Data processing apparatus, data processing method, and data processing program

Abstract

To provide a data processing technique for reducing an amount of deviation from time series data before noise removal processing, which occurs when the noise removal processing of time series data is performed.SOLUTION: The data processing apparatus includes: a fitting processing unit for performing the fitting processing of a pre-selected function based on an error between acquired time series data and a pre-selected function; a subtraction unit for subtracting the acquired time series data by using the function subjected to the fitting processing; a noise removal unit for removing noise by performing low-pass filter processing on the time series data after the subtraction; and an addition unit for adding the function subjected to the fitting processing to the time series data after the noise removal and outputting the result to an output device.SELECTED DRAWING: Figure 8

Description

Technical Field

[0001] The present disclosure relates to a data processing apparatus, a data processing method, and a data processing program.

Background Art

[0002] A noise removal method (a noise removal method by low-pass filter processing) is known in which processing such as discrete cosine transform is performed on time-series data, and inverse discrete cosine transform is calculated using terms up to a predetermined number of terms to perform noise removal processing. In this noise removal method, by performing noise removal processing after removing the change in the entire time-series data in advance, it is possible to reduce the amount of deviation of the time-series data after noise removal processing from the time-series data before noise removal processing.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] On the other hand, in the case of the above noise removal method, if the change in the entire time-series data cannot be accurately removed before noise removal processing, the effect of reducing the deviation amount may not be sufficiently obtained.

[0005] The present disclosure provides a data processing technique for reducing the amount of deviation from time-series data before noise removal processing that occurs during noise removal processing of time-series data.

Means for Solving the Problems

[0006] A data processing apparatus according to one aspect of the present disclosure includes a fitting processing unit that performs fitting processing on the preselected function based on an error between the acquired time-series data and the preselected function, a subtraction unit that subtracts the acquired time-series data using the function on which the fitting processing has been performed, a noise removal unit that removes noise by performing low-pass filter processing on the time-series data after subtraction, and an addition unit that adds the function on which the fitting processing has been performed to the time-series data after noise removal and outputs the result to an output device.

[0007] A data processing method according to one aspect of the present disclosure includes steps in which a computer performs fitting processing on the preselected function based on an error between the acquired time-series data and the preselected function, subtracts the acquired time-series data using the function on which the fitting processing has been performed, removes noise by performing low-pass filter processing on the time-series data after subtraction, and adds the function on which the fitting processing has been performed to the time-series data after noise removal and outputs the result to an output device.

[0008] A data processing program according to one aspect of the present disclosure causes a computer to perform steps in which fitting processing is performed on the preselected function based on an error between the acquired time-series data and the preselected function, the acquired time-series data is subtracted using the function on which the fitting processing has been performed, noise is removed by performing low-pass filter processing on the time-series data after subtraction, and the function on which the fitting processing has been performed is added to the time-series data after noise removal and output to an output device.

Advantages of the Invention

[0009] According to the present disclosure, it is possible to reduce the amount of deviation from the time-series data before noise removal processing that occurs during noise removal processing of time-series data.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2A

Figure 2B

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12A

Figure 12B

Figure 13

Embodiments for Carrying Out the Invention

[0011] Hereinafter, each embodiment will be described with reference to the accompanying drawings. In the present specification and the drawings, for components having substantially the same functional configuration, the same reference numerals are given and redundant descriptions are omitted.

[0012] [First Embodiment] <System Configuration of Data Processing System> First, the system configuration of a data processing system including a data processing apparatus according to the first embodiment will be described. FIG. 1 is a diagram showing an example of the system configuration of the data processing system.

[0013] The data processing system 100 includes a manufacturing apparatus 110, a data processing apparatus 120, and an output apparatus 130.

[0014] The manufacturing apparatus 110 is an arbitrary apparatus that manufactures an object and outputs time-series data. The time-series data output by the manufacturing apparatus 110 includes any process data acquired for measurement, control, or management of the manufacturing process during or after the manufacture of the object. Note that the process data here includes at least any one of temperature data, pressure data, flow rate data, voltage data, image data, and audio data.

[0015] The data processing apparatus 120 acquires the time-series data (time-series data before noise removal processing) output by the manufacturing apparatus 110, and performs noise removal processing after removing the change in the entire acquired time-series data, thereby outputting the time-series data after noise removal processing.

