Content determination program, content determination device, and content determination method

Abstract

Conventional methods for determining the presence of substances were unable to perform X-ray diffraction measurements with sufficient accuracy on non-standard samples. [Solution] The present invention's substance detection program causes a computer to perform the following: a classification process (S11) which classifies the analysis result images read in the data reading process using a classifier configured to classify multiple previously acquired analysis result images into multiple types based on the difference between the presence and absence of the target substance; a determination probability calculation process (S12) which calculates a determination probability indicating the certainty of the classification result of the classification process; a first result determination process (S14) which, if the determination probability is greater than or equal to a preset threshold, records the classification result as a provisional detection result for the analysis result image that has undergone the classification process; and a second result determination process (S15) which, if the determination probability is less than the threshold, records an indeterminate result as a provisional detection result for the analysis result image that has undergone the classification process.

Description

Technical Field

[0001] The present invention relates to an inclusion determination program, an inclusion determination device, and an inclusion determination method. For example, it relates to an inclusion determination program, an inclusion determination device, and an inclusion determination method for determining whether a detection target object to be detected is contained in a test sample by using an XRD (X-Ray Diffraction) image generated by an X-ray analyzer.

Background Art

[0002] In order to prevent health hazards to the human body in an environment where dust is generated, measurements are being made on specific substances that may cause health hazards. One such measurement is the measurement of asbestos. For the qualitative and quantitative analysis of this asbestos, an X-ray analyzer is used. An example of a method for measuring asbestos using an X-ray analyzer is disclosed in Patent Document 1.

[0003] The analyzer described in Patent Document 1 is an analyzer that identifies substances contained in a sample by using a waveform pattern obtained by analyzing the sample, and includes a database in which standard waveform patterns of a large number of known substances are recorded, a measured waveform pattern obtained by analyzing the sample, and an evaluation of the similarity between the standard waveform pattern of each substance recorded in the database and the measured waveform pattern, and a candidate extraction unit that extracts a substance showing a standard waveform pattern with high similarity as an identification candidate, a first display column that displays information on one or more identification candidates extracted by the candidate extraction unit, and a second display column that displays information on the substance including the evaluation result of similarity in the candidate extraction unit for one or more specific substances determined in advance among the substances recorded in the database, and a display processing unit that creates a display screen arranged and displays it on a display unit.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

[0005] However, samples collected at actual work sites often contain many substances other than asbestos. Therefore, even when X-ray diffraction measurements are performed on samples containing asbestos, the resulting waveform patterns may differ from those obtained from standard samples containing a limited number of known substances, leading to a problem of low accuracy in determining the presence or absence of asbestos. [Means for solving the problem]

[0006] The present invention provides a computer to perform the following steps: a data reading process that reads an analysis result image obtained by an X-ray diffraction measurement performed on a sample by an X-ray analyzer; a classification process that classifies the analysis result image read in the data reading process using a classifier configured to classify a plurality of previously acquired analysis result images based on the difference between the presence and absence of the target substance; a definite probability calculation process that calculates a definite probability indicating the certainty of the classification result of the classification process; a first result determination process that records the classification result as a provisional detection result for the analysis result image after the classification process if the definite probability is greater than or equal to a preset threshold; and a second result determination process that records "uncertain" as the provisional detection result for the analysis result image after the classification process if the definite probability is less than the threshold.

[0007] The substance detection device according to the present invention includes a memory and a calculation unit that performs calculation processing using the memory and outputs a provisional detection result. The calculation unit performs the following: a data reading process that reads an analysis result image obtained by an X-ray diffraction measurement performed on a test sample by an X-ray analyzer; a classification process that classifies the analysis result image read in the data reading process using a classifier configured to classify a plurality of previously acquired analysis result images based on the difference between the presence and absence of the target substance; a definite probability calculation process that calculates a definite probability indicating the certainty of the classification result of the classification process; a first result determination process that records the classification result as the provisional detection result for the analysis result image after the classification process if the definite probability is greater than or equal to a preset threshold; and a second result determination process that records "uncertain" as the provisional detection result for the analysis result image after the classification process if the definite probability is less than the threshold.

