receiver 212 measures its own device position by single positioning or the like using a code phase. The GPS receiver 212 transmits the carrier phase and the positioning result to the line concentrator 40.

  The observation station 22 includes a GPS antenna 221 and a GPS receiver 222. The GPS antenna 221 receives a positioning signal (GPS signal) transmitted from each of the plurality of positioning satellites SAT1, SAT2, SAT3, and SAT4, and outputs it to the GPS receiver 222. The GPS receiver 222 observes carrier phases of positioning signals from a plurality of positioning satellites SAT1, SAT2, SAT3 and SAT4, respectively. Also, the GPS receiver 222 measures its own device position by single positioning or the like using a code phase. The GPS receiver 222 transmits the carrier phase and the positioning result to the line aggregator 40.

  The reference station 30 comprises a GPS antenna 301 and a GPS receiver 302. The GPS antenna 301 receives a positioning signal (GPS signal) transmitted from each of the plurality of positioning satellites SAT1, SAT2, SAT3, and SAT4, and outputs it to the GPS receiver 302. The GPS receiver 302 observes carrier phases of positioning signals from a plurality of positioning satellites SAT1, SAT2, SAT3 and SAT4, respectively. Also, the GPS receiver 302 measures its own device position by single positioning or the like using a code phase. The GPS receiver 302 transmits the carrier phase and the positioning result to the line aggregator 40.

  The line concentrator 40 transmits, to the server 50, the carrier phase and the positioning result from each of the observation station 21, the observation station 22, and the reference station 30. The server 50 is realized by, for example, an FTP server.

  The analysis device 60 includes an arithmetic unit 61 and a storage unit 62. The arithmetic unit 61 has the function of the data processing device 10 described above. In addition, the calculation unit 61 acquires the carrier wave phase and the positioning result from each of the observation station 21, the observation station 22, and the reference station 30 from the server 50, and calculates a baseline vector that connects the stations. The calculation unit 61 sequentially calculates baseline vectors at preset intervals of observation time of baseline vectors. Further, the calculation unit 61 sequentially calculates the displacement amount of the baseline vector. For example, the displacement amount of this baseline vector corresponds to "time-series data" of the present invention.

  For example, in the case of a landslide observation system, this corresponds to the fluctuation amount of the ground at the position where the observation station 21 and the observation station 22 are installed. Such ground fluctuation does not change stably unless there is ground fluctuation due to an earthquake, heavy rainfall, or the like. On the other hand, when ground changes occur, changes occur over a wide range depending on the factors. Therefore, by using the carrier wave phase, such a change can be observed with high accuracy. Further, by using the configuration of the data processing apparatus 10 of the present invention, the influence of periodic noise due to multipath, diffraction, etc. can be effectively prevented. Can be suppressed.

  In the above description, the aspect of using the analysis data by static positioning is shown. However, the above configuration and processing are also applied to analysis data by kinematic positioning such as RTK (real-time kinematic) or precision single positioning (PPP). Can be applied.

  Moreover, although the aspect provided with the line | wire aggregator 40 was shown in the above-mentioned description, this can also be abbreviate | omitted. In this case, the observation stations 21 and 22 and the reference station 30 may output the carrier wave phase and the positioning result to the server 50. Also, the server 50 can be omitted. In this case, the line concentrator 40 or the observation stations 21 and 22 and the reference station 30 may output the carrier wave phase and the positioning result to the analysis device 60. The analysis device 60 stores these in the storage unit 62 and uses them for the above-described processing.

  Further, the carrier phase and the positioning result may be aggregated by any of the observation stations 21 and 22 and the reference station 30, and the displacement amounts of the base line vector and the base line vector may be calculated. In this case, the station that has calculated the displacement amount of the baseline vector may output the displacement amount of the baseline vector to the server 50.

  Furthermore, in the above description, the aspect in which the analysis device 60 is provided separately from the observation stations 21 and 22 and the reference station 30 has been described. However, the analysis device 60 is not limited to any of the observation stations 21 and 22 and the reference station 30. It may be built in.

1: Observation system 10, 10A: Data processing device 11: Periodic noise extraction unit 12, 12A, 12B: Periodic noise smoothing unit 13: Periodic noise suppression unit 21, 22: Observation station 30: Reference station 40: Line aggregator 50: Server 60: Analysis device 61: Calculation unit 62: Storage unit 121: Memory 122: Setting units 123, 123B, 124: Low-pass filter 125, 126: Switch 131: Subtractor 132: Smoothing filters 211, 221 and 301: GPS antennas 212, 222, 302: GPS receiver 500: Network D: Analysis data Dn: Periodic noise Dns: Smoothed periodic noise SAT1, SAT2, SAT3, SAT4: Positioning satellite

Claims (9)

A periodic noise extraction unit that extracts periodic noise included in each of the time series data;
Among the periodic noises arranged in time series, a periodic noise smoothing unit that performs smoothing processing for each of the plurality of periodic noises separated by a period of the periodic noise;
A data processing apparatus. The data processing apparatus according to claim 1, wherein
The periodic noise smoothing unit includes:
A storage unit for storing periodic noises arranged in time series;
A setting unit for setting the cycle;
A smoothing filter for smoothing the plurality of periodic noises separated by the periodicity stored in the storage unit;
A data processing apparatus. The data processing apparatus according to claim 1, wherein
The periodic noise smoothing unit includes:
A plurality of smoothing filters comprising a number based on the period;
A switch for selecting the plurality of smoothing filters;
A setting unit that switches the switch using the period;
A data processing apparatus. A data processing device according to any one of claims 1 to 3,
The periodic noise smoothing unit includes:
When periodic noise of a predetermined time is missing in the plurality of periodic noises separated by the period, smoothing is performed using a predetermined number of periodic noises before or after the periodic noise of the predetermined time in the time series. Process
Data processing device. A data processing device according to any one of claims 1 to 4,
The periodic noise suppression unit further includes a subtraction unit that subtracts the smoothed periodic noise from the time-series data,
Data processing device. The data processing apparatus according to claim 5, wherein
The periodic noise suppressor is
It further comprises a smoothing filter that subtracts the smoothed periodic noise from the time-series data and then smoothes it.
Data processing device. The configuration of the data processing device according to claim 5 or 6,
A plurality of stations connected to the data processing device, each receiving a positioning signal, and observing the carrier wave phase respectively;
Displacement observation system with Extract the periodic noise contained in each of the time series data,
Among the periodic noises arranged in time series, smoothing processing is performed for each of the plurality of periodic noises that are separated by the periodic noise period.
Data processing method. A process of extracting periodic noise included in each of the time series data;
Among the periodic noises arranged in time series, a process of smoothing for each of the plurality of periodic noises separated by the periodic noise period;
Is a data processing program for causing an information processing apparatus to execute.

Images