Information processing methods, apparatus, programs, and inspection systems

Abstract

To accurately determine a foreign matter included in an object or abnormality of the object.SOLUTION: An information processing apparatus includes: an acquisition unit which acquires a captured image obtained by capturing an object disposed between two polarizing plates; an abnormality detection unit which detects a foreign matter included in the object or abnormality of the object on the basis of the acquired captured image; and an abnormality determination unit which, if abnormality is detected, uses an abnormal part image in a region where the abnormality is detected and a preliminarily obtained correspondence relation between color information and thickness information to determine the foreign matter included in the object or abnormality of the object.SELECTED DRAWING: Figure 6

Description

Technical Field

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

Background Art

[0002] Conventionally, for example, humans have visually checked and determined whether foreign matter is mixed in a container such as a bottle containing medicines or the like. For this reason, it has been difficult to accurately determine the presence or absence of foreign matter in the container.

[0003] Patent Document 1 proposes a defect detection apparatus for a patterned retardation film that detects defects in a patterned retardation film in which a plurality of striped first and second retardation regions, each functioning as a λ / 4 wavelength plate and having slow axes substantially orthogonal to each other, are alternately arranged. This defect detection apparatus includes: first and second polarizing plates arranged in crossed Nicols so as to sandwich the patterned retardation film; a light source unit that irradiates inspection light onto the patterned retardation film via the first polarizing plate; an imaging device that captures an image of the patterned retardation film via the second polarizing plate to obtain a luminance image; and a defect detection unit that detects defects from the luminance image. In the first and second polarizing plates, the direction of one of the polarization transmission axes is adjusted such that in a state where the one polarization transmission axis is substantially parallel to the optical axis of either one of the first and second retardation regions, respective luminance values when the normal first and second retardation regions are imaged by the imaging device become the same level near the extinction state.

Prior Art Literature

Patent Literature

[0004]

Patent Literature 1

Summary of the Invention

Problem to be Solved by the Invention

[0005] The defect detection device described in Patent Document 1 can be used to visualize the surface shape of a film. However, there is room for improvement in its use for detecting abnormalities in shape.

[0006] The present invention has been made in view of these circumstances, and its purpose is to provide an information processing method, apparatus, program, and inspection system that can accurately determine foreign matter contained in an object or abnormalities in an object. [Means for solving the problem]

[0007] The information processing method according to the present invention involves an information processing device acquiring an image of an object placed between two polarizing plates, detecting foreign matter or abnormalities in the object based on the acquired image, and if an abnormality is detected, determining whether foreign matter or abnormalities in the object are present in the object using an abnormal partial image of the area where the abnormality was detected, and a predetermined correspondence between color information and thickness information. [Effects of the Invention]

[0008] As described above, according to the present invention, it is possible to accurately determine foreign matter contained in an object or abnormalities in the object. [Brief explanation of the drawing]

[0009] [Figure 1] This is a schematic diagram illustrating the outline of the inspection system according to this embodiment. [Figure 2] This is a block diagram showing the hardware configuration of the information processing device according to this embodiment. [Figure 3] This figure shows an example of an image taken of an object being inspected. [Figure 4] This diagram illustrates the distance light travels through a film when the film is tilted. [Figure 5] This is a diagram showing an example of a polarized color chart. [Figure 6] This figure shows the configuration of the information processing device according to this embodiment. [Figure 7] This flowchart shows the contents of the anomaly detection process by the information processing device according to this embodiment. [Modes for carrying out the invention]

[0010] An example of an embodiment of the disclosed technology will be described below with reference to the drawings. In each drawing, identical or equivalent components and parts are given the same reference numerals. Furthermore, the dimensional ratios in the drawings are exaggerated for illustrative purposes and may differ from actual ratios.

[0011] <Configuration of the inspection system according to this embodiment> Figure 1 is a schematic diagram illustrating the overview of the inspection system according to this embodiment. The inspection system according to this embodiment is a system that determines, for example, whether or not there is an abnormality in a packaging material such as a film. The inspection system according to this embodiment is configured to include an information processing device 1, a camera 3, a first polarizing plate 4, a second polarizing plate 5, and a backlight 6, etc.

