Counting method, counting system and counting program

Abstract

To provide technology for accurately counting an object included in a laminate.SOLUTION: A counting method acquires image data on a lamination surface of a laminate, partitions the image data into a plurality of predetermined regions with respect to a lamination direction and a width direction, generates a signal profile indicating each thickness component in the predetermined region, acquires a thickness component estimated from each predetermined region, and counts an object, on the basis of the thickness component.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a method, a system, and a program for counting objects contained in a stack. [Background technology]

[0002] Plate-like or sheet-like articles are stored in a stacked state. Conventionally, there has been known a technique for counting such articles by capturing an image of the articles in a stacked state.

[0003] Patent Document 1 discloses a technique capable of simply and accurately counting the number of layers of a layered pattern on various counting objects having a layered pattern surface. Patent Document 1 discloses a counting method including a step of acquiring a captured image 11 which is an image obtained by capturing the surface of a counting object 10 having a layered pattern surface, a step of extracting a counting target area 11A which is a partial area of ​​the captured image 11, a step of identifying a base point within the counting target area 11A, a step of determining an analysis area 15 based on the base point, and a step of counting the number of layers of the layered pattern based on the analysis area 15.

[0004] Patent Document 2 discloses a technique that enables easy and highly accurate counting of substrates. Patent Document 2 discloses a substrate counting method that includes the steps of acquiring an image of the end faces of a plurality of substrates stacked in one direction, integrating the brightness of each pixel constituting the image for each pixel row in the other direction perpendicular to the one direction in the image to calculate an integrated value, detecting substrates in the image based on the integrated value, and counting the number of detected substrates to obtain a count value. Patent Document 2 describes that the brightness of each pixel is integrated for each pixel row in the horizontal direction (X'-axis direction) of the image to obtain an integrated value (horizontal projection value), and furthermore, that the integrated value is obtained as an edge projection value, thereby calculating the thickness of the substrate, the boundary position of the substrate, and the center position of the substrate.

[0005] Patent Document 3 discloses a technique for measuring the number of stacked plywood sheets, which can automate the work of measuring the number of stacked plywood sheets and improve the measurement accuracy of the number of stacked plywood sheets more easily and economically than before. Patent Document 3 discloses a device 10 for measuring the number of stacked plywood sheets and a method thereof, which can measure the number of stacked plywood sheets in a state where a plurality of plywood sheets 11 are stacked, each plywood sheet being made by bonding together a plurality of types of thin plate materials 12, 13 having different colors, and a method for measuring the number of stacked plywood sheets, which discloses imaging means 15 to image the side of plywood 11, and information processing means 17 to use the brightness value obtained from the image of stacked plywood 11 imaged by imaging means 15 to determine the type of thin plate materials 12, 13 constituting each plywood 11 and calculate the number of stacked plywood sheets 11. Patent document 3 discloses that image data of the surface of the laminate is processed by performing various image processing operations to emphasize the boundaries of the laminate, and then calculating the distance between the plate-like bodies and the reference pitch, and that the reference pitch is a value obtained by adding a predetermined tolerance to the thickness of the substrate. It also discloses that a candidate point P is assigned to identify each plate-like body by referring to the calculated reference pitch, and the candidate point P that is determined to be the starting position is designated as a fixed point P', and the number of consecutive candidate points are processed as fixed points to count the number of plate-like bodies.

[0006] Patent Document 4 discloses a method in which multiple folded sheets of paper 1 are aligned, light is shone on them to produce bright areas 5 on the spine 2 and shadow areas 6 in the recesses between the spines, the images are captured by a television camera 7, and the resulting image signals are converted into multi-valued color intensity signals for each color by an analog-to-digital conversion circuit 9. The color intensity signals of each component point constituting one horizontal scanning line are compared with the color intensity of each component point against two calculated reference curves, and when a component point exceeds the lower reference value, exceeds the upper reference value upward without exceeding the lower reference value downward, this is counted as one non-shadow area, and the comparison is repeated up to the other end to count the number of non-shadow areas, and the non-shadow areas are counted as the spine portion of the paper, i.e., the number of sheets of paper.

[0007] In Patent Document 5, a reference pitch is calculated based on the thickness of the glass plate G, candidate points equal to the reference pitch are calculated as fixed points, and all the calculated fixed points are counted. In other words, a plate counting device for a plate stack is disclosed that recognizes the presence of glass plates G in an image received by a line sensor 48 and counts the number of glass plates G.