[0016] The output device 130 outputs the time-series data after the noise removal process performed by the data processing device 120. The output method by the output device 130 is arbitrary. The output device 130 may generate a graph based on the time-series data after the noise removal process, and for example, display the generated graph to an operator working in the manufacturing process. Alternatively, the output device 130 may record the generated graph in a predetermined recording device so that the operator can analyze it. Note that the output device 130 may also display or record the time-series data itself after the noise removal process (that is, the numerical data before graphing).

[0017] <Concerns about General Noise Removal Methods> Next, before explaining the details of the data processing device 120 according to the present embodiment, concerns about general noise removal methods will be explained. FIGS. 2A and 2B are the first and second diagrams for explaining the concerns about general noise removal methods.

[0018] Both FIGS. 2A and 2B show the time-series data (here, voltage data) before the noise removal process output by the manufacturing device 110, and the time-series data after the noise removal process that is obtained by performing a discrete cosine transform process on the time-series data before the noise removal process and calculating the inverse discrete cosine transform up to a predetermined number of terms. Among these, FIG. 2A shows the case where the noise removal process is performed by performing the discrete cosine transform process and calculating the inverse discrete cosine transform up to the 8th term (showing the case where noise is removed by performing a low-pass filter process). On the other hand, FIG. 2B shows the case where the noise removal process is performed by performing the discrete cosine transform process and calculating the inverse discrete cosine transform up to the 16th term.

[0019] As shown in FIG. 2A, when the number of terms for calculating the inverse discrete cosine transform is small, although the noise is sufficiently removed, in the time range indicated by reference numeral 201 (in the time range where the change in the entire time-series data is steep), the amount of deviation from the time-series data before the noise removal process tends to increase.

[0020] On the one hand, as shown in FIG. 2B, when the number of terms for calculating the inverse discrete cosine transform is large, in the time range indicated by reference numeral 202 (the time range where the change in the entire time series data is steep), the amount of deviation from the time series data before the noise removal process tends to be small. However, when the number of terms for calculating the inverse discrete cosine transform is large, it becomes difficult to sufficiently remove the oscillation of the noise.

[0021] Thus, in the case of a noise removal method that performs discrete cosine transform processing and calculates the inverse discrete cosine transform using terms up to a predetermined number of terms to perform noise removal processing, whether the number of terms for calculating the inverse discrete cosine transform is large or small, each has its own advantages and disadvantages.

[0022] On the other hand, a configuration can be considered in which the number of terms for calculating the inverse discrete cosine transform is reduced so that noise can be sufficiently removed, and the amount of deviation is reduced by removing the steep change in the entire time series data in advance.

[0023] However, in the case of such a noise removal method, there is a concern that if the steep change in the entire time series data cannot be accurately removed before the noise removal process, it is difficult to sufficiently obtain the effect of reducing the amount of deviation.

[0024] Therefore, in the data processing apparatus 120 according to the present embodiment, an optimal conversion function is selected for the time series data before the noise removal process, and fitting processing is performed using the selected conversion function. Thereby, according to the data processing apparatus 120 according to the present embodiment, it becomes possible to accurately remove the steep change in the entire time series data. As a result, according to the data processing apparatus 120 according to the present embodiment, it is possible to reliably reduce the amount of deviation from the time series data before the noise removal process that occurs during the noise removal process of the time series data. Hereinafter, the details of the data processing apparatus 120 according to the present embodiment will be described.

[0025] <Hardware Configuration of Data Processing Apparatus> First, the hardware configuration of the data processing device 120 will be described. FIG. 3 is a diagram showing an example of the hardware configuration of the data processing device.

[0026] The data processing device 120 includes a processor 301, a memory 302, an auxiliary storage device 303, an I / F (Interface) device 304, a communication device 305, and a drive device 306. Each piece of hardware of the data processing device 120 is interconnected via a bus 307.

[0027] The processor 301 includes various arithmetic devices such as a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit). The processor 301 executes various programs by reading various programs (for example, data processing programs, etc.) onto the memory 302.

[0028] The memory 302 includes main memory devices such as a ROM (Read Only Memory) and a RAM (Random Access Memory). The processor 301 and the memory 302 form a so-called computer, and the computer realizes various functions by the processor 301 executing various programs read onto the memory 302.

[0029] The auxiliary storage device 303 stores various programs and various data used when various programs are executed by the processor 301. For example, the conversion function storage unit 410 described later is realized in the auxiliary storage device 303.

