Information processing apparatus, information processing program, and measurement support method

Abstract

To provide a technique capable of appropriately determining the usefulness of a measurement result by a measuring device without requiring a burden on a device user.SOLUTION: An information processing device 2 used by being connected to a measurement device 1 capable of outputting a signal waveform indicating a change in signal intensity with respect to a predetermined physical quantity as a measurement result of an object to be measured includes a data acquisition unit 21 that acquires profile data obtained by imaging the signal waveform output from the measurement device 1, a reference setting unit 22 that prepares a signal waveform obtained by imaging a reference signal waveform as training data, and an AI determination unit 23 that determines the usefulness of a measurement result related to the profile data based on a comparison result between the profile data and the training data.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to an information processing device, an information processing program, and a measurement support method. [Background technology]

[0002] For example, measuring devices such as an inductively coupled plasma (ICP) optical emission spectrometer, an ICP mass spectrometer, or an atomic absorption spectrometer may be used to analyze elements contained in a sample (see, for example, Patent Document 1). Such measuring devices are capable of outputting a signal waveform that indicates a change in signal intensity relative to a predetermined physical quantity (such as light wavelength or time) as a measurement result for the sample to be measured. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2021-135297 Summary of the Invention [Problem to be solved by the invention]

[0004] Measurement results obtained by a measurement device may contain, for example, spectral interference or superimposed noise components, which may impair the usefulness of the measurement results and, as a result, may have a negative impact on the reliability of the measurement results.

[0005] The present disclosure provides a technique that enables the usefulness of measurement results obtained by a measurement device to be determined appropriately and without imposing a burden on the device user. [Means for solving the problem]

[0006] According to one aspect of the present disclosure, An information processing device used in connection with a measuring device capable of outputting a signal waveform indicating a change in signal intensity relative to a predetermined physical quantity as a measurement result for an object to be measured, a data acquisition unit that acquires profile data obtained by imaging the signal waveform output from the measurement device; a reference setting unit that prepares an image of a reference signal waveform as training data; an AI determination unit that determines the usefulness of a measurement result related to the profile data based on a comparison result between the profile data and the teacher data; An information processing device comprising:

[0007] According to another aspect of the present disclosure, a computer connected to a measuring device capable of outputting a signal waveform showing a change in signal strength relative to a predetermined physical quantity as a measurement result for an object to be measured; a data acquisition unit that acquires profile data obtained by imaging the signal waveform output from the measurement device; a reference setting unit that prepares an image of a reference signal waveform as training data; an AI determination unit that determines the usefulness of a measurement result related to the profile data based on a comparison result between the profile data and the teacher data; A program is provided to function as a

[0008] According to yet another aspect of the present disclosure, a data acquisition step for acquiring profile data that visualizes the signal waveform output from a measurement device capable of outputting a signal waveform that indicates a change in signal intensity relative to a predetermined physical quantity as a measurement result of an object to be measured; a reference setting procedure in which an image of a reference signal waveform is prepared as training data; an AI determination procedure for determining the usefulness of the measurement results related to the profile data based on the comparison result between the profile data and the teacher data; An information processing method is provided, including: [Effects of the Invention]

[0009] According to the technology of the present disclosure, it is possible to appropriately determine the usefulness of measurement results obtained by a measurement device without imposing a burden on the device user. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a block diagram schematically illustrating an example configuration of a measurement control system according to an embodiment of the present disclosure. [Figure 2] FIG. 10 is a flow diagram illustrating a specific example of a procedure for information processing performed by an information processing device according to an embodiment of the present disclosure. [Figure 3] 10 is an explanatory diagram illustrating a specific example of a comparison result between profile data and teacher data by an information processing device according to an embodiment of the present disclosure. FIG. [Figure 4] FIG. 10 is an explanatory diagram showing a specific example of an output mode of a measurement result by an information processing device according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, an embodiment of the present disclosure will be described with reference to the drawings.

[0012] (1) Measurement control system configuration First, the configuration of a measurement control system according to an embodiment of the present disclosure will be described. FIG. 1 is a block diagram schematically showing an example of the configuration of a measurement control system according to this embodiment.

[0013] 1, the measurement control system according to this embodiment is configured to include at least a measurement device 1 and an information processing device 2. The measurement device 1 and the information processing device 2 are communicatively connected via a wired or wireless communication line (not shown).