[0008] Patent document 6 discloses a stack number counting device in which a two-dimensional image sensor camera 12 outputs an image of the side of a stack of compact discs 2 as two-dimensional image data, and an image processing device 14 calculates the number of compact discs from this two-dimensional image data. [Prior art documents] [Patent documents]

[0009] [Patent Document 1] Patent Publication No. 2021-071921 [Patent Document 2] JP 2015-228094 A [Patent Document 3] JP 2018-036950 A [Patent Document 4] JP 2014-002701 A [Patent Document 5] JP 2014-032431 A [Patent Document 6] Japanese Patent Application Publication No. 05-197851 Summary of the Invention [Problem to be solved by the invention]

[0010] There was room for improvement in the accuracy of counting the number of plate- or sheet-like objects contained in a stack using image recognition. In particular, objects made of materials such as steel plates were subject to in-plane (width direction) bending, which created gaps between the objects, making it difficult to count them accurately.

[0011] In view of the above-mentioned problems, an object of the present invention is to provide a technique for counting objects contained in a stack with high accuracy. [Means for solving the problem]

[0012] In order to solve the above-mentioned problems, the present invention provides a method for counting objects contained in a laminate, the method including an image acquisition process, a generation process, a thickness estimation process, and a counting process, the image acquisition process acquiring image data of the stacking surface of the laminate, the generation process dividing the image data into a plurality of predetermined regions in the stacking direction and width direction and generating a signal profile indicating each thickness component in the predetermined regions, the thickness estimation process acquiring thickness components estimated from each of the predetermined regions, and the counting process counting the objects based on the thickness components.

[0013] With this configuration, by counting based on thickness components estimated from specific regions in the stacking direction and width direction of the stacking surface, it is possible to achieve accurate counting of objects corresponding to thicknesses that vary from region to region.

[0014] In a preferred form of the present invention, the thickness estimation process generates a classification result of signal peaks extracted based on the thickness components contained in the signal profile, and the counting process counts the objects based on the thickness components corresponding to each of the signal peaks of the classification results. By using this configuration, the signal peaks can be classified so as to increase the probability that the signal peaks are true thickness components, and accurate counting can be performed using the thickness components corresponding to the signal peaks resulting from the optimal classification.

[0015] In a preferred embodiment of the present invention, the thickness estimation process determines a signal peak which is a valid value or an intensity peak which is an invalid value included in a signal profile in the classification result according to the intensity of a thickness component equal to or greater than a predetermined value, and if a predetermined number or more of boundary lines of the specified region are detected based on the thickness components in the classification result, the invalid values ​​are determined to be valid values, and the counting process counts the objects based on the thickness components indicated by the signal profile determined to be valid values. With this configuration, by adopting an appropriate effective value included in the classification result, it is possible to perform counting with high accuracy using the optimal thickness component.

[0016] In a preferred embodiment of the present invention, the method includes a model estimation step of calculating model parameters from input values ​​including thickness components of each specified region and estimating a counting model having the parameters, and the counting step counts the objects based on the thickness components estimated by the counting model. With this configuration, by using a counting model that reflects each thickness component of a predetermined region, it is possible to realize counting that refers to the optimal thickness component.

[0017] In a preferred embodiment of the present invention, the method includes a model estimation step of calculating model parameters based on input values ​​including a thickness component of each specified region, estimating a plurality of counting models having the parameters, each corresponding to each classification result, and determining an optimal counting model from the plurality of counting models corresponding to the classification result based on a difference value between the real value of the thickness component of each specified region and the calculated value of the thickness component by the counting model, and the counting step counts the object based on the thickness component estimated by the counting model. With this configuration, an optimal counting model can be adopted based on the difference between the calculated value of the thickness component and the actual value, thereby achieving more accurate counting. Effect of the Invention

[0018] According to the present invention, it is possible to provide a technique for counting objects contained in a stack with high accuracy. [Brief description of the drawings]

[0019] [Figure 1] FIG. 1 is a block diagram of a system according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a diagram showing the hardware configuration of the present embodiment. [Diagram 3] 3 is an overall process flowchart of the counting device of the present embodiment. [Figure 4] 13 is an image of signal profile generation in this embodiment. [Diagram 5] 4 is a process flowchart of a generation process according to the present embodiment. [Figure 6] FIG. 4 is a schematic diagram of a signal profile according to the present embodiment. [Figure 7] 4 is a process flowchart of a thickness estimation process according to the embodiment. [Figure 8] FIG. 4 is a schematic diagram of signal peak classification according to the present embodiment. [Figure 9] 4 is a processing flowchart of a model estimation step according to the present embodiment. [Figure 10] 13 is an image of a quadratic surface in this embodiment. [Figure 11] 1 is a block diagram of a system according to different embodiments; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0020] Hereinafter, a counting system and a counting method according to an embodiment of the present invention will be described with reference to the drawings. Note that the embodiment shown below is an example of the present invention, and the present invention is not limited to the embodiment shown below, and various configurations can be adopted.