[0012] In the inspection system according to this embodiment, the camera 3, the first polarizing plate 4, the second polarizing plate 5, and the backlight 6 are arranged in this order, and the camera 3 is connected to the information processing device 1 via, for example, a communication cable. The object to be inspected, such as a film, is placed between the first polarizing plate 4 and the second polarizing plate 5, and with light from the backlight 6 illuminating the object from behind via the second polarizing plate 5, the camera 3 takes a photograph of the object to be inspected via the first polarizing plate 4. The image taken by the camera 3 of the object to be inspected is provided to the information processing device 1, and the information processing device 1 performs processing to detect foreign matter contained in the object to be inspected or abnormalities in the object to be inspected based on the image taken by the camera 3.

[0013] Camera 3 comprises optical elements such as lenses and an image sensor such as a CCD (Charge Coupled Device) or CMOS (Complementary Metal Oxide Semiconductor). The camera 3 in this embodiment may be capable of receiving and capturing at least visible light, and may also be capable of capturing invisible light such as infrared or ultraviolet light. Camera 3 provides the captured image to the information processing device 1 as digital image data. Furthermore, the information processing device 1 controls camera 3's shooting, such as changing the shutter speed, adjusting the brightness, and focusing.

[0014] The first polarizer 4 and the second polarizer 5 are engineering components that transmit light vibrating in a predetermined direction and block light vibrating in other directions. The direction in which a polarizer transmits light is called the polarization axis (transmission axis), and in this embodiment, the first polarizer 4 and the second polarizer 5 are arranged so that their polarization axes intersect, and more preferably so that their polarization axes are orthogonal (90°). As a result, the first polarizer 4 and the second polarizer 5 constitute a so-called crossed Nicol optical system. Furthermore, for example, either the first polarizer 4 or the second polarizer 5 may be made rotatable, and the angle between the polarization axis of the first polarizer 4 and the polarization axis of the second polarizer 5 may be made variable. In this case, for example, a mechanism such as a motor and gears for rotating the polarizers is provided, and the rotation of the motor is controlled by the information processing device 1.

[0015] The combination of the first polarizer 4 and the second polarizer 5 may be, for example, a combination using two linear polarizers, or a combination using two circular polarizers, for example, a linear polarizer with a quarter-phase difference plate superimposed on it. When using circular polarizers, one of the first polarizer 4 and the second polarizer 5 is a right-rotating circular polarizer, and the other is a left-rotating circular polarizer.

[0016] The backlight 6 is configured using a light-emitting element such as an LED (Light Emitting Diode). In the present embodiment, the backlight 6 emits at least visible light, and may further emit invisible light such as infrared rays or ultraviolet rays.

[0017] The information processing apparatus 1 is obtained by installing a computer program that performs information processing according to the present embodiment on a general-purpose computer such as a personal computer or a server apparatus. The information processing apparatus 1 according to the present embodiment acquires a captured image obtained by capturing the inspection object by the camera 3, and performs abnormality detection processing on the acquired captured image using a trained learning model (so-called AI (Artificial Intelligence)).

[0018] Figure 2 is a block diagram showing the hardware configuration of the information processing apparatus 1 of the present embodiment.

[0019] As shown in Figure 2, the information processing apparatus 1 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a storage 14, an input unit 15, a display unit 16, and a communication interface (I / F) 17. Each component is connected to be capable of mutual communication via a bus 19.

[0020] The CPU 11 is a central processing unit that executes various programs and controls each unit. That is, the CPU 11 reads a program from the ROM 12 or the storage 14, and executes the program using the RAM 13 as a work area. The CPU 11 performs control of the above-described components and various arithmetic processes in accordance with programs stored in the ROM 12 or the storage 14. In the present embodiment, the ROM 12 or the storage 14 stores an information processing program that performs processing for detecting foreign matter contained in an inspection object or an abnormality of the inspection object. The information processing program may be a single program, or may be a program group composed of a plurality of programs or modules.