[0008] The present invention provides a method for determining whether or not a target substance is present in a sample by computer calculation, comprising: a data reading process in which an analysis result image obtained by an X-ray diffraction measurement performed on the sample by an X-ray analyzer; a classification process in which a classifier configured to classify a plurality of previously acquired analysis result images based on the difference between the presence and absence of the target substance is used to classify the analysis result image read in the data reading process; a definite probability calculation process in which a definite probability indicating the certainty of the classification result of the classification process is calculated; a first result determination process in which, if the definite probability is greater than or equal to a preset threshold, the classification result is recorded as a provisional detection result for the analysis result image that has undergone the classification process; and a second result determination process in which, if the definite probability is less than the threshold, the computer performs the following:

[0009] The substance content determination program, substance content determination device, and substance content determination method according to the present invention can improve the accuracy of the first result determination process by performing a second result determination process. [Effects of the Invention]

[0010] According to the present invention, the substance content determination program, substance content determination method, and substance content determination device make it possible to perform substance content determination processing with high determination accuracy using a computer. [Brief explanation of the drawing]

[0011] [Figure 1] This is a flowchart illustrating the flow of the target substance detection process according to Embodiment 1. [Figure 2] This figure illustrates the XRD image obtained by X-ray diffraction measurement. [Figure 3] This is a block diagram of the substance content determination device according to Embodiment 1. [Figure 4] This is a flowchart illustrating the flow of the substance determination process according to Embodiment 1. [Figure 5] This graph illustrates the relationship between the accuracy rate of the provisional judgment result and the final judgment threshold according to Embodiment 1. [Modes for carrying out the invention]

[0012] For clarity of explanation, the following descriptions and drawings have been omitted and simplified as appropriate. Furthermore, each element shown in the drawings as a functional block performing various processes can be composed of a CPU (Central Processing Unit), memory, and other circuits in hardware terms, and implemented in software terms by programs loaded into memory. Therefore, it will be understood by those skilled in the art that these functional blocks can be implemented in various ways using hardware alone, software alone, or a combination thereof, and are not limited to any one of these. In each drawing, the same elements are denoted by the same reference numeral, and redundant explanations have been omitted where necessary.

[0013] Furthermore, the program described above includes, when loaded into a computer, a set of instructions (or software code) for causing the computer to perform one or more of the functions described in the embodiments. The program may be stored in a non-temporary computer-readable medium or a physical storage medium. Examples, but not limited to, include random-access memory (RAM), read-only memory (ROM), flash memory, solid-state drive (SSD) or other memory technologies, CD-ROM, digital versatile disc (DVD), Blu-ray® disc or other optical disc storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage devices. The program may be transmitted over a temporary computer-readable medium or a communication medium. Examples, but not limited to, include temporary computer-readable medium or a communication medium that includes electrically, optically, acoustically, or otherwise propagating signals.

[0014] Embodiment 1 The substance detection device 10 according to Embodiment 1 is used in part of the process of detecting a target substance in a test sample. In particular, the substance detection device 10 according to Embodiment 1 is used in part of the asbestos detection process, which determines whether or not asbestos is present in the test sample, with asbestos being the target.

[0015] Asbestos is found in various materials and has been reported to cause health problems in humans, creating a need to determine whether or not a material contains asbestos. In such asbestos detection processes, multiple inspections are performed by different personnel or multiple inspections of different types are performed to improve the accuracy of the inspection. The asbestos content determination device 10 according to Embodiment 1 generates a provisional detection result that can be used as one of these multiple inspections. Therefore, Figure 1 shows a flowchart illustrating the flow of the target substance detection process according to Embodiment 1.

[0016] In the following description, polarizing microscopy and electron microscopy are shown as the inspections performed by the person in charge in steps S4 and S8 of FIG. 1. However, this is just an example, and the types of inspections performed by the person in charge vary depending on the type of sample.

[0017] As shown in FIG. 1, in the target substance detection process according to Embodiment 1, first, an inspection sample is inspected with an X-ray analyzer to generate an analysis result image (for example, an XRD (X-Ray Diffraction) image) (step S1). Subsequently, the XRD image generated in step S1 is saved in the inclusion determination device (step S2).

[0018] Then, in the inclusion determination device, an inclusion determination process is performed using the XRD image to generate a provisional determination result indicating whether the detection target substance is contained in the inspection sample (step S3). Also, in the target substance detection process according to Embodiment 1, it is inspected by polarizing microscopy by the person in charge whether the inspection sample contains the detection target substance (for example, asbestos) (step S4).