[0014] The measuring device 1 measures a sample to be measured, and is, for example, an inductively coupled plasma optical emission spectrometer (ICP-OES) that performs optical emission spectroscopic analysis using a high-frequency inductively coupled plasma (ICP) as a light source. However, the measuring device 1 is not limited to an ICP-OES, and may be of other types (e.g., an ICP mass spectrometer, an atomic absorption spectrometer, etc.). In either type, a known specific configuration of the measuring device 1 can be used, and therefore a detailed description thereof will be omitted here.

[0015] There are no particular limitations on the sample that can be measured by the measurement device 1, and it can be a powder, a solid, a liquid (fluid), a mixture of these, etc. In the following explanation, an example will be given in which the sample is a powder containing a metal element.

[0016] The measurement results obtained by the measuring device 1 may be used, for example, to analyze elements contained in a sample, but are not limited to a specific use and may be used for other purposes. The types of measurement results obtained by the measuring device 1 will be described in detail below.

[0017] The information processing device 2 performs information processing to support the measurement processing in the measurement device 1, and has computer functions. Specifically, the information processing device 2 is configured to have a display unit 11 such as a display, an operation unit 12 such as a keyboard or mouse operated by the user, an output unit 13 such as a printer, a memory unit 14 that stores predetermined programs and various data, and a control unit 20 that reads out and executes the predetermined programs from the memory unit 14.

[0018] Of these, the control unit 20 is configured to have, for example, a CPU (Central Processing Unit), but may also be configured to have a dedicated processing circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array). In either configuration, the control unit 20 executes a predetermined program (software), and the information processing by that software is specifically realized using hardware resources.

[0019] For example, the control unit 20 executes a predetermined program (software) to realize functions as a data acquisition unit 21, a standard setting unit 22, an AI (Artificial Intelligence) determination unit 23, and an output control unit 24. The specific contents of the functions realized by each of these units 21 to 24 will be described in detail later.

[0020] As described above, the functions of the data acquisition unit 21, the standard setting unit 22, the AI ​​determination unit 23, and the output control unit 24 are specifically realized by the execution of a predetermined program (software) in the control unit 20. In other words, the predetermined program that realizes the functions of these units 21 to 24 corresponds to one embodiment of the "information processing program" in the present disclosure. In this case, the information processing program is provided from outside via a communication line or a storage medium and installed in the storage unit 14 of the information processing device 2.

[0021] (2) Measurement results obtained by measuring device 1 Next, a brief description will be given of the measurement results obtained by the measurement device 1. Here, a case where the measurement device 1 is an ICP-OES will be taken as an example.

[0022] ICP-OES performs measurements to analyze the elements contained in a sample, which is the object to be measured. Specifically, ICP-OES uses plasma as a light source, and introduces an atomized sample solution into the plasma to emit element-specific spectra. The types of component elements are determined (qualitative) from the positions of these spectra (wavelengths of light), and the concentration of each element is determined (quantitative) from the light emission intensity.

[0023] By performing such measurements, ICP-OES can output the emission intensity value and the concentration value calculated from the emission intensity for each wavelength of light (absorption wavelength). The absorption wavelength is the dominant wavelength (wavelength at which a peak is observed) of the element being analyzed. Both the emission intensity value and the concentration value are peak values ​​at the absorption wavelength.

[0024] Furthermore, ICP-OES can also output, as measurement results, the emission spectrum relating to these values, that is, the signal waveform showing the change in emission intensity over the wavelength range including the above-mentioned absorption wavelength.

[0025] The signal waveform referred to here is continuous linear data that can be electromagnetically recorded and indicates the change in signal intensity relative to a predetermined physical quantity. The predetermined physical quantity corresponds, for example, to a wavelength range including the above-mentioned absorption wavelength. However, this is not limited to this, and for example, a time axis with a fixed interval can also correspond to the predetermined physical quantity. In other words, the signal waveform is composed of linear data indicating the change in signal intensity continuously acquired at a fixed interval. The signal intensity can correspond to the above-mentioned emission intensity. Note that the unit system indicating the change in signal intensity is uniquely determined in the data space to be processed. The data format of such signal waveforms is not limited to a specific format, as long as it can be obtained as an independent data group directly or through signal conversion. For example, numerical data can be in either plain text format or binary format. Furthermore, non-linguistic data such as images can be obtained as continuous linear data directly or through data conversion, so long as it has a data structure that can ultimately be obtained as numerical data. Specifically, if the signal waveform is a specific physical quantity at regular intervals, continuous linear data such as a histogram can be used in the case of a wavelength axis, and continuous linear data such as a chromatogram can be used in the case of a time axis at regular intervals.

