Information Processing Apparatus, Information Processing Method, and Program

Abstract

To provide a technique that can accurately detect inspection objects that meet predetermined conditions from among inspection objects.SOLUTION: A learning result acquiring unit 120 acquires, via a communication unit 59, learning result information transmitted from a learning device 2. A target information acquiring unit 122 acquires, via the communication unit 59, target information for each of analysis objects for measurement, which is transmitted from an analysis device 1. A determining unit 126 determines the nature or quality of each of the analysis objects for measurement for which the target information is acquired based on the learning result information acquired by the learning result acquiring unit 120 and the target information acquired by the target information acquiring unit 122. An output unit 128 outputs results of determination performed by the determining unit 126.SELECTED DRAWING: Figure 4

Description

Technical Field

[0001] The present invention relates to an information processing apparatus, an information processing method, and a program.

Background Art

[0002] Conventionally, in the field of foods and the like, when there is a problem with the quality of the food itself or when the food contains foreign substances or the like, detecting and removing them is an important issue. In contrast, for example, a method has been proposed in which secondary differential processing is performed on an absorption spectrum obtained by irradiating light on a food or the like to be inspected, and a secondary differential spectroscopic image is created to detect foreign substances contained in the food or the like (see, for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in the technique described in Patent Document 1 above, it takes time for various arithmetic processes and acquisition of spectroscopic images. Further, the technique described in Patent Document 1 above simply applies predetermined arithmetic processes and performs foreign substance detection, and depending on the type of the object to be inspected and the data acquisition status, the detection accuracy may be insufficient. Furthermore, most of the conventional techniques including the technique described in Patent Document 1 above are mainly aimed at detecting foreign substances contained in the object to be inspected, and there are few that aim at detecting the quality of the object to be inspected.

[0005] The present invention has been made in view of such circumstances, and an object thereof is to provide a technique capable of accurately detecting an object satisfying predetermined conditions from among objects.

[0006] To achieve the above object, an information processing apparatus according to one aspect of the present invention is an information processing apparatus used for sorting an object, an acquisition means for acquiring information regarding the intensity of light in a predetermined wavelength region related to the object to be sorted, a determination means for applying statistical means to information regarding the intensity of reflection or scattering in a predetermined wavelength region in one or more normal objects and defective objects, and determining the property or quality of the object to be sorted based on the learned result, an output means for outputting the result of determination by the determination means, and includes.

[0007] An information processing method and program according to one aspect of the present invention are also provided as an information processing method or program corresponding to the information processing apparatus according to one aspect of the present invention.

Advantages of the Invention

[0008] According to the present invention, it is possible to provide a technique for accurately detecting an object that satisfies a predetermined condition from among objects.

Brief Description of the Drawings

[0009]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Mode for Carrying Out the Invention

[0010] <Explanation of the Outline> Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a diagram showing the configuration of an information processing system (hereinafter referred to as "this system") according to an embodiment of the present invention.

[0011] Here, prior to the explanation of FIG. 1, Potato Taste Defect (hereinafter referred to as "PTD") related to this system will be explained. PTD refers to poor-quality coffee beans that pose problems, especially in Rwanda, Burundi, the Republic of the Congo, the Democratic Republic of the Congo, Uganda, etc., and various related issues. Although the exact cause of PTD has not been clearly identified, it is thought to be caused by, for example, stress caused by pests such as stink bugs stimulating the coffee beans or bacteria contained in the saliva of stink bugs, etc. Specifically, coffee beans with the characteristics of PTD (hereinafter referred to as "PTD beans") are known to emit a strong strange smell during roasting and lose the flavor unique to coffee, which is different from normal coffee beans (hereinafter referred to as "normal beans"). Therefore, in the above-mentioned countries, when shipping coffee beans, it is generally practiced to select PTD beans at multiple stages after harvesting coffee cherries (fruits). However, at present, the selection of normal beans and PTD beans is mostly carried out manually (by visual inspection and touch) in many cases, which is a very time-consuming task and it is also difficult to completely separate them. Although this system can be applied to general inspection objects, it is particularly useful for efficiently selecting PTD beans and normal beans.

[0012] Figure 1 is a diagram showing the configuration of this system. As shown in Figure 1, this system is composed of an analyzer 1, a learning device 2, and an inference device 3. The analyzer 1, the learning device 2, and the inference device 3 are interconnected via a predetermined network N such as the Internet. Note that the network N is not an essential component, and for example, NFC (Near Field Communication), Bluetooth (registered trademark), LAN (Local Area Network), etc. may be used.