[0021] In this embodiment, the configuration, operation, etc. of the counting system and the counting device are described, but a similarly configured counting method, a computer program, and a program recording medium on which the program is recorded also provide the same effects. By using a program recording medium, for example, the program can be installed in a computer. A series of processes according to this embodiment described below is provided as a computer-executable program, and can be provided via a non-transitory computer-readable recording medium such as a CD-ROM or a flexible disk, or even via a communication line.

[0022] The counting system is composed of a computer device. The computer device has an arithmetic device such as a CPU (Central Processing Unit) and a storage device. The computer device can function as a counting device by executing a counting program stored in the storage device with the arithmetic device. The counting method is realized by processing of the computer device including the counting device.

[0023] In this embodiment, the objects to be counted are plate- or sheet-shaped objects that can be stored in a stacked state. The material of the objects is not particularly limited, but the present invention can achieve accurate counting even for objects that are prone to in-plane bending, such as steel plates.

[0024] In this embodiment, the stack includes the objects to be counted. The stack refers to a plurality of objects that are stacked together. The objects are stacked in one direction to form the stack. The number of layers of the objects can be observed on the side of the stack (hereinafter referred to as the stacking surface), and the number of objects can be counted based on the number of layers on the stacking surface. The stacking direction is not limited and may be any direction, such as vertical or horizontal.

[0025] In this embodiment, the objects are counted based on image data obtained by capturing an image of the stacking surface. The image data used for counting is preferably data obtained by capturing an image of the entire stack at least in the stacking direction of the stacking surface. However, the image data used for counting is not limited to data obtained by capturing an image of a portion of the stack in the stacking direction. In addition, the image data used for counting may be data that allows the entire image of the stack to be grasped by combining a plurality of partial data, for example. The thickness of the objects is not particularly limited as long as it is a thickness that allows each object to be distinguished by the pixels of the image data. Therefore, the thickness of the objects that can be counted depends on the imaging method (distance, etc.) of the image data and the resolution of the imaging device.

[0026] 1 shows a block diagram of a counting system 1. The counting system 1 includes a counting device 2 and an imaging device 3. The counting device 2 is configured to be able to acquire image data captured by the imaging device 3 through wired or wireless communication. Note that the counting device 2 may have the imaging device 3 built-in and acquire the image data.

[0027] The counting device 2 includes, as functional components, an image acquisition unit 21 that acquires image data, a generation unit 22 that generates a signal profile based on the image data, a thickness estimation unit 23 that acquires thickness components estimated based on the signal profile, a model estimation unit 24 that estimates a counting model based on the thickness components, a counting unit 25 that counts objects based on the thickness components, and an output unit 26 that outputs the counting results. The counting device 2 also includes a storage unit DB as a database. The storage unit DB is provided inside or outside the counting device 2, and is not limited to the configuration of the illustrated example as long as it is capable of communicating various data with the counting device 2.

[0028] The control unit 27 is connected to each component and controls each component by issuing control instructions. The control unit 27 is also connected to the storage unit DB and is configured to be able to store and read various data. In the following description, the intervention of the control unit 27 will be omitted unless the operation of each functional component in the counting device 2 becomes unclear.

[0029] The counting device 2 is configured to be able to accept operation input by a user. In response to the operation input, the counting device 2 controls the image acquisition unit 21 via the control unit 27 to realize a process of acquiring image data from the imaging device 3, a process of selecting image data to be counted, and the like.

[0030] The imaging device 3 is a camera or a terminal device equipped with a camera function, and can capture image data. The imaging device 3 can transmit the image data to the counting device 2 by wired communication, wireless communication, or the like. The image data captured by the imaging device 3 may be acquired by the counting device 2 via an Internet communication network, or the like. Although only one imaging device 3 is shown in FIG. 1, a plurality of imaging devices 3 may be present.

[0031] 2 shows a hardware configuration diagram of the counting device 2. The counting device 2 includes, as its hardware configuration, a control device 201, a storage device 202, a communication device 203, an input device 204, and an output device 205. In this embodiment, the counting device 2 may use a computer device such as a server device or a personal computer. Note that the counting device 2 is not limited to the configuration shown in FIG. 2 as long as it is made up of multiple computer devices and can realize the above-mentioned functional components (21-27) as a whole.

[0032] The control device 201 is configured with one or more processors such as a CPU, and executes a counting program, an OS (Operating System), and other applications to control the overall processing in the counting device 2. The storage device 202 is a hard disk drive (HDD), a solid state drive (SSD), a flash memory, a RAM (Random Access Memory), or the like, and stores the counting program and various data. The communication device 203 is a communication interface for wired communication, wireless communication, etc., and controls data communication with external devices including the imaging device 3. The communication device 203 is also capable of performing data communication with external devices by controlling communication with a communication network NW.

[0033] The input device 204 is an input interface that accepts operational input by a user, and is composed of at least one of a touch panel, a mouse, a keyboard, etc. The output device 205 is composed of at least one of a display that outputs the processing result by the control device 201, a speaker, etc.