[0019] Here, in the target substance detection process according to Embodiment 1, if the certainty of the provisional determination result generated in step S3 is above a certain level (the YES branch of step S5), it is determined whether the provisional determination result of step S3 and the result of the polarizing microscopy inspection of step S4 match (step S6). And when the provisional determination result and the result of the polarizing microscopy inspection match in step S6, the result of step S3 or step S4 is recorded as the inspection result for the inspection sample (step S7).

[0020] On the other hand, when the certainty of the preliminary determination result in step S5 does not reach a certain level (the NO branch in step S5), or when the preliminary determination result and the result of polarized light microscopy do not match in step S6 (the NO branch in step S6), in the target substance detection process according to Embodiment 1, another test is performed (step S8). This another test may be, for example, polarized light microscopy by a person other than the person in charge of step S4, or electron microscopy. And in the target substance detection process according to Embodiment 1, the test result for the test sample is recorded based on the test results of steps S4 and S8.

[0021] Thus, in the target substance detection process according to Embodiment 1, if the certainty of the preliminary determination result generated in the inclusion determination process based on the XRD image is above a certain level, by adopting the preliminary determination result as the result of one of the two tests, the burden on the test personnel is reduced. Therefore, the inclusion determination device that performs the inclusion determination process will be described in detail below.

[0022] In the inclusion determination system 1 according to Embodiment 1, a primary determination is made to determine the presence or absence of the detection target substance based on the analysis result image obtained by examining the test sample using an X-ray analyzer. Therefore, the analysis result image obtained from the X-ray analyzer will be described.

[0023] In the X-ray analyzer, X-ray diffraction measurement is performed on the test sample. In this X-ray diffraction measurement, while changing the irradiation angle of X-rays with respect to the test sample, the diffraction intensity of the X-rays reflected by the test sample is measured for each irradiation angle. Therefore, FIG. 2 shows a diagram for explaining the XRD image obtained by the X-ray diffraction measurement. In FIG. 2, four XRD images classified into four types are shown. The four types shown in FIG. 2 are also the types when classifying the XRD image with a classifier described later. Also, in FIG. 2, for two detection target substances, four types are set to classify the difference in whether they exist or not, but the number of types varies according to the number of detection target substances.

[0024] As shown in Figure 2, the analysis result image (e.g., XRD image) used as input in the substance detection system is a graph with the X-ray irradiation angle on the horizontal axis and the diffraction intensity on the vertical axis. The irradiation angle only needs to include at least the range containing the largest peak obtained from the substance to be detected.

[0025] In the example shown in Figure 2, amphibole and serpentinite minerals were used as the target materials. For amphibole minerals, diffraction intensity peaks are observed at irradiation angles of approximately 10.5, 20, and 30 degrees. In the example shown in Figure 2, the lower limit of the irradiation angle was set to the range that includes the highest diffraction intensity peak at 10.5 degrees. For serpentinite minerals, diffraction intensity peaks are observed at irradiation angles of approximately 12.1, 24, and 36 degrees. In the example shown in Figure 1, the upper limit of the irradiation angle was set to the range that includes the highest diffraction intensity peak at 12.1 degrees. Specifically, in the example shown in Figure 2, the irradiation angle was set to 10.3 to 12.5 degrees. By narrowing the range of irradiation angles in this way, which is narrower than the range that includes all the peaks of the target material, it is possible to reduce the time required to acquire XRD images. For example, if the irradiation angle is set to include a total of six peaks from amphibole and serpentinite minerals, it takes about 20 minutes per sample. However, by narrowing the irradiation angle to include just one peak each from amphibole and serpentinite minerals, it becomes possible to complete the measurement in about 3 minutes per sample.

[0026] Amphibole minerals include crocidolite, amosite, anthophyllite, tremolite, and actinolite. Serpentinite minerals include chrysotile.