[0021] ROM 12 stores various programs and various data. RAM 13 temporarily stores programs or data as a working area. The storage 14 is configured of an HDD (Hard Disk Drive) or an SSD (Solid State Drive), and stores various programs including an operating system, and various data.

[0022] The input unit 15 includes a pointing device such as a mouse, and a keyboard, and is used for performing various inputs including a captured image captured by the camera 3. For example, a captured image obtained by capturing an inspection object as shown in FIG. 3 is input to the input unit 15. The captured image is, for example, an RGB image.

[0023] The display unit 16 is, for example, a liquid crystal display, and displays various types of information. The display unit 16 may adopt a touch panel method and function as the input unit 15.

[0024] The communication interface 17 is an interface for communicating with other devices, and standards such as Ethernet (registered trademark), FDDI, and Wi-Fi (registered trademark) are used, for example.

[0025] In the present embodiment, the inspection object is a packaging member formed of a transparent member, and the captured image is, for example, an image obtained by capturing a film that is a packaging member formed of a transparent member as shown in FIG. 3, and in this case, the color of the film indicates how far light travels through the film.

[0026] For example, in the case of a film, the standard color is purple, but other colors are observed in wrinkled portions. As the angle of the film becomes steeper, the color changes in the order of "purple" → "blue" → "green" → "yellow". At this time, the color depends on both the lower film and the upper film. This is because the more wrinkled a portion is, the longer distance light consequently travels through the film. In addition, the color also changes depending on the thickness of the film.

[0027] For example, as shown in Figure 4, when the film is tilted, the light travels a longer distance through the film than when it is flat, resulting in a change in color. There is a specific rule governing the order in which the colors change, and the distance traveled through the film can be estimated from the color.

[0028] Furthermore, while there is a pattern to the color changes, similar colors can sometimes be observed. Let's explain this by assuming that one sheet of film is 0.1 mm thick. For example, when there is one sheet of film (0.1 mm), an orange color is observed, but even in areas with severe wrinkles (approximately 0.5 mm), an orange-ish color is also shown. This is because similar colors appear periodically, as shown in the polarization color chart in Figure 5.

[0029] In Figure 5, the vertical axis represents retardation, the horizontal axis represents thickness, and the diagonal line represents the refractive index difference Δn of the two intrinsic polarized light after passing through the object being inspected. Assuming the refractive index difference Δn is known, the thickness can be determined from the color information.

[0030] In the lower right rectangular area of ​​Figure 3(B), due to misalignment, there is only one film sheet between the bottom surface and camera 3. In the upper right rectangular area of ​​Figure 3(B), blue and green colors are observed due to film entanglement. This is caused by wrinkles in the film on the bottom and top sides. If the wrinkles are larger, orange may also be observed.

[0031] Furthermore, if an object is placed inside a film-formed packaging material, the protrusion of that object may cause some deformation of the film's shape. For example, in the case of packaging material for instant ramen toppings, various colors can be observed on the film due to the unevenness caused by the ingredients. Also, abnormalities on the outer edge of the film can lead to serious problems such as holes, so careful inspection is necessary, but often colors different from what is expected are observed. For example, the angle may be uneven due to the effects of heat sealing, or there may be subtle differences in the unevenness of each package.

[0032] Given these circumstances, in order to use the visualization of surface shape for anomaly detection, it is necessary to determine which areas of the packaging material, and within which range of colors, should be observed to be considered normal.

[0033] Therefore, in this embodiment, first, a learning model for anomaly detection is trained to recognize normal conditions. At this time, a machine learning model for object detection or segmentation (for example, Mask-RCNN) is used to obtain the range of the object to be inspected in the image.

[0034] Furthermore, information such as shape and printed characters is used to determine the upper, lower, left, right, and central parts of the object being inspected.