[0034] FIG. 3 is a flowchart showing the overall processing performed by the counting device 2.

[0035] The image acquisition unit 21 acquires image data of the stacking surface of the stacked body imaged by the imaging device 3 (S11, image acquisition step). The image acquisition unit 21 may detect the stacking surface included in the image data as a target area, and further acquire image data related to the target area. Here, the image acquisition unit 21 detects the area of ​​the stacking surface used for counting and other areas not used for counting by image recognition, and sets the area of ​​the stacking surface used for counting as the target area. A machine-learned model such as a neural network can be used for the image recognition. Note that there is no limitation on the type of model. The image acquisition unit 21 can input the image data to the machine-learned model and acquire image data related to the target area as an output.

[0036] The image acquisition unit 21 may execute image correction processing to make the image data easier to process in the subsequent steps. The image correction processing includes keystone correction, etc., and includes processing to correct a layered surface included in the image data as a trapezoidal region into a rectangular region when the layered surface is a trapezoidal region.

[0037] The generating unit 22 divides the acquired image data into a plurality of predetermined regions in the stacking direction and the width direction (S12, generating step). The plurality of predetermined regions are set by an arbitrary number of divisions in each of the vertical and horizontal directions (stacking direction, width direction). For example, when the image data is divided into 2 vertically and 1 horizontally, the number of divisions is 2x1, and the number of division images after division is 2. The number of division images is not limited as long as it is multiple, and for example, 1x2, 2x2, 4x4, 6x4, etc. can be set. Here, one image divided from the image data is defined as a division image (predetermined region) in order to distinguish it from the image data. It is preferable that the number of divisions is multiple (2 or more vertically and 2 or more horizontally) in each of the stacking direction and the width direction.

[0038] The generating unit 22 generates a signal profile of each thickness component of the section image (S12, generating step). The generating step S12 will be described in detail later with reference to FIG.

[0039] In this embodiment, the signal profile is obtained by performing a two-dimensional Fourier transform on the grayscale data of the section image, with the horizontal axis indicating frequency (1 / z) and the vertical axis indicating the intensity (I) of the thickness component. The frequency indicates the inverse of the thickness component (z). The thickness component (z) indicates the thickness value per object. The intensity (I) of the thickness component in the signal profile is an intensity that indicates the likelihood that it is that thickness component in the section image.

[0040] Fig. 4(a) is a reference diagram of the image data. In Fig. 4(a), the width direction of the laminate is x and the stacking direction is y. The positions in the width direction and stacking direction in the image data are defined as image components (x, y).

[0041] Fig. 4(b) is a reference diagram of a partition image in which image data is partitioned into predetermined regions. In Fig. 4(b), the number of partitions is 4x4, and the number of partition images is 16. Here, the partition images are each area partitioned by dashed lines. Note that the partitions by dashed lines are added for the purpose of explanation, and dashed lines do not exist in the actual partition image.

[0042] FIG. 4(c) shows the grayscale data of one of the partition images in FIG. 4(b). The grayscale data shows the grayscale of the Y coordinate image at the center of the X coordinate of the partition image. In FIG. 4(c), the vertical axis shows the Y coordinate, and the horizontal axis shows the grayscale value. The grayscale data is obtained as 16 pieces of data corresponding to the 16 partition images in FIG. 4(b).

[0043] FIG. 4(d) is a reference diagram of a signal profile of a thickness component of a partition image. In this embodiment, the signal profile is obtained by performing a two-dimensional Fourier transform on the grayscale data. Here, 16 signal profiles corresponding to the 16 grayscale data in FIG. 4(c) are obtained. The signal profile is converted into data in which the horizontal axis is frequency (1 / z) and the vertical axis is the intensity (I) of the thickness component. In this way, the thickness component in each predetermined region is obtained based on the signal profile. In an ideal predetermined region without inner surface deflection or image distortion of the object, one thickness component is obtained as one intensity peak, but in reality, multiple thickness components (z) are often obtained as intensity peaks corresponding to each thickness component due to inner surface deflection or image distortion. Since the multiple thickness components (z) included in the signal profile are indiscriminate and uncertain, it is preferable to narrow down the candidates suitable as the true thickness component (z) to one or more.

[0044] The thickness estimation unit 23 acquires thickness components estimated from each of the predetermined regions (S13, thickness estimation step). The thickness estimation unit 23 classifies the signal peaks extracted based on the thickness components of each of the predetermined regions indicated by the signal profile, and generates one or more classification results. The thickness estimation unit 23 can output estimates of multiple thickness components as classification results of the signal peaks of the signal profile. The thickness estimation step S13 will be described in detail later with reference to FIG. 7.