[0026] ICP-OES quantifies elements by measuring their emission spectra at specific absorption wavelengths. However, different elements have different absorption wavelengths. For example, phosphorus (P) has four dominant wavelengths: 213.617 nm, 214.914 nm, 178.221 nm, and 177.434 nm. Zinc (Zn) has five dominant wavelengths: 202.548 nm, 206.200 nm, 213.857 nm, 330.258 nm, and 334.502 nm. Therefore, ICP-OES sometimes measures each element at multiple absorption wavelengths and selects the results to use as the basis for quantification. This is because, depending on the type of element contained in the sample, spectral interference or noise components may be superimposed on the measurement results.

[0027] For example, among the main wavelengths of P mentioned above, copper (Cu) peaks can be observed at wavelengths of 213.617 nm, 214.914 nm, and 177.434 nm. Therefore, if Cu is included as a constituent element in the sample, spectral interference may occur at these wavelengths. Also, for example, among the main wavelengths of Zn mentioned above, the wavelength of 202.548 nm completely overlaps with the wavelength of Cu, so measuring only this wavelength may result in misinterpretation.

[0028] However, spectral interference occurs depending on the concentration of the interference source, and does not necessarily occur at adjacent wavelengths. It is not possible to determine whether spectral interference exists unless measurements are taken at multiple absorption wavelengths for each element. Furthermore, the superposition of noise components does not occur at specific absorption wavelengths.

[0029] Therefore, when quantifying the constituent elements in a sample, in order to prevent the usefulness of the ICP-OES measurement results from being impaired by eliminating the effects of spectral interference and noise components, measurements are taken at multiple absorption wavelengths for each element, and then the measurement results on which quantification is based are selected from these.

[0030] In this way, when multiple measurement results with different measurement parameters such as absorption wavelength are output from ICP-OES, the data format of each measurement result can be any format that allows the final data to be obtained independently, and each measurement result is considered to be a unique data set.

[0031] When multiple measurement results are obtained using ICP-OES, it becomes necessary to select the measurement result on which quantification is based. This selection is generally performed manually, based on empirical judgment, after referring to each measurement result. However, leaving the selection of measurement results to human judgment results in a situation where the judgment relies too heavily on the experience of the person making the judgment, which may result in an inappropriate judgment. Inappropriate selection of measurement results reduces the usefulness of the ICP-OES measurement results, which in turn adversely affects the reliability of the measurement results. Furthermore, leaving the selection of measurement results to human judgment may place an excessive burden on the person making the judgment, for example, when there is a large amount of information to process.

[0032] In consideration of the above, in the measurement control system according to this embodiment, the information processing device 2 performs the information processing described below on the measurement results obtained by the ICP-OES, which is the measurement device 1.

[0033] (3) Information processing procedure by the information processing device 2 Next, a specific description will be given of the procedure of information processing performed by the information processing device 2. Note that the procedure of information processing described below is mainly controlled by the control unit 20 of the information processing device 2. FIG. 2 is a flowchart showing a specific example of the procedure of information processing performed by the information processing device 2 according to this embodiment.

[0034] Prior to information processing by the information processing device 2, the measurement device 1 performs measurements on the sample to be measured. At this time, the measurement device 1 performs measurements at each of a plurality of absorption wavelengths, as described above. In other words, the measurement device 1 uses each of the plurality of absorption wavelengths as a measurement parameter. Therefore, when a measurement is performed on a certain object to be measured, the measurement device 1 outputs a plurality of measurement results obtained using a plurality of measurement parameters.

[0035] When measurement results are output from the measurement device 1, the information processing device 2 acquires the measurement results (step 101; hereinafter, step will be abbreviated as "S"). If multiple measurement results are output from the measurement device 1, each measurement result is acquired.