[0013] Here, the object to be analyzed by the analyzer 1 is, for example, food, fruits, drugs, beverages, seasonings, etc. As described above, it is typically coffee beans. The analyzer 1 irradiates the object to be analyzed, including normal beans or PTD beans, with light of a predetermined wavelength, and determines whether the object to be analyzed is a normal bean or a PTD bean based on the obtained reflection spectrum. Specifically, the analyzer 1 is provided with a spectroscope 11, a light source 12, a control unit 13, and a storage unit 14.

[0014] The spectroscope 11 disperses the reflected light passing through the object to be analyzed and measures the intensity of the spectrum. The spectroscope 11 is composed of a general-purpose spectroscope, a CCD camera, an RGB camera, a hyperspectral camera, a multispectral camera, etc.

[0015] The light source 12 is composed of, for example, a halogen lamp, etc. The light source 12 irradiates, for example, the object to be analyzed with light in a predetermined wavelength range.

[0016] The control unit 13 executes various arithmetic processes on the information regarding the intensity of the spectrum obtained by the spectroscope 11 (hereinafter referred to as "spectrum information"). The control unit 13 is composed of a CPU (Central Processing Unit), etc.

[0017] The storage unit 14 stores various data. The storage unit 14 is composed of a general-purpose HDD (Hard Disk Drive), SSD (Solid State Drive), etc.

[0018] The learning device 2 executes various arithmetic processes for learning based on various data and outputs the results of learning. The learning device 2 is composed of a general-purpose PC (Personal Computer), etc.

[0019] The inference device 3 executes various processes for determining whether the object to be analyzed is a normal bean or a PTD bean based on a program including the learning results output by the learning device 2, and outputs the results. The inference device 3 is composed of a general-purpose PC (Personal Computer), etc.

[0020] <Hardware Configuration> FIG. 2 is a block diagram showing an example of the hardware configuration of the learning device in the information processing system of FIG. 1. The learning device 2 is composed of a general-purpose PC (Personal Computer) or the like. As shown in FIG. 2, the learning device 2 includes a control unit 21, a ROM (Read Only Memory) 22, a RAM (Random Access Memory) 23, a bus 24, an input / output interface 25, an output unit 26, an input unit 27, a storage unit 28, a communication unit 29, and a drive 30.

[0021] The control unit 21 is composed of a microcomputer including a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), and a semiconductor memory, and executes various processes according to a program recorded in the ROM 22 or a program loaded from the storage unit 28 to the RAM 23. In the RAM 23, information and the like necessary for the control unit 21 to execute various processes are also appropriately stored.

[0022] The control unit 21, the ROM 22, and the RAM 23 are interconnected via the bus 24. The input / output interface 25 is also connected to this bus 24. The output unit 26, the input unit 27, the storage unit 28, the communication unit 29, and the drive 30 are connected to the input / output interface 25.

[0023] The output unit 26 is composed of various liquid crystal displays, speakers, etc., and outputs various information as images and sounds.

[0024] The input unit 27 is composed of a keyboard, a mouse, etc., and inputs various information.

[0025] The storage unit 28 is composed of an HDD (Hard Disk Drive), an SSD (Solid State Drive), etc., and stores various data. In this embodiment, for example, various information including various programs and various databases is stored.

[0026] The communication unit 29 controls communication with other information processing devices and the like via a network N including the Internet.

[0027] The drive 30 is provided as needed. A removable medium 41 made of a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is appropriately mounted on the drive 30. The program read from the removable medium 41 by the drive 30 is installed in the storage unit 28 as needed. Also, the removable medium 41 can store various data stored in the storage unit 28 in the same manner as the storage unit 28.

[0028] FIG. 3 is a block diagram showing an example of the hardware configuration of the inference device in the information processing system of FIG. 1. The inference device 3 is composed of a general-purpose PC (Personal Computer) or the like. As shown in FIG. 3, the inference device 3 includes a control unit 51, a ROM (Read Only Memory) 52, a RAM (Random Access Memory) 53, a bus 54, an input / output interface 55, an output unit 56, an input unit 57, a storage unit 58, a communication unit 59, and a drive 60. Note that a removable medium 71 made of a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is appropriately mounted on the drive 60. However, since the hardware configuration of each part of the inference device 3 can be basically the same as the hardware configuration of each part of the learning device 2, the description is omitted here.

[0029] FIG. 4 is a block diagram showing an example of the functional configuration of the analysis device of FIG. 1, the learning device of FIG. 2, and the inference device of FIG. 3.