[0045] The model estimation unit 24 calculates model parameters using input values ​​including thickness components of each predetermined region, and estimates a counting model having the parameters (S14, model estimation step). In this embodiment, the counting model can output a thickness component (z) for image components (x, y). In this embodiment, the model estimation unit 24 can estimate one or more counting models corresponding to each classification result generated by the thickness estimation unit 23. When there are multiple estimated counting models, the model estimation unit 24 determines an optimal counting model from the multiple estimated counting models. Note that when there is only one estimated counting model, the model estimation unit 24 can adopt that counting model. The model estimation step S14 will be described in detail later with reference to FIG. 9.

[0046] The counting unit 25 counts the objects based on the thickness component of each predetermined region (S15, counting step). The counting unit 25 counts the objects for each predetermined region, for example, by detecting the boundary line of the objects based on the thickness component of each predetermined region. The counting unit 25 can calculate the total number of objects by adding up the counting results of the predetermined regions in the stacking direction. Note that at this time, the model estimation step S14 may be omitted, and the objects are counted based on the thickness component of each predetermined region estimated by the thickness estimation step S13. Furthermore, when the model estimation step S14 is omitted, the model estimation unit 24 becomes unnecessary.

[0047] The counting unit 25 can count the objects based on the thickness component estimated by the counting model (S15, counting step). The counting model indicates an index of the thickness component that can change corresponding to the coordinates (x, y) in the image data of the object. The counting unit 25 can calculate the total number of objects by referring to the thickness component (z) derived by the counting model for each coordinate (x, y) in the image data of the object and identifying the boundary of the object in the thickness direction (stacking direction).

[0048] The output unit 26 outputs based on the counting result of the counting unit 25 regarding the total number of objects (S16, output step). The output unit 26 outputs the counting result as a display on the output device 205 or as a sound generated from a speaker or the like.

[0049] FIG. 5 is a process flowchart of the generation process in the generation step S12.

[0050] The generating unit 22 divides the image data into a plurality of predetermined regions in the stacking direction and width direction (S21).

[0051] The generating unit 22 executes image processing to binarize each of the partitioned predetermined regions (S22). Here, the image processing is not limited to a specific processing mode as long as it is processing for making the data suitable for the conversion processing described later.

[0052] The generating unit 22 converts each of the predetermined regions into a signal profile indicating frequency (S23). In this embodiment, the generating unit 22 converts the image components (x, y) into a frequency (1 / z) that is the inverse of the thickness component (z) by performing a two-dimensional Fourier transform process on the predetermined regions. Through the frequency conversion, the signal profile becomes data with the frequency (1 / z) on the horizontal axis and the intensity (I) of the thickness component on the vertical axis.

[0053] The generating unit 22 determines the signal peaks of each signal profile (S24). In this embodiment, a predetermined threshold is set for the intensity (I) of the thickness component, and the generating unit 22 determines the intensity (I) of the thickness component that exceeds the threshold as a signal peak. The signal peak represents a predetermined thickness component having a high thickness component intensity that exceeds the threshold. The signal peak is extracted from the intensity peaks included in the signal profile by the threshold.

[0054] The generation unit 22 determines whether the processes of S22 to S24 have been completed for all the predetermined regions (sections), and if so (YES in S25), ends the process. If the generation unit 22 has not completed the processes for all the sections (NO in S25), it returns to S22 and executes the process for the next section.

[0055] Fig. 6(a) illustrates a signal profile in a predetermined region. Here, in the signal profile, the horizontal axis indicates frequency (1 / z) and the vertical axis indicates the intensity (I) of the thickness component. Since the thickness component is the inverse of the frequency, the thickness component (z) becomes larger toward the left side of Fig. 6(a) and becomes smaller toward the right side.

[0056] 6(b) shows a schematic diagram of a signal profile in which the signal peaks of S24 are determined. The thresholds for the intensity of the thickness component are exemplified by a first threshold I1 and a second threshold I2. For example, when the first threshold I1 is used, the signal peak P1 is determined, and when the second threshold I2 is used, the signal peaks P1 and P2 are determined. The thresholds for the intensity of these thickness components may be arbitrarily set by the user.

[0057] FIG. 7 is a process flowchart showing the thickness estimation process in the thickness estimation step S13.

[0058] The thickness estimation unit 23 classifies signal peaks included in the signal profile generated by the generation unit 22 (S31). The thickness estimation unit 23 classifies the signal peaks based on the thickness components (z) in each predetermined region included in the signal profile, and generates a classification result. The thickness estimation unit 23 classifies signal peaks whose thickness components (z) are similar between signal profiles into one group. Therefore, the thickness estimation unit 23 classifies the signal peaks based on the thickness components or thickness component ranges associated with the classification results between signal profiles. The thickness estimation unit 23 can output the thickness components corresponding to the signal peaks. The classification method can employ an existing clustering algorithm, such as the k-means method.