[0027] In the examples shown in Figure 2, the upper left shows the XRD image when both amphibole and serpentinite minerals are undetectable, the upper right shows the XRD image when amphibole minerals are undetectable and serpentinite minerals are detected, the lower left shows the XRD image when amphibole minerals are detected and serpentinite minerals are undetectable, and the lower right shows the XRD image when both amphibole and serpentinite minerals are detected. As can be seen from these four XRD images, the presence or absence and position of peaks in the XRD image differ depending on whether or not the target substance is contained in the sample. Furthermore, although the examples shown in Figure 2 are XRD images obtained from a single sample, even when the target substance is contained in the sample, there may be slight shifts in the position of the diffraction intensity peaks or differences in the height of the diffraction intensity peaks. In addition, diffraction intensity peaks may occur even when the sample does not contain the target substance. In particular, in non-standard samples where it is unclear whether or not the target substance is present, and where other unspecified substances are present in the sample, the variation in the position and height of the diffraction intensity peaks becomes larger.

[0028] In the substance detection device of the substance detection system according to Embodiment 1, the XRD image is classified into one of several pre-set categories, including the variability of the XRD image. Furthermore, in the substance detection device of the substance detection system according to Embodiment 1, the certainty of the classification result is calculated as a certainty probability, making it possible to decide not to use XRD images with classification results that do not meet a certain certainty probability as the primary result of the inspection.

[0029] Therefore, the substance content determination system 1, including the substance content determination device according to Embodiment 1, will be described in detail. Figure 3 shows a block diagram of the substance content determination device 10 according to Embodiment 1. Note that Figure 1 illustrates a configuration in which the substance content determination device 10 is included as part of the substance content determination system 1.

[0030] As shown in Figure 3, the substance detection system 1 according to Embodiment 1 includes a substance detection device 10, an X-ray analyzer 20, and an external storage 30. The X-ray analyzer 20 performs X-ray diffraction measurement on the sample to generate an analysis result image (e.g., an XRD image). The external storage 30 is a storage unit that stores a trained model (e.g., trained parameters) used in the classifier described later. In Figure 3, the external storage 30 is shown as a separate block from the substance detection device 10, but the external storage 30 may be provided integrally with the substance detection device 10, or it may be installed in a different location via a network line.

[0031] The substance detection device 10 can be configured as dedicated hardware, but it can also be implemented using a computer equipped with memory and a processing unit. The following description will explain an example in which the substance detection device 10 is implemented by executing a substance detection program in the processing unit of a computer.

[0032] As shown in Figure 3, the substance detection device 10 includes a memory (e.g., a storage unit 13) and a calculation unit 14. The substance detection device 10 may also have a separate memory (not shown) used by the calculation unit 14 for calculation processing. The storage unit 13 includes a monitoring folder 13a where the X-ray analyzer 20 stores XRD images, and an output folder 13b where the calculation unit 14 stores preliminary determination results. Furthermore, an input unit 11 and a display unit 12 are connected to the substance detection device 10 as a user interface.

[0033] The arithmetic unit 14 performs data reading, classification, probability calculation, and result determination by executing the substance determination program. These processes may be performed by a single processor, different processors, or different computers. In particular, the classification process can be performed by a computer located in a different location from the substance determination device 10. The example shown in Figure 3 illustrates an example in which data reading, classification, probability calculation, and result determination are performed by a single processor (for example, the arithmetic unit 14). Figure 3 shows a diagram in which a data reading processing unit 15 performs data reading, a classifier 16 performs classification, a probability calculation processing unit 17 performs probability calculation, and a result determination processing unit 18 performs result determination, all implemented by executing a program.

[0034] The data loading processing unit 15 performs data loading processing to read the analysis result images obtained by the X-ray diffraction measurement performed on the inspection sample by the X-ray analyzer 20. The classifier 16 is configured to classify multiple previously acquired XRD images into multiple types based on the presence or absence of the target substance, and performs classification processing to classify the XRD images loaded in the data loading process. Details of this classifier 16 will be described later.

[0035] The confirmation probability calculation processing unit 17 performs a confirmation probability calculation process to calculate a confirmation probability that indicates the certainty of the classification result of the classification process. The result determination processing unit 18 performs a first result determination process and a second result determination process. The first result determination process is performed when the confirmation probability is equal to or greater than a predetermined threshold (for example, a confirmation determination threshold TH), and records the classification result as a provisional detection result for the XRD image that has undergone classification processing. The second result determination process is performed when the confirmation probability is less than the confirmation determination threshold TH, and records indetermination as a provisional detection result for the XRD image that has undergone classification processing.