[0030] The I / F device 304 is a connection device that connects an operation device 311 and a display device 312, which are examples of a user interface device for an administrator of the data processing device 120 to operate the data processing device 120. The communication device 305 is a communication device for communicating with external devices (for example, the manufacturing device 110 and the output device 130) via a network (not shown).

[0031] The drive device 306 is a device for setting the recording medium 313. The recording medium 313 here includes media that record information optically, electrically, or magnetically, such as CD-ROMs, flexible disks, magneto-optical disks, etc. The recording medium 313 may also include semiconductor memories that record information electrically, such as ROMs, flash memories, etc.

[0032] Note that various programs installed in the auxiliary storage device 303 are installed, for example, when the distributed recording medium 313 is set in the drive device 306 and the various programs recorded on the recording medium 313 are read by the drive device 306. Alternatively, various programs installed in the auxiliary storage device 303 may be installed by being downloaded from a network via the communication device 305.

[0033] <Functional Configuration of the Data Processing Device in the Function Selection Phase> Next, the functional configuration of the data processing device 120 will be described. The data processing device 120 includes functions realized in the function selection phase and functions realized in the data processing phase.

[0034] The function selection phase is a phase in which an optimal conversion function used when performing fitting processing on the time-series data before noise removal processing is selected in advance. On the other hand, the data processing phase is a phase in which, by using the conversion function selected in the function selection phase to perform fitting processing on the time-series data before noise removal processing, the changes in the entire time-series data are removed in advance, and then noise removal processing is performed. Here, the functional configuration of the data processing device 120 in the function selection phase will be described.

[0035] FIG. 4 is a diagram showing an example of the functional configuration of the data processing apparatus in the function selection phase. A data processing program is installed in the data processing apparatus 120, and when the program is executed, the data processing apparatus 120 functions as a data acquisition unit 401, a fitting processing unit 402, and a selection unit 403 in the function selection phase.

[0036] The data acquisition unit 401 acquires the time-series data before noise removal processing (for example, see reference numeral 411) output by the manufacturing apparatus 110. The data acquisition unit 401 notifies the acquired time-series data before noise removal processing to the fitting processing unit 402 and the selection unit 403, respectively.

[0037] The fitting processing unit 402 sequentially reads out a plurality of types of conversion functions stored in advance in the conversion function storage unit 410, and performs fitting processing on the time-series data before noise removal processing. Specifically, the fitting processing unit 402 adjusts the parameters included in the read conversion function so that the error obtained by summing the squares of the differences between the values at each time of the read conversion function and the values at each time of the time-series data before noise removal processing over the entire time range is minimized.

[0038] The fitting processing unit 402 adjusts the parameters for each of the plurality of types of conversion functions by performing the same processing. Reference numeral 412 shows a state in which a logarithmic function is read out as a conversion function and fitting processing is performed on the time-series data before noise removal processing by adjusting the parameters.

[0039] The selection unit 403 selects the optimal conversion function by comparing the errors with the time-series data before noise removal processing for each of the plurality of types of conversion functions on which fitting processing has been performed.

[0040] <Details of the Functional Configuration of the Data Processing Apparatus in the Function Selection Phase> Next, details of each part included in the functional configuration of the data processing apparatus 120 in the function selection phase will be described.

[0041] (1) Details of the conversion function storage unit 410 FIG. 5 is a diagram showing a specific example of a conversion function stored in the conversion function storage unit of the data processing apparatus in the function selection phase. The conversion function storage unit 410 stores, as the conversion function, at least any one of a logarithmic function, an N-th order polynomial function, an exponential function, a rational function, a trigonometric function, a hyperbolic function, an inverse trigonometric function, a normal distribution function, or any function obtained by combining them.

[0042] When performing fitting processing on voltage data Vf, which is an example of time-series data before noise removal processing, each of the plurality of types of conversion functions used includes, as parameters, ·a 0 、a 1 、···a N 、 ·b 0 、b 1 、···b M 、 ·N, M, σ, μ, etc. are included.

[0043] (2) Details of the fitting processing unit 402 FIG. 6 is a diagram showing a specific example of the processing executed by the fitting processing unit of the data processing apparatus in the function selection phase. The fitting processing unit 402 sequentially reads out the conversion function from the conversion function storage unit 410, and performs fitting processing on the time-series data (reference numeral 411) before noise removal processing by adjusting the parameters included in the read conversion function.