[0059] The thickness estimation unit 23 refers to the thickness component associated with each classification result and determines whether the value is invalid for all signal profiles (S32). For example, in the case of a classification result classified as a thickness component of 0.2, if the target signal profile contains an intensity peak corresponding to a thickness component of 0.2, the value is valid, and if the intensity peak is not contained, the value is invalid.

[0060] If the target signal profile is an invalid value (YES in S32), the thickness estimation unit 23 executes a boundary detection process in a predetermined region corresponding to the signal profile (S33). The boundary detection process is a detection process by image scanning in the stacking direction of the predetermined region, and determines whether a boundary can be detected in the predetermined region by the thickness component associated with the classification result. If a boundary can be detected, the number of boundary lines is determined.

[0061] The thickness estimation unit 23 determines whether the number of boundary lines exceeds a threshold value in S33 (S34). This threshold value is a value for the number of boundary lines, and may be adjustable according to the thickness of the object or the image size of the predetermined region.

[0062] When the number of boundary lines exceeds the threshold (YES in S34), the thickness estimation unit 23 determines that the signal peaks included in the signal profile determined to be invalid values ​​are valid values ​​(S35). When the number of boundary lines does not exceed the threshold (NO in S34), the thickness estimation unit 23 determines that the signal peaks included in the signal profile determined to be invalid values ​​are invalid values.

[0063] The thickness estimation unit 23 determines whether the invalid value determination process of S32 has been completed for the signal profiles of all the predetermined sections classified as identical, and if completed (YES in S36), proceeds to the next process of S37. If the thickness estimation unit 23 has not completed the process for all the sections (NO in S36), it returns to S32 and executes the invalid value determination process for the next section.

[0064] The thickness estimation unit 23 determines whether the invalid value determination process of S32 has been completed for the signal profiles of all the predetermined sections of all the classification results, and if completed (YES in S37), ends the process. If the thickness estimation unit 23 has not completed the process for all the classification results (NO in S37), it returns to S32 and executes the invalid value determination process for the other classification results.

[0065] Fig. 8 is a schematic explanatory diagram of classification of signal peaks included in a signal profile in a thickness estimation process. In Fig. 8, an example of a signal profile with 3x3 partition images is shown, but the present invention is not limited to this. In Fig. 8(c) to (f), the shaded signal profiles indicate that signal peaks that fall into the same classification are valid values.

[0066] Fig. 8(a) shows the results of determining signal peaks for the signal profiles in S24 of Fig. 5. Signal peaks P1 and P2 are determined in signal profile S1, signal peak P3 is determined in signal profile S2, and signal peak P4 is determined in signal profile S3.

[0067] Fig. 8(b) shows the results of classifying the signal peaks of each signal profile in S31 of Fig. 7. In signal profile S1, signal peak P1 is classified into the first signal classification G1, signal peak P2 is classified into the second signal classification G2, in signal profile S2, signal peak P3 is classified into the second signal classification G2, and in signal profile S3, signal peak P4 is classified into the third signal classification G3. Here, signal peaks P2 and P3 have similar thickness components (z), and therefore are classified into the same second signal classification G2.

[0068] 8(c) shows the classification result of the first signal classification G1. The classification result of the first signal classification G1 defines the signal peaks corresponding to the first signal classification G1 included in the signal profiles S1 and S5 as valid values, and defines the intensity peaks corresponding to the first signal classification G1 included in the other signal profiles as invalid values.

[0069] 8(d) shows the classification result of the second signal classification G2, in which the signal peaks corresponding to the second signal classification G2 included in the signal profiles S1, S2, and S5 are set as valid values, and the intensity peaks corresponding to the second signal classification G2 included in the other signal profiles are set as invalid values.

[0070] 8(e) shows the classification result of the third signal classification G3. The classification result of the third signal classification G3 sets the signal peaks corresponding to the third signal classification G3 included in the signal profiles S3, S4, S6, S7, S8, and S9 as valid values, and sets the intensity peaks corresponding to the third signal classification G3 included in the other signal profiles as invalid values.

[0071] FIG. 8(f) shows the classification result of the first signal classification G1 after the process of FIG. 7 for validating the invalid values ​​included in the classification result of the first signal classification G1 of FIG. 8(c) is executed. Here, the signal peak P5 is newly determined as a valid value for the signal profile S3. The intensity peak corresponding to the first signal classification G1 of the signal profile S3 was an invalid value, but as a result of executing the boundary detection of S33 assuming that it is a valid signal peak, it is determined as valid in S34, and the invalid intensity peak is updated to a valid signal peak in S35. In FIG. 8(f), in addition to the signal peaks corresponding to the first signal classification G1 of the signal profiles S1 and S5 that were valid values ​​from the beginning, the intensity peaks corresponding to the first signal classification G1 other than the signal profile S2 that were invalid values ​​at the beginning are determined as valid signal peaks P5. In this way, the process of FIG. 7 has the effect of detecting the true thickness component by reviving (validating) the intensity peaks overlooked in S24 as valid signal peaks.