[0030] As shown in FIG. 4, the control unit 21 of the learning device 2 is provided with a measurement information acquisition unit 80, an arithmetic processing unit 82, and a learning processing unit 84. In addition, a learning result DB 300 is provided in a region of the storage unit 28 of the learning device 2.

[0031] The measurement information acquisition unit 80 acquires information regarding the intensity of reflection or scattering in a predetermined wavelength region for one or more normal objects and defective objects. Specifically, the measurement information acquisition unit 80 acquires information (hereinafter referred to as "measurement information") regarding the measured results of an analysis object for learning (hereinafter referred to as "learning analysis object") transmitted from the analysis device 1 via the communication unit 29. Note that the measurement information is, for example, information regarding the intensity of the spectrum of the reflected light with respect to the light irradiated to the learning analysis object. Note that it is desirable that the measurement information acquired here functions as teacher data for learning. That is, the measurement information may include information as to whether the measurement information is acquired from normal beans or PTD beans.

[0032] The arithmetic processing unit 82 executes various arithmetic processes on the measurement information. Specifically, the arithmetic processing unit 82 is provided with a smoothing processing unit 90, a normalization processing unit 92, a centering processing unit 94, and a wavelength selection unit 96.

[0033] The smoothing processing unit 90 executes a process of smoothing the measurement information at each measurement point. Note that the smoothing process performed by the smoothing processing unit 90 is, for example, for the purpose of removing noise and detecting differences without being caught by local trends. In addition, the arithmetic processing unit 82 generates an average spectrum obtained by smoothing the respective spectral intensities of the measurement information acquired by the measurement information acquisition unit 80.

[0034] The normalization processing unit 92 and the centering processing unit 94 each perform arithmetic processing to normalize and center the average spectrum generated by the smoothing processing unit 90. It is known that by performing such processing, statistical differences in various types of information can be emphasized. Regarding the result of the specific centering processing by the centering processing unit 94 and the like, it will be described later with reference to FIG. 9 and the like.

[0035] The wavelength selection unit 96 selects a wavelength region with a high priority that contributes to the determination of whether the beans are normal or PTD beans using various statistical methods. Specifically, in this embodiment, the wavelength selection unit 96 adopts a method called PLS-β, which is a variable selection method related to the partial least squares (PLS) method, to select a wavelength region with a high priority that contributes to the determination of whether the beans are normal or PTD beans. Here, the PLS method and the PLS-β method will be briefly described. The PLS method is a method for generating a linear regression model using latent variables, similar to the principal component regression method. However, when determining the latent variables, unlike the principal component regression, it is determined such that the covariance between the latent variable and the target variable is maximized. The PLS method is often useful when there is a high correlation between the explanatory variables and is suitable for application to data such as spectral intensities in this system. The PLS-β method is a method for converting the PLS regression model into a linear regression model and selecting important explanatory variables based on the magnitude of the absolute value of the regression coefficient. In this system, the PLS-β method is adopted to select variables (wavelength regions) that contribute significantly to the determination of whether the beans are normal or PTD beans. Regarding the specific method and result of wavelength selection by the wavelength selection unit 96, it will be described later with reference to FIG. 10 and the like.

[0036] The learning processing unit 84 performs learning by applying statistical means to the information acquired by the data acquisition means. Specifically, the learning processing unit 84 executes learning processing on the measurement information in each wavelength region selected by the wavelength selection unit 96 and generates a learning result. The learning processing unit 84 stores the generated learning result in the learning result DB 300 and transmits information regarding the learning result (hereinafter referred to as "learning result information") to the inference device 3. Note that as a learning method, this system can execute learning processing by combining any statistical methods such as neural networks, deep learning, and various regression models. Also, the learning result information may include, for example, a so-called learned model and various information regarding various programs, algorithms, mathematical formulas, etc. used for the determination of normal beans and PTD beans.

[0037] As shown in FIG. 4, the control unit 51 of the inference device 3 is provided with a learning result acquisition unit 120, a target information acquisition unit 122, an arithmetic processing unit 124, a determination unit 126, and an output unit 128. Also, a learning result DB 400 is provided in an area of the storage unit 58 of the inference device 3.

[0038] The learning result acquisition unit 120 applies statistical means to information regarding the intensity of reflection or scattering in a predetermined wavelength region of one or more normal objects and defective objects to acquire the learned result. Specifically, the learning result acquisition unit 120 acquires the learning result information transmitted from the learning device 2 via the communication unit 59. Also, the learning result acquisition unit 120 stores the acquired learning result information in the learning result DB 400.