[0035] The appearance of each region, visualized using a crossed-donicol optical system with circularly polarized light, is observed using samples without anomalies. Each acquired image is then used as a normal sample to train a learning model for anomaly detection (such as PaDiM PatchCore). For example, several hundred samples are needed, and a larger number is preferable.

[0036] Next, anomaly detection is performed. At this time, a machine learning model for object detection and segmentation (e.g., Mask-RCNN) is used to obtain the extent of the object to be inspected in the image. Then, using information such as shape and printed characters, the upper, lower, left, right, and central parts of the object to be inspected are determined. The images of each area, visualized using a crossed-donicol optical system with circular polarization, are then input into the learning model for anomaly detection to determine whether they look statistically abnormal.

[0037] Then, for each pixel in the area where an anomaly is detected, the color information of the pixel is converted into thickness information based on a polarization color chart, and based on the shape represented by the conversion result for each pixel, it is determined whether there is a foreign object contained in the object being inspected or an anomaly in the object being inspected.

[0038] Next, we will describe the functional configuration of the information processing device 1. Figure 6 is a block diagram showing an example of the functional configuration of the information processing device 1.

[0039] Functionally, the information processing device 1 comprises a learning model storage unit 20, an acquisition unit 22, an identification unit 24, a position determination unit 26, an anomaly detection unit 28, and an anomaly determination unit 30, as shown in Figure 6.

[0040] The learning model memory unit 20 stores a learning model that has been trained through machine learning to accept an input image and output information regarding the presence or absence of foreign objects or abnormalities in the object being inspected as captured in the image.

[0041] Specifically, the learning model is a learning model that has been machine-learned for each location of the object being inspected. For example, for the top, bottom, left side, right side, and center of the object being inspected, a learning model that has been machine-learned to accept partial images of that part as input and output information regarding the presence or absence of foreign objects or abnormalities in the object being inspected as captured in the partial image is stored in the learning model storage unit 20.

[0042] The acquisition unit 22 acquires images of the object to be inspected, which are captured by the camera 3.

[0043] The identification unit 24 identifies the image region in the acquired captured image in which the object to be inspected is captured. Specifically, it uses machine learning models for object detection and segmentation (e.g., Mask-RCNN) to identify the image region in the captured image in which the object to be inspected is captured.

[0044] The position determination unit 26 determines the position of the object to be inspected for each partial image within the identified image region. Specifically, for each partial image within the image region, it determines whether the position of the object to be inspected represented by that partial image is in the upper, lower, left, right, or central part of the object to be inspected.

[0045] More specifically, the shape of the object being inspected and information such as printed characters represented in the partial image are used to determine whether the area being inspected is the top, bottom, left side, right side, or center of the object.

[0046] The abnormality detection unit 28 detects foreign matter contained in the object being inspected or abnormalities in the object being inspected based on the acquired captured images.

[0047] Specifically, the system receives captured images as input and uses machine learning to input information about whether foreign objects or abnormalities are present in the object being inspected as captured in the images. The acquired images are then input to a machine learning model, the information output by the model is retrieved, and based on the retrieved information, foreign objects or abnormalities in the object being inspected are detected.

[0048] More specifically, for each partial image of the captured image, the partial image is input into a learning model corresponding to the determination result of the location of the object to be inspected represented by that partial image, and foreign matter or abnormalities in the object to be inspected represented by that partial image are detected.

[0049] If an abnormality is detected, the abnormality detection unit 30 uses the partial image in which the abnormality was detected, and a pre-determined correspondence between color information and thickness information, to determine whether the foreign object contained in the object being inspected or an abnormality in the object being inspected is represented by the partial image.

[0050] Specifically, when the abnormality determination unit 30 determines whether a foreign object is contained in the object to be inspected or whether an abnormality exists in the object to be inspected, it converts the color information of each pixel in the partial image where an abnormality is detected into thickness information using a predetermined correspondence between color information and thickness information based on a polarization color chart, and determines whether a foreign object is contained in the object to be inspected or whether an abnormality exists in the object to be inspected based on the shape represented by the conversion result for each pixel. For example, if the rate of change of thickness information between adjacent pixels is greater than or equal to a threshold, it is determined that the object to be inspected contains a foreign object or that the object to be inspected is abnormal.