[0044] In the example of FIG. 6, as the conversion function, a logarithmic function is read out, and fitting processing is performed by adjusting the parameters a 0 、a 1 included in the logarithmic function, and the state of calculating the fitting data (reference numeral 601) is shown.

[0045] The fitting processing unit 402 performs the same processing on the N - th polynomial function, exponential function, rational function, trigonometric function, hyperbolic function, inverse trigonometric function, normal distribution function, or a function combining them, which are stored in the conversion function storage unit 410, as the conversion function other than the logarithmic function. Thereby, the fitting processing unit 402 can calculate the fitting data of each conversion function.

[0046] (3) Details of the selection unit 403 FIG. 7 is a diagram showing a specific example of the processing executed by the selection unit of the data processing device in the function selection phase. As shown by reference numeral 700, the selection unit 403 calculates the error by summing the squares of the differences at each time between the time - series data before the noise removal process and the fitting data of each conversion function over a predetermined time range.

[0047] The predetermined time range mentioned here may be the entire time range of the time - series data before the noise removal process acquired by the data acquisition unit 401, or a partial time range. When the predetermined time range is a partial time range, it may be a time range where the change in the entire time - series data before the noise removal process is steep. Specifically, the predetermined time range may be the time range indicated by reference numeral 201 in FIG. 2A.

[0048] The example of FIG. 7 shows a state where the error in the entire time range between the fitting data (reference numeral 601) when the conversion function is a logarithmic function and the time - series data before the noise removal process is calculated (see reference numeral 701). Also, the example of FIG. 7 shows a state where the selection unit 403 selects the logarithmic function as the conversion function for which the fitting data with the minimum error is calculated.

[0049] <Functional configuration of the data processing device in the data processing phase> Next, the functional configuration of the data processing device 120 in the data processing phase will be described. FIG. 8 is a diagram showing an example of the functional configuration of the data processing device in the data processing phase. As described above, a data processing program is installed in the data processing device 120. When the program is executed, the data processing device 120 functions as a data acquisition unit 401, a fitting processing unit 801, a subtraction unit 802, a noise removal unit 803, and an addition unit 804 in the data processing phase.

[0050] Among these, since the data acquisition unit 401 has already been described with reference to FIG. 4, the description thereof will be omitted here.

[0051] The fitting processing unit 801 performs fitting processing on the time-series data before noise removal processing using the conversion function (here, a logarithmic function) selected in the function selection phase. Specifically, the fitting processing unit 801 calculates the sum of the squares of the differences between the values at each time of the time-series data (reference numeral 811) before noise removal processing acquired by the data acquisition unit 401 in the data processing phase and the values at each time of the logarithmic function selected in the function selection phase over the entire time range, and adjusts the parameters included in the logarithmic function so that the error is minimized. Thereby, the fitting processing unit 801 can calculate fitting data (reference numeral 812).

[0052] The subtraction unit 802 subtracts the time-series data (reference numeral 811) before noise removal processing acquired by the data acquisition unit 401 in the data processing phase using the fitting data (reference numeral 812). Thereby, the subtraction unit 802 can remove the overall steep changes from the time-series data before noise removal processing.

[0053] Reference numeral 813 is an example of the time-series data after the steep changes in the entire time-series data before noise removal processing have been removed by the subtraction unit 802.

[0054] The noise removal unit 803 performs noise removal processing on the time series data indicated by reference numeral 813. Specifically, the noise removal unit 803 performs discrete cosine transform processing on the time series data indicated by reference numeral 813, and performs noise removal processing by calculating the inverse discrete cosine transform for terms up to a predetermined number of terms.

[0055] Reference numeral 814 is an example of the time series data after noise removal processing, on which noise removal processing has been performed by the noise removal unit 803.

[0056] The addition unit 804 can generate the time series data before noise removal processing (reference numeral 811), which has been restored to the state before the steep changes were removed, by adding the fitting data indicated by reference numeral 812 to the time series data after noise removal processing, on which noise removal processing has been performed by the noise removal unit 803.

[0057] Thereby, it is possible to generate the time series data after noise removal processing (reference numeral 815) from which noise has been removed from the time series data before noise removal processing (reference numeral 811).

[0058] <Details of the functional configuration of the data processing apparatus in the data processing phase> Next, details of each part included in the functional configuration of the data processing apparatus 120 in the data processing phase will be described.