[0036] Then, when a plurality of measurement results are acquired, the control unit 20 of the information processing device 2 extracts a measurement result of interest from the acquired measurement results (S102). The method for extracting the measurement results is not particularly limited, but it is possible to extract the measurement results based on a predetermined priority order for each measurement parameter (e.g., absorption wavelength).

[0037] When the measurement results of interest are extracted, the control unit 20 of the information processing device 2 functions as the data acquisition unit 21, which acquires profile data, which is data that visualizes the signal waveforms included in the extracted measurement results (S103). The data acquisition unit 21 acquires the profile data as follows. For example, if the signal waveforms are output as numerical data, the data acquisition unit 21 plots the numerical data in a two-dimensional coordinate space to visualize the signal waveforms and generate the profile data. Also, for example, if the signal waveforms are output as non-verbal data such as images, the data acquisition unit 21 uses the output data as is as the profile data. In this way, the data acquisition unit 21 acquires the profile data.

[0038] Meanwhile, in the information processing device 2, the function of the control unit 20 as the standard setting unit 22 prepares teacher data to be compared with the profile data acquired by the data acquisition unit 21 (S104). The teacher data may be prepared after the profile data is acquired, or may be prepared prior to the acquisition of the profile data, as long as it is prepared before the AI ​​determination (S105) described below.

[0039] The training data is an image of a signal waveform that serves as a reference for the AI ​​judgment (S105) described below. An example of a reference signal waveform is a signal waveform obtained by measuring a standard solution. The attributes of the standard solution (such as the type and concentration of the contained components) are known in advance. In other words, the training data is, for example, a signal waveform of a standard solution whose attributes are known, and is an image of the signal waveform in a state where no spectral interference or the like occurs with respect to the attribute. Therefore, there may be multiple types of training data for different attributes. Note that the signal waveform represented by the training data is not necessarily limited to the signal waveform of the standard solution, and may be any other signal waveform as long as it serves as a reference for the AI ​​judgment (S105) described below. An example of such a signal waveform is a signal waveform in a state where spectral interference occurs.

[0040] Such teacher data may be obtained, for example, by measuring a standard solution or the like using the measurement device 1. The obtained teacher data is then stored in the memory unit 14 of the information processing device 2, and the teacher data is prepared by reading it from the memory unit 14 during AI determination (S105), which will be described later. Note that the preparation of teacher data is not limited to the above-described method, and other methods may also be used. Another possible method is, for example, accessing an external device such as a cloud server and reading the data from the external device. Another possible method is, for example, storing previously acquired signal waveforms, generating one or more signal waveforms based on the stored contents, and using the generated signal waveforms as teacher data.

[0041] After acquiring the profile data (S103) and preparing the teacher data (S104), the AI ​​determination unit 23 functioning in the control unit 20 of the information processing device 2 compares the acquired profile data with the prepared teacher data (S105). Then, the AI ​​determination unit 23 determines the usefulness of the measurement results related to the profile data based on the comparison result between the profile data and the teacher data (S105).

[0042] FIG. 3 is an explanatory diagram showing a specific example of the results of comparing profile data with training data. The example shows the results of comparing measurement results when the wavelength of 202.548 nm, one of the main wavelengths of Zn, is used as the measurement parameter. In the two-dimensional coordinate space shown in the figure, the horizontal axis represents the wavelength range including the absorption wavelength, which is a predetermined physical quantity, and the vertical axis represents the signal intensity for each wavelength. In other words, in this two-dimensional coordinate space, signal waveforms (continuous linear data) showing changes in signal intensity relative to the predetermined physical quantity are shown as profile data and training data.

[0043] The AI ​​judgment unit 23 compares the profile data with the training data while they are both visualized. Therefore, the AI ​​judgment unit 23 can compare the profile data with the training data while focusing on the image features of each.

[0044] The image feature may be, for example, at least one of the central wavelength at the signal waveform peak position, the signal intensity at the signal waveform peak position, the half-width of the signal waveform peak, the number of signal waveform peaks, etc. Of course, other features may also be used as the image feature, or a combination of these may be used as appropriate.