[0039] The target information acquisition unit 122 acquires information regarding the intensity of reflection or scattering in a predetermined wavelength region of the target object to be sorted. Specifically, the target information acquisition unit 122 acquires information (hereinafter referred to as "target information") regarding the intensity of the spectrum of each reflected light with respect to the analysis target object to be sorted (hereinafter referred to as "measurement analysis target object") transmitted from the analyzer 1 via the communication unit 59. Note that the target information is, for example, information regarding the intensity of the spectrum of the reflected light with respect to the light irradiated on the measurement analysis target object, etc.

[0040] The arithmetic processing unit 124 performs various statistical processes (such as normalization processing and centering processing) on the target information.

[0041] The determination unit 126 applies statistical means to the information regarding the intensity of reflection or scattering in a predetermined wavelength region of one or more normal target objects and defective target objects, and determines whether the target object to be sorted is a normal target object or a defective target object based on the learned result. Specifically, the determination unit 126 determines whether each measurement analysis target object from which the target information is acquired is a normal bean or a PTD bean based on the learning result information acquired by the learning result acquisition unit 120 and the target information acquired by the target information acquisition unit 122.

[0042] The output unit 128 executes various processes for outputting the determination result performed by the determination unit 126.

[0043] FIG. 5 is a diagram showing an example of an image including visual features of typical normal beans and PTD beans. Specifically, as shown in FIG. 5, an image of a typical normal bean and an image of a typical PTD bean are displayed. First, as shown in FIG. 5(A), the normal bean has no particular discoloration or internal damage and can be recognized as a natural bean by people. In contrast, as shown in FIG. 5(B), the PTD bean has holes (caused by the invasion of stink bugs) shown. In addition to the presence of such unnatural holes, the PTD bean may also have green or black discoloration inside the bean, and whether or not there is such a finding can be a criterion for determining whether it is a normal bean or a PTD bean. Also, even if there are holes on the surface of the bean, there are some that have no internal impact. Although the example of FIG. 5 illustrates a very easy-to-understand example, in reality, depending on the size and angle of the holes, etc., it may be difficult to distinguish, it may be difficult to distinguish by image recognition, etc., or it may take time for visual confirmation. Therefore, if such PTD beans can be excluded efficiently and accurately, it is considered that the damage caused by PTD beans can be significantly reduced.

[0044] FIG. 6 is a diagram showing an example of a method by which the analyzer of FIG. 4 acquires various information. As shown in FIG. 6, this system acquires various information (such as the intensity of the reflection spectrum) at a total of six points, three points on the front side and three points on the back side of the coffee bean. Thereby, it is considered that this system can acquire information having a rough spatial resolution and can perform determination between normal beans and PTD beans with higher accuracy.

[0045] Next, with reference to FIGS. 7 to 9, an example of the result of performing various statistical processes on the acquired measurement information will be described. One of the purposes of the statistical processes shown in FIGS. 7 to 9 is to accentuate the statistical differences in the measurement information and confirm that the information acquired by this method reflects the differences in the spectral intensity trends of normal beans and PTD beans. First, FIG. 7 is a diagram showing an example of an average spectrum generated by averaging measurement information. That is, FIG. 7 is a graph of an average spectrum generated by averaging the measurement information measured at the six points shown in FIG. 6.

[0046] In the example of FIG. 7, in normal beans, the graph has a steep rise in the wavelength range generally from 600 nm to 1100 nm, whereas no steep rise was observed in PTD beans.

[0047] Next, FIG. 8 is a graph showing the result of differentiating the average spectrum shown in FIG. 7 after performing processing such as smoothing. In the example of FIG. 8, the trend is more clearly shown. In the first derivative data of normal beans, a decreasing trend is observed in the wavelength range generally from 600 nm to 1100 nm, whereas in the first derivative data of PTD beans, an increasing trend is observed in the wavelength range generally from 600 nm to 1100 nm.

[0048] Similarly, FIG. 9 is a graph showing the result of performing normalization processing and centering processing on the average spectrum shown in FIG. 7 after performing processing such as smoothing. Also in the example of FIG. 9, particularly in the wavelength range of normal beans generally from 600 nm to 1100 nm, the numerical values change in a downward slope to the right, whereas in the same wavelength range of PTD beans, the tendency for the numerical values to decrease to the right is not observed. Rather, a tendency for the values to rise from the lower left to the upper right as a whole is observed.

[0049] FIG. 10 is a diagram showing an example of the result of further performing a predetermined statistical process on the result of the statistical process of FIG. 9. That is, FIG. 10 is a graph showing an example of the result of further applying the PLS-β method to the result of the centering process of FIG. 9.