[0072] FIG. 9 is a process flowchart showing the model estimation process in the model estimation step S14.

[0073] The model estimation unit 24 calculates model parameters and estimates a counting model having the parameters (S41). The counting model is used to derive a thickness component (z) at coordinates (x, y) of an image component of image data. In one embodiment, the model estimation unit 24 calculates model parameters using input values ​​including the thickness component of each predetermined region, and estimates a counting model having the parameters. Here, the input values ​​include reference values ​​of the image components of each predetermined region. The reference values ​​of the image components can be, for example, the center coordinates of each predetermined region, but are not limited thereto, and may be any coordinates set uniformly in each predetermined region.

[0074] The counting model may be a function or a machine-learned model. Here, the counting model is not limited by the type of model as long as it can obtain the thickness component (z).

[0075] In this embodiment, the model employs a function capable of generating a quadratic surface. The function according to this embodiment is as shown in formula (1). Here, x and y in formula (1) correspond to the coordinates (x, y) in the image before division into the predetermined regions, and z in formula (1) corresponds to the thickness component (z) at the coordinates (x, y). In addition, α, β, γ, δ, ε, and ζ in formula (1) respectively indicate coefficients. For example, the image component of the predetermined region employs the value at the center position of the predetermined region. The model estimation unit 24 uses all (x, y, z) of the signal peaks belonging to the same signal classification as inputs for deriving the parameters of formula (1), and adjusts the coefficients so as to approximate each input. By adjusting the coefficients, the parameters α', β', γ', δ', ε', and ζ' of the counting model are derived, and a counting model that is valid for a specific signal classification is generated.

[0076]

number

[0077] In this embodiment, the computational model having the derived parameters represents a quadratic surface of the estimated thickness component (z) that varies in response to the image components (x, y).

[0078] The model estimation unit 24 derives a difference value based on the real value of the thickness component and the calculated value of the thickness component by the counting model (S42). The difference value is used to evaluate the generated counting model. Specifically, the model estimation unit 24 obtains the calculated value of the thickness component from the counting model by inputting the reference value (x, y) of the image component of the predetermined region to the estimated counting model. The model estimation unit 24 holds the real value of the thickness component of the corresponding predetermined region used to generate the counting model. The real value of the thickness component may be the thickness component corresponding to the signal peak. The model estimation unit 24 calculates the difference between the real value of the thickness component and the calculated value of the thickness component. The model estimation unit 24 calculates the difference in all the predetermined regions in the same manner, and derives the difference value as the sum of these differences. The difference value can be an accuracy index of the counting model.

[0079] The model estimation unit 24 executes the processes of S41 and S42 for all the classification results (S43).

[0080] The model estimation unit 24 determines an optimal count model based on the difference value corresponding to each classification result (S44). For example, if there are three classification results, three count models are estimated, and the respective difference values ​​are derived in S42. The model estimation unit 24 can adopt the count model with the smallest difference value as the one with the smallest difference from the real value and the highest accuracy.

[0081] In this manner, by adopting a mode in which an optimal counting model is adopted from among a plurality of counting models estimated from the classification results of a plurality of thickness components, it is possible to improve the counting accuracy.

[0082] The counting unit 25 performs counting of the objects while referring to the counting model. FIG. 10 is an image diagram of the counting model showing a quadratic surface. In FIG. 10, a thickness component (z) corresponding to an image component (x, y) is shown as a quadratic surface. The counting unit 25 performs counting of the objects while referring to an estimated value (z) of the thickness component for the image component (x, y) in the image data using the quadratic surface. In this way, by referring to the counting model, it is possible to accurately count the objects in accordance with the thickness (z) that changes depending on the width direction (x) and stacking direction (y).

[0083] In this embodiment, the counting unit 25 refers to the quadratic surface to count the objects in the image data. Specifically, the counting unit 25 refers to the quadratic surface in detecting the boundary line of the object in the image data. As an example, the counting unit 25 first calculates the thickness component (z) of the reference position near the center of the image data by substituting the coordinates (x, y) of the image component into the quadratic surface. Next, the counting unit 25 detects the boundary line of the object having the thickness of the laminated plate that matches the calculated thickness component (z) by searching the image data in the stacking direction using the calculated thickness component as a clue. Next, the counting unit 25 calculates the thickness component (z) at the position by substituting the coordinates (x, y) of the detected boundary line near the quadratic surface. Note that the coordinates (x, y) near the boundary line may be a position obtained by adding z / 2 (half the thickness component) from the boundary line in the stacking direction. The counting unit 25 searches for the next boundary line at a position obtained by adding the thickness component (z) calculated from the detected boundary line in the stacking direction. By repeating these processes up to the top and bottom ends in the stacking direction, it is possible to count all objects included in the stack with high accuracy using the thickness component that changes according to the image component.