[0045] The AI ​​determination unit 23 then determines whether or not the image features of interest are consistent with each other. For example, if the image feature of interest is the central wavelength of the signal waveform peak position, the AI ​​determination unit 23 recognizes the difference between the central wavelength of the profile data and the central wavelength of the training data, and if the magnitude of that difference is within a preset tolerance, it determines that the image features of interest are consistent with each other; otherwise, it determines that the image features of interest are inconsistent with each other. The AI ​​determination unit 23 performs this consistency determination for the image feature of interest. If multiple image features are being considered, it performs this determination for all of them.

[0046] As a result, if consistency with the teacher data is ensured for the image feature of interest, the AI ​​judgment unit 23 judges that the measurement results related to the profile data compared with this are useful. On the other hand, if consistency with the teacher data is not ensured for at least one image feature, the AI ​​judgment unit 23 determines that the measurement results related to the profile data compared with this are not useful, since there is a possibility that they are affected by spectral interference, noise components, etc. In this way, the AI ​​judgment unit 23 judges the usefulness of the measurement results related to the profile data based on the comparison result between the profile data and the teacher data.

[0047] The above-described series of processes (S102 to S105) from extracting the measurement results of interest to determining the usefulness of the measurement results is repeated until (S106) it is completed for all of the multiple measurement results obtained from the measurement device 1. As a result, for each of the multiple measurement results obtained from the measurement device 1, the usefulness determination result by the AI ​​determination unit 23 is assigned as variable information such as a flag.

[0048] Here, the example of determining usefulness focuses on the image feature quantities of the profile data and the training data, but this is not necessarily limited to this. For example, the AI ​​determination unit 23 and the standard setting unit 22, which prepares the training data used by the AI ​​determination unit 23, may perform data processing as described below to determine the usefulness of the measurement results related to the profile data acquired by the data acquisition unit 21.

[0049] For example, profile data (i.e., a unique dataset of continuous linear data) may be analyzed using a mathematical model based on chemoinformatics technology constructed from training data that appropriately expresses the identity or difference between samples. Then, based on the analysis results, the AI ​​determination unit 23, which functions as an artificial intelligence, may accurately determine whether the profile data is good or bad (i.e., the usefulness of the measurement results related to the profile data).

[0050] In this case, training data that adequately expresses the identity or difference between samples serves as a high-quality dataset with prediction errors removed as much as possible. A high-quality dataset with prediction errors removed as much as possible is a model standardized (auto-scaled) with ideal linear data that removes possible noise.

[0051] Mathematical models constructed from such data sets are robust models, and ideal linear data with noise removed to enable estimation performance is preferable. When using a mathematical model constructed from ideal linear data, it is an interpretation-oriented method model that understands and predicts the structure of the data. For example, continuous linear data can be statistically analyzed using regression analysis such as multivariate analysis to estimate the attributable data.

[0052] When using a model trained on ideal linear data, it is a result-oriented model that pursues the accuracy of predicting unknown data. For example, by comparing the results of the model created by machine learning with the expected value and brushing up the training dataset, you can verify with actual data whether there is overfitting, and by considering the results, you can improve the accuracy of the training set.

[0053] To improve the accuracy of a model trained on ideal linear data, identify the causes of discrepancies and improve the underlying dataset. To improve the underlying dataset, assign new classes to points that were difficult to classify or did not perform as expected. For example, the first choice for improving a model is to modify the dataset, such as by increasing the amount of input data to improve comprehensiveness. Therefore, if adding data does not produce any positive results, modify parameters such as class weights to address discrepancies. The effectiveness of machine learning requires eliminating the possibility that changing one result to match the expectation will result in other results that do not meet expectations. Simply improving the underlying dataset may not improve the accuracy of a model trained on ideal linear data. Intentionally avoiding data that may be considered noise, such as by limiting the spectral wavelengths used, may improve accuracy.

[0054] A characteristic of elemental spectral datasets is that when the wavelengths of continuous linear data are close, the intensities such as absorbance are similar. A characteristic of elemental spectral datasets is that they contain noise. A characteristic of elemental spectral datasets is that data other than the maximum value of signal intensity is also important. A characteristic of elemental chromatogram datasets is that, in addition to the characteristic of spectral datasets, the amount of data increases over time.

[0055] In this way, regression analysis is performed on a spectral or chromatographic data set. Similar techniques can be applied to data preprocessing to improve classification estimation performance.