[0050] In the example of FIG. 10, the random numbers used for data splitting of the training data and the verification data are changed, and the frequency (number of times, vertical axis) of each wavelength region selected from the results of 2500 trials is displayed as a result. Therefore, in the example of FIG. 10, the wavelength regions with a high frequency of wavelengths on the vertical axis can be said to be the wavelength regions with high priority when verifying whether they contribute to the determination of normal beans or PTD beans. Specifically, in the example of FIG. 10, it can be inferred that 370 nm, 410 nm, 450 nm, 490 nm, 520 nm, 550 nm, 565 nm, 580 nm, 600 nm, 620 nm, 635 nm, 660 nm, 675 nm, 730 nm, 760 nm, 800 nm, 875 nm, 935 nm, 960 nm, 995 nm, 1020 nm, 1080 nm, 1090 nm, 1100 nm, 1120 nm, 1140 nm, 1155 nm are the wavelengths with a large contribution. From these data, the inventors of the present application, for example, estimated that the wavelength regions of 550 nm to 580 nm, 600 nm to 680 nm, 715 nm to 745 nm, 785 nm to 815 nm, 935 to 995 nm, 1005 nm to 1035 nm, and 1080 nm to 1120 nm are the wavelength regions with particularly large contributions.

[0051] An example when learning is performed by changing the combination of selected wavelengths is shown in FIG. 11. FIG. 11 is a diagram showing an example of the results of performing various statistical processes and learning on only the spectral intensity at the selected wavelengths from the raw data, and confirming the accuracy, and is a diagram showing an example different from the examples of FIGS. 7 to 10. In the example of FIG. 11, information such as the number of wavelengths selected as a result of each process, the length of the selected wavelengths, the correct answer rate for normal beans, the correct answer rate for abnormal beans (PTD beans), and the correct answer rate for the whole (normal beans and PTD beans) is summarized as a list. The inventors of the present application further conducted studies including many other samples and combinations of statistical processes, and in particular, by selecting four wavelength regions of 550 nm to 580 nm, 600 nm to 680 nm, 715 nm to 745 nm, and 785 nm to 815 nm, it was clarified that the determination of normal beans and PTD beans can be stably and highly accurately performed. Also, for example, when two wavelengths of 875 nm and 935 nm are selected, it was clarified that the determination of normal beans and PTD beans can be performed with sufficient accuracy.

[0052] Here, the inventors of the present invention were able to know from the observations during the measurement experiment that PTD beans have a part that is no different from normal beans and a part that has the spectral characteristics peculiar to PTD beans. In this regard, this system can perform learning processing and wavelength region selection using the result of arithmetic processing in which the influence of the spectral characteristics peculiar to PTD beans remains even when a process that reduces the spatial resolution such as averaging processing is performed. Then, this system performs learning processing using only the data in the corresponding wavelength region of data with an arbitrarily high spatial resolution to examine whether the specified predetermined wavelength region identified from the result is effective, examines the accuracy of determining whether it is a normal bean or a PTD bean, and examines the conditions under which high-accuracy determination can be realized. By such a method, this system can reduce the load of learning processing and wavelength region selection, and can widely secure options for arithmetic processing to be adopted in combination.

[0053] FIG. 12 is a flowchart for explaining the flow of learning processing executed by the learning device of FIG. 4.

[0054] In step S1, the measurement information acquisition unit 80 of the learning device 2 acquires a total of six points of measurement information including the outside and inside of the coffee bean transmitted from the analysis device 1 via the communication unit 29.

[0055] In step S2, the smoothing processing unit 90 of the learning device 2 generates an average spectrum obtained by smoothing the respective spectral intensities of the measurement information acquired by the measurement information acquisition unit 80.

[0056] In step S3, the normalization processing unit 92 and the centering processing unit 94 each execute arithmetic processing to normalize and center the average spectrum generated by the smoothing processing unit 90.

[0057] In step S4, the wavelength selection unit 96 of the learning device 2 applies various statistical methods (e.g., PLS-β) to the result of the arithmetic processing executed by the centering processing unit 94 to select a wavelength region with a high priority that contributes to the determination of whether the beans are normal or PTD beans.

[0058] In step S5, the learning processing unit 84 of the learning device 2 executes a learning process on the measurement information in each set of wavelength regions selected by the wavelength selection unit 96 and generates a learning result. Here, the measurement information used by the learning processing unit 84 for learning may be the acquired measurement information as it is, or the average spectrum based on the measurement information. In this embodiment, it is assumed that the learning processing unit 84 executes the learning process using the non-averaged measurement information at the selected wavelengths determined based on the averaged measurement information.