[0051] <Operation of the information processing device according to this embodiment> Next, the operation of the information processing device 1 will be explained.

[0052] Figure 7 is a flowchart showing the flow of anomaly detection processing by the information processing device 1. The CPU 11 reads an information processing program from the ROM 12 or storage 14, loads it into the RAM 13, and executes it to perform anomaly detection processing. Images captured by the camera 3 are also input to the information processing device 1. Furthermore, the learning model storage unit 20 of the information processing device 1 is assumed to store learning models, each of which has been machine-learned for each position of the object to be inspected.

[0053] In step S100, the CPU 11, acting as the acquisition unit 22, acquires the image captured by the camera 3 of the object to be inspected.

[0054] In step S102, the CPU 11, acting as a specific unit 24, identifies the image region in the acquired captured image in which the object to be inspected is depicted.

[0055] In step S104, the CPU 11, acting as a position determination unit 26, determines the position of the object to be inspected for each partial image of the identified image region.

[0056] In step S106, the CPU 11, acting as an anomaly detection unit 28, inputs each partial image of the captured image into a learning model corresponding to the determination result of the position of the object to be inspected represented by that partial image.

[0057] In step S108, the CPU 11, acting as an anomaly detection unit 28, acquires information output by the learning model for each partial image of the captured image.

[0058] In step S110, the CPU 11, acting as an anomaly detection unit 28, determines whether a foreign object contained in the object to be inspected or an anomaly in the object to be inspected represented by the partial image has been detected in at least one partial image. If a foreign object contained in the object to be inspected represented by the partial image or an anomaly in the object to be inspected is detected in at least one partial image, the process proceeds to step S112. On the other hand, if no foreign object contained in the object to be inspected or an anomaly in the object to be inspected is detected in any of the partial images, the process proceeds to step S120.

[0059] In step S112, the CPU 11, acting as an abnormality determination unit 30, determines whether there is a foreign object contained in the object to be inspected or an abnormality in the object to be inspected. In this unit, it converts the color information of each pixel in the partial image where an abnormality is detected into thickness information.

[0060] In step S114, the CPU 11, acting as an abnormality determination unit 30, determines whether there is a foreign object contained in the object being inspected or an abnormality in the object being inspected, based on the shape represented by the conversion result for each pixel.

[0061] In step S116, the CPU 11 determines whether or not there are foreign substances in the object being inspected or abnormalities in the object being inspected. If it is determined that there are foreign substances in the object being inspected or abnormalities in the object being inspected, the process proceeds to step S118. On the other hand, if it is not determined that there are foreign substances in the object being inspected or abnormalities in the object being inspected, the process proceeds to step S120.

[0062] In step S118, the CPU 11 notifies the user via the display unit 16 that there is a foreign object in the object being inspected or an abnormality in the object being inspected, and terminates the abnormality detection process.

[0063] In step S120, the CPU 11 notifies the display unit 16 that there are no foreign objects in the object being inspected or any abnormalities in the object being inspected, and terminates the abnormality detection process.

[0064] As described above, the inspection system according to this embodiment acquires an image of an object to be inspected placed between two polarizing plates, detects foreign matter or abnormalities in the object to be inspected based on the acquired image, and if an abnormality is detected, determines the foreign matter or abnormality in the object to be inspected using the abnormal area image of the area where the abnormality was detected, and a pre-determined correspondence between color information and thickness information. This makes it possible to accurately determine foreign matter or abnormalities in the object to be inspected.

[0065] <Variation> It should be noted that the present invention is not limited to the embodiments described above, and various modifications and applications are possible without departing from the spirit of the invention.