[0036] Here, we will explain the classifier 16 in detail. The classifier 16 classifies the input XRD image into one of the four types shown in Figure 2. More specifically, if the target substance includes a first target substance (e.g., amphibole minerals) and a second target substance (e.g., serpentinite minerals), the classifier 16 classifies the XRD image into one of the following categories: Category 1 (upper left graph in Figure 2), where neither amphibole nor serpentinite minerals are present; Category 2 (lower left graph in Figure 2), where amphibole minerals are present and serpentinite minerals are absent; Category 3 (upper right graph in Figure 2), where amphibole minerals are absent and serpentinite minerals are present; or Category 4 (lower right graph in Figure 2), where both amphibole and serpentinite minerals are present.

[0037] A classifier 16 that performs such classification processing can be implemented using a predetermined function, but it can also utilize artificial intelligence (AI) technology. Therefore, the following explanation describes an example of implementing classifier 16 using AI technology. When classifying classifier 16 using AI technology, pre-trained parameters are required, as shown in Figure 3. These pre-trained parameters are generated using machine learning. Specifically, pre-trained parameters are generated by applying untrained parameters to classifier 16, inputting pre-prepared training data, and updating the training parameters so that the classification result of the input XRD image matches the training data. This training data is, for example, a combination of an XRD image and the results of polarized light microscopy examination performed on the inspection sample from which the XRD image was obtained. By performing machine learning using such training data, pre-trained parameters are generated such that the probability of a classification result obtained from polarized light microscopy examination results for an XRD image is greater than or equal to a threshold (e.g., a determination threshold TH). Then, classifier 16 uses these pre-trained parameters to perform classification processing on newly loaded XRD images.

[0038] Here, the probability of certainty is the probability that the classification result is correct, and it takes a value between 0% and 100%. For example, it can be calculated by the distance between the output value that classifier 16 outputs when a certain XRD image is input and the output value that classifier 16 outputs when the probability of certainty is 100%. It is preferable that the probability of certainty TH is the probability at which the classification result by classifier 16 becomes error-free (the accuracy rate becomes 100%).

[0039] Next, Figure 4 shows a flowchart illustrating the flow of the substance detection process according to Embodiment 1. The flowchart shown in Figure 4 also explains the operation of the substance detection device 10.

[0040] As shown in Figure 4, in the substance detection device 10, the data reading processing unit 15 reads the XRD images stored in the monitoring folder 13a by the X-ray analyzer 20 (step S10). Next, the data reading processing unit 15 inputs the read XRD images to the classifier 16, which classifies the input XRD images into one of several types (for example, the four types shown in Figure 2) (step S11). The substance detection device 10 also uses the probability calculation processing unit 17 to calculate the probability of certainty indicating the certainty of the classification result in step S11 (step S12). After that, the substance detection device 10 uses the result determination processing unit 18 to perform the determination processing in steps S13 to S15.

[0041] In step S13, the result determination processing unit 18 determines whether the probability of determination calculated in step S12 is equal to or greater than a predetermined threshold (determination threshold TH) (step S13). If the result determination processing unit 18 determines in step S13 that the probability of determination is equal to or greater than the determination threshold TH, the result determination processing unit 18 performs a first result determination process to record the classification result as a provisional detection result for the XRD image that has undergone classification processing (step S14).

[0042] The provisional determination result is digital data such as a text file, and the classification result is shown in a predetermined format by the inclusion determination device 10. The provisional determination result is stored in the output folder 13b. In the provisional determination result generated in this first result determination process, for example, a flag indicating the presence or absence of amphibole minerals and serpentinite minerals may be recorded. For example, the presence flag may be indicated by text data such as "○" or "Present," and the absence flag by text data such as "×" or "Absent."

[0043] On the other hand, if the probability of confirmation in step S13 is determined to be less than the confirmation threshold, the result determination processing unit 18 performs a second result determination process to record indetermination as a provisional detection result of the XRD image that has undergone classification processing (step S15).

[0044] The provisional determination result is digital data such as a text file, and the classification result is shown in a predetermined format by the content determination device 10. The provisional determination result is stored in the output folder 13b. In the provisional determination result generated in this second result determination process, for example, a flag indicating that the presence of amphibole minerals and serpentinite minerals is uncertain may be recorded. The uncertain flag may be indicated by text data such as "△" or "uncertain".