[0059] In step S6, the learning processing unit 84 of the learning device 2 transmits learning result information regarding the generated learning result to the inference device 3. Thereby, the learning process executed by the learning device 2 ends.

[0060] Subsequently, the flow of the inference process executed by the inference device 3 will be described. FIG. 13 is a flowchart for explaining the flow of the inference process executed by the inference device of FIG. 4.

[0061] In step S21, the learning result acquisition unit 120 of the inference device 3 acquires the learning result information transmitted from the learning device 2 via the communication unit 59.

[0062] In step S22, the target information acquisition unit 122 of the inference device 3 acquires, via the communication unit 59, the target information of the selected wavelengths in each measurement analysis target object transmitted from the analysis device 1.

[0063] In step S23, the arithmetic processing unit 124 executes various statistical processes (such as normalization processing and centering processing) on the target information.

[0064] In step S24, the determination unit 126 of the inference device 3 determines whether each measurement analysis target object from which the target information is obtained is a normal bean or a PTD bean based on the learning result information obtained by the learning result acquisition unit 120 and the target information obtained by the target information acquisition unit 122.

[0065] In step S25, the output unit 128 of the inference device 3 outputs the determination result made by the determination unit 126. Thereby, the inference process executed by the inference device 3 ends.

[0066] Although one embodiment of the present invention has been described above, the present invention is not limited to the above-described embodiment, and modifications, improvements, etc. within the range capable of achieving the object of the present invention are included in the present invention.

[0067] Here, an example of an application example of the present system will be shown with reference to FIG. 14. FIG. 14 is an example of a plot diagram showing the results of performing various statistical preprocessings.

[0068] In the example of FIG. 14, a graph in which each spectrum subjected to arithmetic processing is projected is displayed on a two-dimensional space with important latent variables as axes. For example, in the example of FIG. 14, a group of normal beans is relatively largely displayed in the upper left direction of the figure, and a group of PTD beans is relatively largely displayed in the lower right direction of the figure. There is also a possibility that the present system can determine whether the measurement analysis target object is a normal bean or a PTD bean by performing various preprocessings on the raw data in this way and using the results as they are.

[0069] Also, in the above-described embodiment, the present system has been described as obtaining various information at a total of six points on the front and back sides of the coffee bean, but it is not limited thereto. The location and number where the present system obtains various information are arbitrary and not limited. Furthermore, the system may adopt methods with higher spatial resolution, such as various cameras and imaging technologies, to obtain various information (such as the intensity of the reflection spectrum).

[0070] Also, in the above-described embodiment, the system has been described as selecting the wavelength region using the PLS-β method, but it is not limited thereto. The system may adopt various statistical methods for variable selection, such as, for example, P LS -VIP (Variable Importance in Projection), LASSO (Least Absolute Shrinkage and Selection Operator), Stepwise, GA-PLS (Genetic Algorithm-PLS), etc.

[0071] Also, in the above-described embodiment, the system has been described as being able to stably and highly accurately determine normal beans and PTD beans by selecting four wavelength regions of 550 nm to 580 nm, 615 nm to 645 nm, 715 nm to 745 nm, and 785 nm to 815 nm. However, these selected wavelengths are merely examples. The system may, for example, select wavelength regions other than the above-described ones as the selected wavelengths, or may select only one or a plurality of the four wavelength regions of 550 nm to 580 nm, 615 nm to 645 nm, 715 nm to 745 nm, and 785 nm to 815 nm as the selected wavelengths.

[0072] Also, although not described in the above-described embodiment, the system may, for example, additionally obtain or label information such as spectral information, image region selection, and bean quality (normal or PTD bean) moisture content for the image captured by the camera. In this case, the system may, for example, obtain these information as part of the measurement information.

[0073] For example, the above system configuration is just an example and is not limiting. In particular, the analyzer 1, the learning device 2, and the inference device 3 do not necessarily have to function as individual hardware. For example, the functions of the analyzer 1 and the learning device 2 may be provided integrally, or the functions of the analyzer 1 and the inference device 3 may be provided integrally, or the functions of the learning device 2 and the inference device 3 may be provided integrally.

[0074] For another example, the above series of processes can be executed by hardware or by software. That is, the functional configuration in FIG. 4 is merely illustrative and not limiting. It is sufficient that the information processing system is equipped with a function capable of executing the above series of processes as a whole, and there are no particular limitations on what functional blocks are used to realize this function. Also, the location of the functional blocks is not limited to the example in FIG. 4 and can be arbitrary. Furthermore, one functional block may be constituted by hardware alone, software alone, or a combination thereof.