[0066] For example, if the inspection target, which is a packaging material, contains multiple types of contents, the location and extent of those contents within the inspection target can be detected using techniques such as object detection. Alternatively, the changes in the film shape (color) around the contents due to the protrusion of those contents can be individually learned, and abnormalities can be detected in the same manner as in the above embodiment. For example, if there are small shrimp in the dried ingredients, because of the thickness, even if there is a significant change in shape around them, it can be considered normal and not abnormal.

[0067] Furthermore, instant noodles sometimes have printed designs on the film, such as on the dried ingredients. In this case, the backlight may not penetrate, making it impossible to visualize the shape using crossed Nicols in that area. In this case, an infrared camera may be used instead of a visible light camera. When using an infrared camera, infrared light often penetrates the ink, reducing the influence of the ink. Since infrared cameras are often monochrome, in actual implementation, it is necessary to place multiple bandpass filters that transmit only light within a certain wavelength range in front of the infrared camera and observe light in various wavelength ranges by switching between them. The polarization color chart is a diagram for color cameras, but in the infrared region as well, the wavelength of light changes according to certain rules.

[0068] The color camera is equipped with three types of RGB bandpass filters on its sensor. The sensitivity of the CMOS or other sensors used in color cameras is approximately 400nm to 1000nm, allowing observation of light within this wavelength range. As an infrared camera, for example, an InGaAS camera can be used. This is an infrared camera that uses a sensor capable of observing light in the approximately 900nm to 1700nm range. For example, by placing multiple types of bandpass filters in front of the infrared camera, similar to the RGB bandpass filters in a color camera, and switching between filters during shooting, the wavelength of light at each location in the image can be measured. Wavelength corresponds to color in a color image, and therefore, a polarization color chart for the infrared region can be created.

[0069] Furthermore, the various processes that the CPU reads and executes in each of the above embodiments may be executed by various processors other than the CPU. Examples of such processors include GPUs (Graphics Processing Units), FPGAs (Field-Programmable Gate Arrays), and other PLDs (Programmable Logic Devices) whose circuit configurations can be changed after manufacturing, as well as dedicated electrical circuits that are processors with circuit configurations specifically designed to execute particular processes, such as ASICs (Application Specific Integrated Circuits). Also, the anomaly detection process may be executed by one of these various processors, or by a combination of two or more processors of the same or different types (for example, multiple FPGAs, and a combination of a CPU and an FPGA). More specifically, the hardware structure of these various processors is an electrical circuit that combines circuit elements such as semiconductor elements.

[0070] Furthermore, although the above embodiments describe a configuration in which the information processing program is pre-stored (installed) in the storage 14, the invention is not limited to this. The program may be provided in a form stored on a non-transitory storage medium such as a CD-ROM (Compact Disk Read Only Memory), DVD-ROM (Digital Versatile Disk Read Only Memory), or USB (Universal Serial Bus) memory. Alternatively, the program may be provided in a form downloaded from an external device via a network.

[0071] The following additional information is disclosed regarding the embodiments described above.

[0072] (Additional note 1) An information processing device, Memory and At least one processor connected to the memory, Includes, The aforementioned processor, By capturing an image of an object placed between two polarizing plates, Based on the acquired captured images, foreign matter or abnormalities in the object are detected. If an anomaly is detected, the foreign matter contained in the object or an anomaly in the object is determined using the anomaly portion image of the area where the anomaly was detected, and the pre-determined correspondence between color information and thickness information. An information processing device configured in such a way.

[0073] (Additional note 2) A non-temporary storage medium that stores a program executable by a computer to perform anomaly detection processing, The aforementioned anomaly detection process is: By capturing an image of an object placed between two polarizing plates, Based on the acquired captured images, foreign matter or abnormalities in the object are detected. If an anomaly is detected, the foreign matter contained in the object or an anomaly in the object is determined using the anomaly portion image of the area where the anomaly was detected, and the pre-determined correspondence between color information and thickness information. Non-transitory storage medium. [Explanation of Symbols]