[0045] Here, we will explain the magnitude of the confirmation threshold TH. As mentioned above, it is preferable that the confirmation threshold TH is the probability of confirmation where the misclassification rate of the classification result is 0%, that is, the accuracy rate of the classification result is 100%. Therefore, Figure 5 shows a graph illustrating the relationship between the accuracy rate of the provisional classification result and the confirmation threshold according to Embodiment 1. In the graph shown in Figure 5, the horizontal axis shows the confirmation threshold from 99.0 to 99.9%, and the vertical axis shows the accuracy rate of the classification result and the uncertainty rate, which indicates the percentage of classification results that were determined to be uncertain.

[0046] As shown in Figure 5, increasing the definitive determination threshold TH increases the rate of indeterminate determination. However, by setting the definitive determination threshold TH to 99.5% or higher, the accuracy rate becomes 100%. In this way, by setting the accuracy rate to 100%, the reliability of the provisional determination result in which the type has been determined can be increased. When the definitive determination threshold TH, which results in an accuracy rate of 100%, is 95.5%, the rate of indeterminate determination is 24.6%, meaning that for approximately 75% of the test samples, the provisional determination result generated by the content determination device 10 can be used as the result of one of the two tests, thus reducing the number of tests performed by the person in charge by one.

[0047] The classifier 16 classifies XRD images obtained from non-standard samples into one of four types. However, even if trained parameters are generated using a vast amount of training data, or if the accuracy of the function used for classification is improved, it is difficult to achieve a 100% accuracy rate for the classification results. However, in the substance detection device 10 according to Embodiment 1, by labeling classification results with a probability of certainty lower than the certainty threshold TH as uncertain, only classification results with a 100% accuracy rate are extracted, making it possible to achieve a 100% accuracy rate for confirmed classification results. As a result, in the target substance detection process using the substance detection device 10 according to Embodiment 1, this certain classification result can be used as a provisional determination result, reducing the number of inspections by the person in charge and thus reducing the time required for the target substance detection process.

[0048] Furthermore, in the substance detection device 10 according to Embodiment 1, if an XRD image is obtained that is limited to a range containing at least one irradiation intensity peak obtained from the substance to be detected in the X-ray diffraction measurement performed by the X-ray analyzer 20, a preliminary determination result with high accuracy (for example, an accuracy rate of 100%) can be obtained. In other words, by using the substance detection device 10 according to Embodiment 1, it is possible to obtain a preliminary determination result with high accuracy while shortening the inspection time in the X-ray analyzer 20.

[0049] Furthermore, XRD images from non-standard samples whose contained substances are not predetermined show shifts in peak height and position compared to XRD images obtained from standard samples whose contained substances are predetermined. Therefore, when trained parameters are generated using training data generated from results obtained from standard samples, there is a problem that the classification accuracy of XRD images obtained from non-standard samples deteriorates. In the substance detection device 10 according to Embodiment 1, when generating trained parameters used by the classifier 16, training data created using non-standard samples is used instead of standard samples. As a result, the substance detection device 10 according to Embodiment 1 can improve the classification accuracy for non-standard samples. In particular, asbestos is often contained in building materials, but since there are countless sample patterns in building materials, when such non-standard samples are classified based on standard samples, the uncertainty rate of the classification result increases, and the effectiveness of generating a provisional judgment result using XRD images decreases. However, by using the substance detection device 10, the usefulness of the provisional judgment result for non-standard samples can be increased.

[0050] Furthermore, the test samples subject to inspection include samples that have deteriorated over time. However, the XRD images obtained from standard samples differ significantly from those obtained from non-standard samples that have deteriorated over time. Using training data that does not include non-standard samples that have deteriorated over time reduces the classification accuracy for these samples. However, in the substance determination device 10 according to Embodiment 1, it is possible to improve the classification accuracy for non-standard samples that have deteriorated over time by including non-standard samples that have deteriorated over time in the training data and generating trained parameters.