[0056] Smoothing a data set reduces noise. However, depending on the smoothing method, there is a risk of losing information about maximum and minimum values. Calculating the slope of a data set using higher-order differentiation corrects the baseline, extracts uncorrected spectral information, and obtains time changes. However, this has the opposite effect to smoothing.

[0057] Well-known data processing methods for linear continuous datasets include the simple moving average method, linearly weighted moving average method, exponentially weighted moving average method, and Savitzky-Golay method. The combination of data processing methods and differential orders for these linear continuous datasets is selected by validating the model using training data or by considering the normal distribution of noise. This data processing method for linear continuous datasets is incorporated, and classification is performed to categorize multiple objects, dividing the data into predefined linear data classes. In the dataset creation phase after classification, classes are prepared for the number of objects to be classified, and datasets labeled as linear continuous data classified into each class are prepared. By repeatedly executing, analyzing, and solving these models, accuracy is improved, and the machine learning judgment results are output to regression.

[0058] In regression, quantitative value data, which is the correct information for each region, is provided as the output result of the measurement device 1, and the results obtained by regression are sequentially compared with the results obtained by the measurement device 1 to make a final acceptance / rejection decision. The output results of the measurement device 1 used in regression utilize information such as the expected quantitative value, signal strength, signal priority, and standard deviation or relative standard deviation indicating signal stability in a specific region. A final acceptance / rejection decision is made based on the calculated value for that expected value, and the acceptance / rejection result is obtained from the frequency of agreement between the expected value and the calculated value. The means for communicating the acceptance / rejection to the person performing the analysis can be any method, such as a unique expression using numbers, letters, symbols, or character modifications.

[0059] Specifically, the AI ​​judgment unit 23 compares (e.g., sequentially matches) the usefulness judgment result by the AI ​​judgment unit 23 with output signals other than the signal waveform from the measurement device 1 (S107). The output signals to be compared are, for example, numerical signals such as values ​​of luminescence intensity or concentration. The usefulness of such numerical signals can be determined by determining whether they fall within a predetermined allowable range. For this reason, the AI ​​judgment unit 23 may compare the usefulness judgment result based on the comparison result between the profile data and the teacher data with the usefulness judgment result of output signals other than the signal waveform, and may perform a final acceptance / rejection decision, such as adopting the result if they match, or rejecting the result if they do not.

[0060] In this way, when determining usefulness, the AI ​​determination unit 23 may compare the output signal from the measurement device 1 with signals other than the signal waveform, and then determine whether or not the usefulness is acceptable.

[0061] Through the data processing described above, the usefulness of the measurement results related to the profile data is determined. That is, the AI ​​determination unit 23 determines the usefulness of the measurement results related to the profile data acquired by the data acquisition unit 21 through an analysis process that utilizes the training data prepared by the standard setting unit 22. In this case, the accuracy of the training data prepared by the standard setting unit 22 can also be improved by using machine learning based on feedback of various predetermined information (for example, the determination results by the AI ​​determination unit 23).

[0062] After the AI ​​judgment unit 23 judges the usefulness of all measurement results obtained from the measurement device 1, the function of the output control unit 24 in the control unit 20 in the information processing device 2 then extracts the measurement results to be selected based on the usefulness judgment results by the AI ​​judgment unit 23 (S107).

[0063] For example, consider a case where multiple measurement results with different measurement parameters are output from the measurement device 1, and priorities have been set for each measurement result in advance. In this case, the output control unit 24 first checks the usefulness determination result by the AI ​​determination unit 23 for the measurement result with the highest priority. If the measurement result is determined to be useful, the output control unit 24 extracts the measurement result with the next highest priority as the measurement result to be selected. On the other hand, if the measurement result is determined to be unuseful, the output control unit 24 repeats the same process for the measurement result with the next highest priority. In this way, the output control unit 24 extracts the measurement result with the highest priority from among the measurement results determined to be useful by the AI ​​determination unit 23 as the measurement result to be selected.

[0064] In other words, when the measuring device 1 is capable of outputting multiple measurement results obtained for a certain object under measurement using multiple measurement parameters, the output control unit 24 extracts a measurement result to be selected from the multiple measurement results obtained by the measuring device 1 based on the usefulness determined by the AI ​​determination unit 23. Note that, although the example given here uses a pre-set priority order to extract the measurement result to be selected, this is not necessarily limited to this, and extraction may be performed using other methods as long as they are based on the usefulness determined by the AI ​​determination unit 23.