[0075] When the series of processes are executed by software, the program constituting the software is installed in a computer or the like from a network or a recording medium. The computer or the like may be a computer incorporated in dedicated hardware. Also, the computer or the like may be a computer capable of executing various functions by installing various programs, such as a general-purpose smartphone or a general-purpose PC (Personal Computer) in addition to a server.

[0076] A recording medium containing such a program may be constituted not only by a removable medium (not shown) distributed separately from the apparatus main body to provide the program to the user, but also by a recording medium or the like provided to the user in a state pre-installed in the apparatus main body.

[0077] In addition, in this specification, the step of describing a program recorded on a recording medium includes not only processes that are performed in chronological order according to that order, but also processes that are executed in parallel or individually, even if they are not necessarily processed in chronological order. That is, some of the steps in the steps of FIGS. 12 and 13 may be appropriately changed or omitted.

[0078] Also, for example, in this specification, the term "system" shall mean an overall device composed of a plurality of devices, a plurality of means, etc.

[0079] In other words, the information processing apparatus to which the present invention is applied only needs to have the following configuration, and can take various embodiments. That is, the information processing apparatus to which the present invention is applied is an information processing apparatus used for sorting objects, an acquisition means (for example, the target information acquisition unit 122) for acquiring information regarding the intensity of light in a predetermined wavelength region related to the object to be sorted; a determination means (for example, the determination unit 126) for applying statistical means to information regarding the intensity of reflection or scattering in a predetermined wavelength region in one or more normal objects and defective objects, and determining the nature or quality of the object to be sorted based on the learned result; an output means (for example, the output unit 128) for outputting the result of the determination by the determination means; is sufficient as an information processing apparatus provided with the above.

[0080] Further, the determination means may apply statistical means to information regarding the intensity of reflection or scattering in a wavelength region from 310 nm to 1180 nm, and determine whether the object to be sorted is a normal object or a defective object based on the learned result.

[0081] Also, the wavelength region may be calculated by executing a statistical method on information regarding the intensity of reflection or scattering in two or more regions of the object to be sorted.

[0082] Further, the statistical method may include a process of reducing the spatial resolution with respect to information on the intensity of reflection or scattering in two or more regions of the object to be sorted.

[0083] The determination means applies statistical means to information on the intensity of reflection or scattering in a wavelength region including at least one of 550 nm to 580 nm, 600 nm to 680 nm, 715 nm to 745 nm, 785 nm to 815 nm, 935 to 995 nm, 1005 nm to 1035 nm, and 1080 nm to 1120 nm, and determines whether the object to be sorted is a normal object or a defective object based on the learned result.

[0084] Further, as a second aspect of the present invention, an information processing apparatus to which the present invention is applied is An information processing apparatus for sorting an object, Data acquisition means (for example, measurement information acquisition unit 80) for acquiring information on the intensity of reflection or scattering in a predetermined wavelength region of one or more normal objects and defective objects, Learning means (for example, learning processing unit 84) for performing learning by adapting statistical means to the information acquired by the data acquisition means, An information processing apparatus provided with the above is sufficient.

[0085] Further, as a third aspect of the present invention, an information processing method to which the present invention is applied is An information processing method executed by a computer used for sorting an object, Acquisition means for acquiring information on the intensity of light in a predetermined wavelength region regarding the object to be sorted, Determination means for applying statistical means to information on the intensity of reflection or scattering in a predetermined wavelength region of one or more normal objects and defective objects, and determining the nature or quality of the object to be sorted based on the learned result, Output means for outputting the result of determination by the determination means, An information processing method including the above is sufficient.

[0086] Also, as a fourth aspect of the present invention, the program to which the present invention is applied is configured to cause a computer used for sorting objects to acquire information regarding the intensity of light in a predetermined wavelength region related to the object to be sorted, apply statistical means to information regarding the intensity of reflection or scattering in a predetermined wavelength region in one or more normal objects and defective objects, and determine the property or quality of the object to be sorted based on the learned result, and output the result of the determination by the determination means, and a program that executes a control process including is sufficient.