[0059] (1) Details of the subtraction unit 802 FIG. 9 is a diagram showing a specific example of the processing executed by the subtraction unit of the data processing apparatus in the data processing phase. The subtraction unit 802 subtracts the fitting data (reference numeral 812) that has been fitted by the fitting processing unit 801 using a selected conversion function (here, a logarithmic function) from the time series data before noise removal processing (reference numeral 811) acquired by the data acquisition unit 401 in the data processing phase, thereby generating the time series data after subtraction (reference numeral 813). Further, the subtraction unit 802 notifies the generated time series data after subtraction (reference numeral 813) to the noise removal unit 803.

[0060] (2) Details of the noise removal unit 803 FIG. 10 is a diagram showing a specific example of processing executed by the noise removal unit of the data processing apparatus in the data processing phase. The noise removal unit 803 performs discrete cosine transform processing on the time-series data (reference numeral 813) after subtraction, and calculates the inverse discrete cosine transform up to a predetermined number of terms to perform noise removal processing. In the example of FIG. 10, by calculating the inverse discrete cosine transform up to the 8th term (that is, by substituting M = 8), time-series data (reference numeral 814) after noise removal processing is generated.

[0061] (3) Details of the addition unit 804 FIG. 11 is a diagram showing a specific example of processing executed by the addition unit of the data processing apparatus in the data processing phase. The addition unit 804 adds the fitting data (reference numeral 812) to the time-series data (reference numeral 814) after noise removal processing, and returns it to the state before the steep change is removed. As a result, the addition unit 804 can calculate the time-series data (reference numeral 815) after noise removal processing and output it to the output device 130.

[0062] <Function selection processing and data processing flow> Next, the function selection processing and data processing flow in the data processing system 100 will be described. FIGS. 12A and 12B are examples of the first and second flowcharts showing the function selection processing and data processing flow.

[0063] As shown in FIG. 12A, when the function selection processing is started, in step S1201, the data processing apparatus 120 acquires the time-series data before noise removal processing output by the manufacturing apparatus 110.

[0064] In step S1202, the data processing apparatus 120 performs fitting processing on the acquired time-series data before noise removal processing using a plurality of types of conversion functions.

[0065] In step S1203, the data processing device 120 calculates the difference at each time between the time-series data before the noise removal process and the fitting data, squares the difference, and calculates the error by summing over a predetermined time range.

[0066] In step S1204, the data processing device 120 selects the conversion function that minimizes the calculated error and ends the function selection process.

[0067] As shown in FIG. 12B, when the data processing starts, in step S1211, the data processing device 120 acquires the time-series data before the noise removal process output by the manufacturing device 110.

[0068] In step S1212, the data processing device 120 performs fitting processing on the time-series data before the noise removal process using the conversion function selected in the function selection process.

[0069] In step S1213, the data processing device 120 removes the sharp changes in the entire time-series data before the noise removal process by subtracting the fitting data from the time-series data before the noise removal process.

[0070] In step S1214, the data processing device 120 performs discrete cosine transform processing on the time-series data after subtraction and calculates the inverse discrete cosine transform up to a predetermined number of terms to perform the noise removal process.

[0071] In step S1215, the data processing device 120 adds the fitting data to the time-series data after the noise removal process to return to the state before the sharp changes were removed.

[0072] In step S1216, the data processing device 120 determines whether the next time-series data has been output from the manufacturing device 110. If it is determined in step S1216 that the next time-series data has been output (if YES in step S1216), the process returns to step S1211.

[0073] On the other hand, if it is determined in step S1216 that the next time-series data has not been output (if NO in step S1216), the data processing is terminated.

[0074] <Effect of the noise removal method by the data processing device 120> Next, the effect of the noise removal method by the data processing device 120 will be described. FIG. 13 is a diagram showing the effect of the noise removal method by the data processing device. Similar to FIGS. 2A and 2B, it shows the time-series data before the noise removal process (also voltage data here) output by the manufacturing device 110, and the time-series data after the noise removal process obtained by performing a discrete cosine transform process on the time-series data before the noise removal process and calculating the inverse discrete cosine transform up to a predetermined number of terms. Note that the time-series data after the noise removal process shown in FIG. 13 is the time-series data after the noise removal process performed by calculating the inverse discrete cosine transform up to the 8th term by performing the discrete cosine transform process.