[0065] Thereafter, the measurement results obtained by the measuring device 1 are output on at least one of the display unit 11 or the output unit 14 of the information processing device 2. At this time, in the information processing device 2, the function as the output control unit 24 in the control unit 20 performs information processing as described below to support the output of the measurement results by the display unit 11 or the output unit 14.

[0066] 4 is an explanatory diagram showing a specific example of an output mode for measurement results. The illustrated example shows a case where measurement results obtained for each of multiple samples 1 to 4, which are multiple objects to be measured, using multiple measurement parameters (e.g., wavelength ○○○ nm, wavelength △△△ nm, ...) are output in a list format. Each measurement result shown in the diagram is divided into subitems, such as concentration (ratio), concentration (relative standard deviation = %RSD), absorbance intensity (c / s), absorbance intensity (relative standard deviation = %RSD), etc.

[0067] Prior to outputting the measurement results in such a list format, the output control unit 24 determines a specific output mode for the list format (S109) based on the usefulness determination result by the AI ​​determination unit 23. More specifically, the output control unit 24 determines the output mode for the list format so that, for example, measurement results determined not to be useful by the AI ​​determination unit 23 are output in a different mode from other results (i.e., measurement results determined to be useful) so that they can be distinguished from the other results.

[0068] In the specific example shown in Figure 4, the display columns for measurement results determined to be unuseful and the display columns for the measurement parameters to which those measurement results correspond are surrounded by a bold frame in an output format, making these display columns distinguishable from other display columns. Note that while the example in the figure uses a bold frame as an example, the output format is not limited to this as long as it makes the usefulness determination results distinguishable. For example, in addition to surrounding the display columns with a bold frame, other output formats may be used, such as using different display colors for display columns and characters, different fonts or sizes for characters, or adding a predetermined mark or hatching to the display columns.

[0069] That is, the output control unit 24 changes the output mode of the measurement results from the measurement device 1 depending on the usefulness determined by the AI ​​determination unit 23. Then, the output control unit 24 generates an image in a format that can be output by at least either the display unit 11 or the output unit 14, while following such an output mode.

[0070] As a result, at least one of the display unit 11 and the output unit 14 will output a list format as shown in the example (an output format that allows the usefulness judgment results by the AI ​​judgment unit 23 to be identified).

[0071] (4) Effects Obtained by the Present Embodiment According to this embodiment, one or more of the following effects can be obtained.

[0072] (a) In this embodiment, in the information processing device 2 used in connection with the measurement device 1, the AI ​​judgment unit 23 judges the usefulness of the measurement results related to the profile data based on the comparison result between the profile data acquired by the data acquisition unit 21 and the teacher data prepared by the standard setting unit 22. As a result, it becomes possible to judge the measurement results obtained by the measurement device 1 as not useful if, for example, spectral interference occurs or noise components are superimposed. Therefore, when utilizing the measurement results obtained by the measurement device 1, it becomes possible to avoid measurement results judged to be not useful, thereby preventing adverse effects on the reliability of the measurement results.

[0073] In this embodiment, such a judgment of usefulness is made based on the results of comparing the visualized profile data with the training data. Therefore, the judgment can be made quantitatively, for example, focusing on image features. This makes it possible to eliminate the risk of the judgment being inappropriate, unlike when the judgment is made artificially based on empirical rules. Moreover, even when the amount of information to be processed is large, there is no risk of the burden on the person making the judgment being excessive.

[0074] In other words, according to this embodiment, it is possible to appropriately determine the usefulness of the measurement results obtained by the measurement device without placing any burden on the device user, and as a result, it is possible to improve the reliability when utilizing the measurement results obtained by the measurement device 1.

[0075] (b) As explained in this embodiment, when the AI ​​determination unit 23 determines usefulness, if the AI ​​determination unit 23 determines whether the usefulness is acceptable or not by comparing the output signal other than the signal waveform from the measurement device 1, the usefulness determination can be made more appropriate. In other words, by going through the comparison, the reliability of the usefulness determination can be further improved.