Explanation of Signs

[0087] 1... Analyzer 2... Learning device 21... Control unit 80... Measurement information acquisition unit 82... Arithmetic processing unit 90... Smoothing processing unit 92... Normalization processing unit 94... Centering processing unit 96... Wavelength selection unit 84... Learning processing unit 300... Learning result DB 3... Inference device 51... Control unit 120... Learning result acquisition unit 122... Target information acquisition unit 124... Arithmetic processing unit 126... Determination unit 128... Output unit 400... Learning result DB

Claims

1. An information processing apparatus used for sorting objects, an acquisition means for acquiring information regarding the intensity of light in a predetermined wavelength region related to the object to be sorted, a determination means for applying statistical means to learning information regarding the intensity of reflection or scattering in a predetermined wavelength region in one or more normal objects and defective objects, and determining the nature or quality of the object to be sorted based on the learned result, an output means for outputting the result of the determination by the determination means, comprising: the determination means generates an average spectrum obtained by smoothing each spectral intensity of the learning information with respect to the learning information acquired at a location including at least the front side and the back side of the coffee bean as the object, and applies a predetermined statistical method to the result of the arithmetic processing obtained by normalizing the average spectrum, thereby selecting a wavelength region with a high priority that contributes to the determination of whether it is a normal bean or a PTD bean, and determines whether the coffee bean as the object is a normal bean or a PTD bean based on an algorithm generated by performing a learning process on the learning information in the selected set of wavelength regions, information processing apparatus.

2. The selection of the wavelength region with a high priority that contributes to the determination of whether it is a normal bean or a PTD bean is performed based on the difference in spectral intensity in the range from 600 nm to 1100 nm, The information processing apparatus according to Claim 1.

3. the determination means applies statistical means to the learning information regarding the intensity of reflection or scattering in the wavelength region from 310 nm to 1180 nm, and determines whether the object to be sorted is a normal object or a defective object based on the learned result, The information processing apparatus according to Claim 1.

4. The wavelength region is calculated by performing a statistical method on the learning information regarding the intensity of reflection or scattering in two or more regions of the object to be sorted, The information processing apparatus according to Claim 3.

5. The statistical method includes a process of reducing the spatial resolution with respect to the learning information regarding the intensity of reflection or scattering in two or more regions of the object to be sorted, The information processing apparatus according to Claim 4.

6. An information processing apparatus for sorting objects, a data acquisition means for acquiring learning information regarding the intensity of reflection or scattering in a predetermined wavelength region in one or more normal objects and defective objects, Learning means for performing learning by applying statistical means to the learning information acquired by the data acquisition means; comprising; The learning means generates an average spectrum obtained by smoothing the respective spectral intensities of the learning information acquired at a location including at least the front and back sides of the coffee beans of the object, and applies a predetermined statistical method to the result of the arithmetic processing of normalizing the average spectrum, thereby selecting a wavelength region with a high priority for determining whether the coffee beans are normal beans or PTD beans, and performs a learning process on the learning information in the selected set of wavelength regions. An information processing apparatus.

7. An information processing method executed by a computer used for sorting an object, an acquisition step of acquiring information regarding the intensity of light in a predetermined wavelength region regarding the object to be sorted; a determination step of applying statistical means to learning information regarding the intensity of reflection or scattering in a predetermined wavelength region in one or more normal objects and defective objects, and determining the nature or quality of the object to be sorted based on the learned result; an output step of outputting the result of the determination in the determination step; including; In the determination step, for the learning information acquired at a location including at least the front and back sides of the coffee beans of the object, an average spectrum obtained by smoothing the respective spectral intensities of the learning information is generated, and a predetermined statistical method is applied to the result of the arithmetic processing of normalizing the average spectrum, thereby selecting a wavelength region with a high priority for determining whether the coffee beans are normal beans or PTD beans, and based on the algorithm generated by performing a learning process on the learning information in the selected set of wavelength regions, determining whether the coffee beans of the object are normal beans or PTD beans. An information processing method.

8. Causing a computer used for sorting an object to an acquisition means for acquiring information regarding the intensity of light in a predetermined wavelength region regarding the object to be sorted; a determination means for applying statistical means to learning information regarding the intensity of reflection or scattering in a predetermined wavelength region in one or more normal objects and defective objects, and determining the nature or quality of the object to be sorted based on the learned result; an output means for outputting the result of the determination by the determination means; execute a control process including; The determination means generates an average spectrum obtained by smoothing the respective spectral intensities of the learning information acquired at a location including at least the front side and the back side of the coffee beans of the object, and applies a predetermined statistical method to the result of the arithmetic processing of normalizing the average spectrum, thereby selecting a wavelength region with a high priority that contributes to the determination of whether the coffee beans are normal beans or PTD beans, and based on the algorithm generated by performing a learning process on the learning information in the selected set of wavelength regions, determines whether the coffee beans of the object are normal beans or PTD beans. Program.

Images