[0075] As is clear from the comparison between FIG. 13 and FIG. 2A, since the number of terms for calculating the inverse discrete cosine transform is small for the time-series data after the noise removal process shown in FIG. 13, similar to FIG. 2A, the noise is sufficiently removed, and after the steep changes in the entire time-series data are appropriately removed, since the noise removal process is performed, it can be seen that the amount of deviation from the time-series data before the noise removal process is small even in the time range with steep changes compared to FIG. 2A.

[0076] In this way, by adopting a configuration in which fitting processing is performed on the time-series data before noise removal processing using an optimal conversion function, according to the data processing apparatus 120 according to the first embodiment, it becomes possible to appropriately remove sharp changes in the entire time-series data. As a result, according to the data processing apparatus 120 according to the first embodiment, it is possible to surely reduce the amount of deviation from the time-series data before noise removal processing that occurs during the noise removal processing of the time-series data.

[0077] <Summary> As is clear from the above description, the data processing apparatus 120 according to the first embodiment performs fitting processing on the pre-selected conversion function based on the error between the time-series data acquired from the manufacturing apparatus 110 and the pre-selected conversion function, subtracts the acquired time-series data using the fitting data of the conversion function on which the fitting processing has been performed, performs discrete cosine transform processing on the time-series data after subtraction, calculates the inverse discrete cosine transform up to a predetermined number of terms to perform noise removal processing, adds the fitting data to the time-series data after noise removal processing, and outputs it to the output apparatus 130.

[0078] In this way, the data processing apparatus 120 according to the first embodiment performs fitting processing on the time-series data before noise removal processing using a pre-selected conversion function. Thereby, according to the data processing apparatus 120 according to the first embodiment, it becomes possible to appropriately remove sharp changes in the entire time-series data.

[0079] As a result, according to the data processing apparatus 120 according to the first embodiment, it is possible to surely reduce the amount of deviation from the time-series data before noise removal processing that occurs during the noise removal processing of the time-series data.

[0080] [Second Embodiment] In the above-described first embodiment, a case where noise removal processing is performed by performing discrete cosine transform processing on the time-series data before noise removal processing and calculating the inverse discrete cosine transform for terms up to a predetermined number of terms has been described. However, the low-pass filter processing for noise removal processing is not limited to this.

[0081] For example, instead of discrete cosine transform, it may be configured to perform noise removal processing by using an orthogonal sequence, such as discrete Fourier transform, discrete sine transform, etc.

[0082] Also, in the above-described first embodiment, when calculating the error between the time-series data before noise removal processing and the fitting data, it has been described as using the least squares error. However, an error other than the least squares error may be calculated.

[0083] Also, in the above-described first embodiment, it has been described as outputting the time-series data (reference numeral 815) after noise removal processing to the output device 130. However, the output method when outputting to the output device 130 is not limited to this. For example, in the case of outputting a graph based on the time-series data (reference numeral 815) after noise removal processing, the time-series data (reference numeral 815) after noise removal processing may be configured to be output by superimposing it on the time-series data (reference numeral 811) before noise removal processing.

[0084] [Other Embodiments] In the above-described first embodiment, a case where the data processing device 120 is configured as a separate body from the manufacturing device 110 has been described. However, the data processing device 120 may be configured as a part of the manufacturing device 110. Also, in the first embodiment, a case where the data processing device 120 is configured as a separate body from the output device 130 has been described. However, the data processing device 120 may be configured integrally with the output device 130.