[0074] 1. Information Processing Device 3 cameras 4. First polarizing plate 5. Second polarizer 6 Backlight 11 CPU 14 Storage 15 Input section 16 Display section 20 Learning Model Memory Unit 22 Acquisition Department 24 Specific part 26 Position determination section 28 Anomaly detection unit 30 Abnormality determination section

Claims

1. Information processing device, An image is obtained of an object placed between two polarizing plates arranged so that their polarization axes intersect. Based on the acquired captured images, foreign matter or abnormalities in the object are detected. If an abnormality is detected, the abnormal portion image of the area where the abnormality was detected from the captured image, and a pre-determined correspondence between color information and thickness information are used to convert the color information of the abnormal portion image into thickness information representing the thickness of the object. Based on the converted thickness information, a determination is made as to whether the object contains foreign matter or whether the object has an abnormality. Information processing methods.

2. The information processing method according to claim 1, in which, when determining foreign matter contained in the object or an abnormality of the object, the color information of each pixel in the abnormal portion image is converted into thickness information, and the foreign matter contained in the object or an abnormality of the object is determined based on the shape represented by the conversion result for each pixel.

3. The information processing device acquires an image captured by illuminating an object placed between the two polarizing plates with a backlight. The information processing method according to claim 1.

4. The aforementioned information processing device The acquired image is input to a machine learning model that accepts captured images as input and outputs information regarding the presence or absence of foreign objects or abnormalities in the objects captured in the images. The information output by the aforementioned learning model is acquired, Based on the acquired information, foreign matter contained in the object or abnormalities in the object are detected. The information processing method according to claim 1.

5. The aforementioned information processing device Furthermore, based on the acquired captured image, the position of the object represented by each partial image of the captured image is determined. The aforementioned learning model is a learning model in which machine learning has been performed for each position of the object, For each partial image of the captured image, the partial image of the captured image is input to the learning model corresponding to the determination result of the position of the object represented by the partial image. The information processing method according to claim 4, which acquires information output by the learning model.

6. The aforementioned information processing device Identify the image region in which the object is captured in the acquired image, Based on the identified image region, foreign matter contained in the object or an abnormality in the object is detected. The information processing method according to claim 1.

7. The object in question is a packaging member made of a transparent material. The information processing method according to any one of claims 1 to 6.

8. An acquisition unit that acquires an image of an object placed between two polarizing plates arranged so that their polarization axes intersect, An abnormality detection unit that detects foreign matter contained in the object or abnormalities in the object based on the acquired captured image, If an abnormality is detected, an abnormality determination unit uses the abnormal portion image of the region where the abnormality was detected from the captured image, and a pre-determined correspondence between color information and thickness information to convert the color information of the abnormal portion image into thickness information representing the thickness of the object, and determines whether there is a foreign object contained in the object or an abnormality in the object based on the converted thickness information. Information processing device including

9. Obtain an image of an object placed between two polarizing plates arranged so that their polarization axes intersect, Based on the acquired captured images, foreign matter or abnormalities in the object are detected. If an abnormality is detected, the abnormal portion image of the area where the abnormality was detected from the captured image, and a pre-determined correspondence between color information and thickness information are used to convert the color information of the abnormal portion image into thickness information representing the thickness of the object. Based on the converted thickness information, a determination is made as to whether the object contains foreign matter or whether the object has an abnormality. An information processing program that causes a computer to perform a task.

10. Two polarizing plates arranged so that their polarization axes intersect, A camera that photographs an object placed between the two polarizing plates, An information processing apparatus having an acquisition unit that acquires an image captured by the camera, an abnormality detection unit that detects foreign matter contained in the object or an abnormality in the object based on the acquired image, and an abnormality determination unit that, when an abnormality is detected, uses an abnormal partial image of the region in the image where the abnormality was detected, and a predetermined correspondence between color information and thickness information to convert the color information of the abnormal partial image into thickness information representing the thickness of the object, and determines whether foreign matter contained in the object or an abnormality in the object based on the converted thickness information. An inspection system equipped with the following features.

Images