[0051] It should be noted that the present invention is not limited to the embodiments described above, and can be modified as appropriate without departing from the spirit of the invention. [Explanation of symbols]

[0052] 1. Substance detection system 10 Content determination device 11 Input section 12 Display section 13 Storage section 14 Arithmetic section 15. Data reading processing unit 16 Classifier 17. Determined Probability Calculation Processing Unit 18 Result determination processing unit 20 X-ray analysis device 30 External storage

Claims

1. The data loading process involves reading the analysis result image obtained from the X-ray diffraction measurement performed on the inspection sample by the X-ray analysis device, and A classification process is performed to classify the analysis result images read in the data reading process using a classifier configured to classify multiple previously acquired analysis result images into multiple types based on the difference between the presence and absence of the target substance, A definite probability calculation process that calculates a definite probability indicating the certainty of the classification result of the aforementioned classification process, If the probability of certainty is greater than or equal to a predetermined threshold, a first result determination process is performed which records the classification result as a provisional detection result for the analysis result image that has undergone the classification process, If the probability of certainty is less than the threshold, a second result determination process is performed to record uncertainty as the provisional detection result of the analysis result image after the classification process, Have the computer perform this task. The substance to be detected includes a first substance to be detected and a second substance to be detected. The classifier processes the analysis result image. A first classification in which neither the first target substance nor the second target substance is present, A second classification in which the first target substance is present and the second target substance is absent, A third classification in which the first target substance is absent and the second target substance is present, A fourth classification in which both the first target substance and the second target substance exist, A substance detection program that classifies substances into one of the following categories.

2. The substance content determination program according to claim 1, wherein the threshold is 95.5%.

3. The content determination program according to claim 1, wherein the substance to be detected includes at least one of amphibole minerals and serpentinite minerals.

4. The content determination program according to claim 1, wherein the classifier is machine-trained using training data which is a combination of the analysis result image and the results of polarized light microscopy examination performed on the inspection sample from which the analysis result image was obtained, so that the probability of the analysis result image being classified by the polarized light microscopy examination results is greater than or equal to the threshold.

5. The aforementioned test sample is a non-standard sample in which it has not been determined whether or not it contains the substance to be detected. The content determination program according to claim 4, wherein the analysis result image used as training data is obtained by the X-ray diffraction measurement performed on a non-standard sample.

6. The content determination program according to claim 1, wherein the analysis result image includes one of a plurality of peaks corresponding to the substance to be detected.

7. Memory and It includes a calculation unit that performs calculation processing using the memory and outputs a provisional detection result, The aforementioned arithmetic unit, The data loading process involves reading the analysis result image obtained from the X-ray diffraction measurement performed on the inspection sample by the X-ray analysis device, and A classification process is performed to classify the analysis result images read in the data reading process using a classifier configured to classify multiple previously acquired analysis result images based on the difference between the presence and absence of the target substance, A definite probability calculation process that calculates a definite probability indicating the certainty of the classification result of the aforementioned classification process, If the probability of certainty is greater than or equal to a predetermined threshold, a first result determination process is performed to record the classification result as the provisional detection result for the analysis result image that has undergone the classification process, If the probability of certainty is less than the threshold, a second result determination process is performed to record uncertainty as the provisional detection result of the analysis result image after the classification process, We will implement the following: The substance to be detected includes a first substance to be detected and a second substance to be detected. The classifier processes the analysis result image. A first classification in which neither the first target substance nor the second target substance is present, A second classification in which the first target substance is present and the second target substance is absent, A third classification in which the first target substance is absent and the second target substance is present, A fourth classification in which both the first target substance and the second target substance exist, A substance determination device that classifies substances into one of the following categories.

8. A method for determining whether or not a target substance is present in a test sample by computer calculation processing, The data loading process involves reading the analysis result image obtained from the X-ray diffraction measurement performed on the inspection sample by the X-ray analysis device, and A classification process is performed to classify the analysis result images read in the data reading process using a classifier configured to classify multiple previously acquired analysis result images based on the difference between the presence and absence of the target substance, A definite probability calculation process that calculates a definite probability indicating the certainty of the classification result of the aforementioned classification process, If the probability of certainty is greater than or equal to a predetermined threshold, a first result determination process is performed which records the classification result as a provisional detection result for the analysis result image that has undergone the classification process, If the probability of certainty is less than the threshold, a second result determination process is performed to record uncertainty as the provisional detection result of the analysis result image after the classification process, The computer is instructed to perform the following: The substance to be detected includes a first substance to be detected and a second substance to be detected. The classifier processes the analysis result image. A first classification in which neither the first target substance nor the second target substance is present, A second classification in which the first target substance is present and the second target substance is absent, A third classification in which the first target substance is absent and the second target substance is present, A fourth classification in which both the first target substance and the second target substance exist, A method for determining the contents of a substance, classifying it into one of the following categories.

Images