[0076] (c) As explained in this embodiment, if the output control unit 24 changes the output mode of the measurement results from the measurement device 1 depending on the usefulness determined by the AI ​​determination unit 23, the usefulness determination result by the AI ​​determination unit 23 can be easily recognized by referring to the output content of the measurement results. This provides great convenience for users of the measurement control system including the measurement device 1. This also leads to improved reliability when utilizing the measurement results obtained by the measurement device 1.

[0077] (d) As described in this embodiment, when the output control unit 24 extracts a measurement result to be selected from multiple measurement results obtained by the measuring device 1 based on the usefulness determined by the AI ​​determination unit 23, it is possible to prevent the extraction of measurement results that, for example, have spectral interference or noise components superimposed thereon. In other words, when the measuring device 1 is capable of outputting multiple measurement results obtained for a certain object under measurement using multiple measurement parameters, it is possible to extract only useful results from these multiple measurement results. This improves the reliability of utilizing the measurement results obtained by the measuring device 1, and also prevents users of the measurement control system including the measuring device 1 from feeling inconvenienced as a result.

[0078] (5) Modifications, etc. Although one embodiment of the present disclosure has been specifically described above, the present disclosure is not limited to the above-described embodiment and can be modified as appropriate without departing from the spirit of the present disclosure.

[0079] For example, in the above-described embodiment, the measurement device 1 is mainly an ICP-OES, but the present invention is not limited to this and can be applied to other types of measurement devices (for example, an ICP mass spectrometer, an atomic absorption spectrometer, etc.). Furthermore, the sample to be measured is not limited to a powder containing a metal element, and may be a powder, solid, liquid (fluid), a mixture of these, etc.

[0080] Furthermore, in the above embodiment, an example has been described in which there is only one measurement device 1 in the system, but this is not limiting, and there may be multiple measurement devices 1 in the system. The same applies to the information processing device 2.

[0081] Furthermore, in the above-described embodiment, specific examples of profile data, teacher data (see Figure 3), output contents of measurement results (see Figure 4), etc. have been given, but these are merely specific examples and are not limited to these specific examples. [Explanation of symbols]

[0082] 1... measuring device, 2... information processing device, 11... display unit, 12... operation unit, 13... output unit, 14... memory unit, 20... control unit, 21... data acquisition unit, 22... reference setting unit, 23... AI determination unit, 24... output control unit

Claims

1. An information processing device used in connection with a measuring device capable of outputting a signal waveform indicating a change in signal intensity relative to a predetermined physical quantity as a measurement result for an object to be measured, a data acquisition unit that acquires profile data obtained by imaging the signal waveform output from the measurement device; a reference setting unit that prepares an image of a reference signal waveform as training data; an AI determination unit that determines the usefulness of a measurement result related to the profile data based on a comparison result between the profile data and the teacher data; An information processing device comprising:

2. The AI ​​determination unit, when determining the usefulness, compares the signal waveform with an output signal from the measuring device other than the signal waveform and determines whether the usefulness is acceptable or not. The information processing device according to claim 1 .

3. an output control unit that changes the output mode of the measurement result by the measurement device depending on the usefulness determined by the AI ​​determination unit; The information processing device according to claim 1 or 2, comprising:

4. When the measurement device is capable of outputting a plurality of measurement results obtained using a plurality of measurement parameters for a certain object to be measured, the output control unit extracts a measurement result to be selected from the plurality of measurement results based on the usefulness determined by the AI ​​determination unit. The information processing device according to claim 3 .

5. a computer connected to a measuring device capable of outputting a signal waveform showing a change in signal strength relative to a predetermined physical quantity as a measurement result for an object to be measured; a data acquisition unit that acquires profile data obtained by imaging the signal waveform output from the measurement device; a reference setting unit that prepares an image of a reference signal waveform as training data; an AI determination unit that determines the usefulness of a measurement result related to the profile data based on a comparison result between the profile data and the teacher data; A program that functions as a

6. a data acquisition step for acquiring profile data that visualizes the signal waveform output from a measurement device capable of outputting a signal waveform that indicates a change in signal intensity relative to a predetermined physical quantity as a measurement result of an object to be measured; a reference setting procedure in which an image of a reference signal waveform is prepared as training data; an AI determination procedure for determining the usefulness of a measurement result related to the profile data based on a comparison result between the profile data and the teacher data; An information processing method including:

Images