[0085] Note that in the disclosed technology, forms such as the following supplementary notes can be considered. (Supplementary Note 1) A fitting processing unit that performs fitting processing on the pre-selected function based on the error between the acquired time series data and the pre-selected function; A subtraction unit that subtracts the acquired time series data using the function on which the fitting processing has been performed; A noise removal unit that removes noise by performing low-pass filter processing on the time series data after subtraction; An addition unit that adds the function on which the fitting processing has been performed to the time series data after noise removal and outputs it to an output device A data processing device having the above. (Appendix 2) The data processing device according to Appendix 1, wherein the function includes at least one of an Nth-degree polynomial function, an exponential function, a logarithmic function, a rational function, a trigonometric function, a hyperbolic function, an inverse trigonometric function, a normal distribution function, or a function obtained by combining them. (Appendix 3) The pre-selected function is A function selected from among a plurality of the functions based on the error in a predetermined time range between the acquired time series data and each of the plurality of functions on which fitting processing has been performed. The data processing device according to Appendix 2. (Appendix 4) The fitting processing unit is The data processing device according to any one of Appendices 1 to 3, which performs the fitting processing by adjusting parameters included in the pre-selected function based on the error between the acquired time series data and the pre-selected function. (Appendix 5) The noise removal unit is Removing noise by performing low-pass filter processing using an orthogonal function sequence on the time series data after subtraction. The data processing device according to any one of Appendices 1 to 4. (Appendix 6) The data processing device according to Appendix 5, wherein the orthogonal function sequence includes at least one of a discrete Fourier transform, a discrete cosine transform, and a discrete sine transform. (Appendix 7) The acquired time-series data includes at least one of temperature data, pressure data, flow rate data, voltage data, image data, and audio data. The data processing device according to any one of Appendices 1 to 6. (Appendix 8) The computer Performing a fitting process of the pre-selected function based on the error between the acquired time-series data and the pre-selected function; Subtracting the acquired time-series data using the function subjected to the fitting process; Removing noise by performing a low-pass filter process on the time-series data after subtraction; Adding the function subjected to the fitting process to the time-series data after noise removal and outputting the result to an output device A data processing method for executing the above steps. (Appendix 9) On the computer Performing a fitting process of the pre-selected function based on the error between the acquired time-series data and the pre-selected function; Subtracting the acquired time-series data using the function subjected to the fitting process; Removing noise by performing a low-pass filter process on the time-series data after subtraction; Adding the function subjected to the fitting process to the time-series data after noise removal and outputting the result to an output device A data processing program for causing the above steps to be executed.

[0086] Note that the configurations and the like described in the above embodiments are not limited to the configurations shown herein, such as combinations with other elements. Modifications can be made without departing from the spirit of the present invention, and can be appropriately determined according to the application form.

Explanation of Reference Numerals

[0087] 100: Data processing system 110: Manufacturing device 120: Data processing device 130: Output device 401: Data acquisition unit 402: Fitting processing unit 403: Selection unit 801: Fitting processing unit 802: Subtraction unit 803: Noise removal unit 804: Addition unit

Claims

1. A fitting processing unit that performs fitting processing on the pre-selected function based on the error between the acquired time-series data and the pre-selected function; A subtraction unit that subtracts the acquired time-series data using the function on which the fitting processing has been performed; A noise removal unit that removes noise by performing low-pass filter processing on the time-series data after subtraction; An addition unit that adds the function on which the fitting processing has been performed to the time-series data after noise removal and outputs the result to an output device A data processing device having the above components.

2. The data processing device according to claim 1, wherein the function includes at least one of an Nth-degree polynomial function, an exponential function, a logarithmic function, a rational function, a trigonometric function, a hyperbolic function, an inverse trigonometric function, a normal distribution function, or a function combined with any of them.

3. The pre-selected function is A function selected from among a plurality of the functions based on the error between the acquired time-series data and each of the plurality of functions on which fitting processing has been performed within a predetermined time range. The data processing device according to claim 2.

4. The fitting processing unit Performs the fitting processing by adjusting the parameters included in the pre-selected function based on the error between the acquired time-series data and the pre-selected function. The data processing device according to claim 1.

5. The noise removal unit Removes noise by performing low-pass filter processing on the time-series data after subtraction using an orthogonal function sequence. The data processing device according to claim 1.

6. The data processing device according to claim 5, wherein the orthogonal function sequence includes at least one of a discrete Fourier transform, a discrete cosine transform, and a discrete sine transform.

7. The acquired time-series data includes at least one of temperature data, pressure data, flow rate data, voltage data, image data, and audio data. The data processing device according to claim 1.

8. A computer Performs a step of performing fitting processing on the pre-selected function based on the error between the acquired time-series data and the pre-selected function; Performs a step of subtracting the acquired time-series data using the function on which the fitting processing has been performed; Performs a step of removing noise by performing low-pass filter processing on the time-series data after subtraction; A step of adding a function obtained by performing the fitting process to the time-series data after noise removal and outputting the result to an output device, and A data processing method for executing the above steps. **Claim 9** On a computer, A step of performing a fitting process on the pre-selected function based on the error between the acquired time-series data and the pre-selected function; A step of subtracting the acquired time-series data using the function obtained by performing the fitting process; A step of removing noise by performing a low-pass filter process on the time-series data after subtraction; A step of adding the function obtained by performing the fitting process to the time-series data after noise removal and outputting the result to an output device, and A data processing program for causing the computer to execute the above steps.

Images