[0084] <Embodiment 2> Fig. 11 shows a block diagram of a counting system 10 in embodiment 2. According to Fig. 11, the counting system 10 includes a counting device 11 and a terminal device 12, which are configured to be able to communicate with each other via a communication network NW. Here, there can be a plurality of terminal devices 12.

[0085] The terminal device 12 may be a smartphone, a tablet terminal, a personal computer, or the like. The terminal device 12 is preferably configured to include an imaging device. In the second embodiment, the terminal device 12 transmits image data to be counted to the counting device 11.

[0086] A computer such as a server device can be used as the counting device 11. The counting device 11 is configured as one device or a combination of multiple devices. The counting device 11 can take the form of, for example, a cloud server.

[0087] The counting device 11 has the same functional components (21-26) as the counting device 2. Note that some of the functional components may be realized by the terminal device 12. The counting device 11 receives image data from the terminal device 12 and counts the objects in the image data. The counting device 11 transmits the counting result to the terminal device 12.

[0088] The terminal device 12 receives the counting result from the counting device 11 and can display the result on a display or the like as an output device.

[0089] The counting system 10 of the second embodiment can, for example, be in the form of a web service, and thus can favorably assist a user in obtaining the counting results of objects using the terminal device 12. [Explanation of symbols]

[0090] 1 Counting System 2 Counting device 21 Image acquisition unit 22 Generation part 23 Thickness Estimation Section 24 Model Estimation Section 25. Counting Department 26 Output section 27 Control Unit DB storage 3. Imaging device 10 Counting System 11 Counting device 12 Terminal Equipment NW Communication network

Claims

1. A method for counting objects contained in a laminate, comprising the steps of: The image acquiring step, the generating step, the thickness estimating step, and the counting step are executed by a computer; The image acquisition step acquires image data of a lamination surface of a laminate, The generating step includes dividing the image data into a plurality of predetermined regions in a stacking direction and a width direction, and generating a signal profile indicating each thickness component in the predetermined regions; The thickness estimation step acquires a thickness component estimated from each of the predetermined regions, The counting method, wherein the counting step counts the objects based on the thickness components.

2. The thickness estimation step generates a classification result of the signal peak extracted based on the thickness component included in the signal profile, The counting method according to claim 1 , wherein the counting step counts the objects based on thickness components corresponding to the signal peaks of the respective classification results.

3. The thickness estimation step includes determining whether a signal peak is a valid value or an invalid value included in a signal profile in the classification result according to an intensity of a thickness component equal to or greater than a predetermined value, When a predetermined number or more of the boundary lines of the predetermined region are detected based on the thickness component in the classification result, the invalid value is determined to be the valid value; The counting method according to claim 2 , wherein the counting step counts the objects based on thickness components indicated by the signal profiles determined to be the valid values.

4. A model estimation step of calculating model parameters using input values ​​including thickness components of the respective predetermined regions and estimating a counting model having the parameters; The counting method according to claim 1 , wherein the counting step counts the objects based on a thickness component estimated by the counting model.

5. Calculating model parameters using input values ​​including thickness components of the respective predetermined regions, and estimating a plurality of counting models having the parameters, the counting models corresponding to the respective classification results; a model estimation step of determining an optimal counting model from a plurality of counting models corresponding to the classification results based on a difference between a real value of a thickness component of each predetermined region and a calculated value of the thickness component using the counting model, The counting method according to claim 2 or 3, wherein the counting step counts the objects based on a thickness component estimated by the counting model.

6. A counting system for objects contained in a stack, comprising: The apparatus includes an image acquisition unit, a generation unit, a thickness estimation unit, and a counting unit, The image acquisition unit acquires image data of a stacking surface of a stacked body, the generating unit divides the image data into a plurality of predetermined regions in a stacking direction and a width direction, and generates a signal profile indicating a thickness component in each of the predetermined regions; the thickness estimator acquires a thickness component estimated from each of the predetermined regions; A counting system, wherein the counting unit counts the objects based on the thickness components.

7. A program for counting objects included in a stack, comprising: A computer is caused to function as an image acquisition unit, a generation unit, a thickness estimation unit, and a counting unit; The image acquisition unit acquires image data of a stacking surface of a stacked body, the generating unit divides the image data into a plurality of predetermined regions in a stacking direction and a width direction, and generates a signal profile indicating a thickness component in each of the predetermined regions; the thickness estimator acquires a thickness component estimated from each of the predetermined regions; The counting unit counts the objects based on the thickness components.

Images