Installation surface inspection device

Abstract

Provided is an installation surface inspection device capable of more appropriately confirming the presence or absence of a fallen object on an installation surface of a device. An installation surface inspection device 50 comprises: a horizontal support part 533 which extends in a horizontal direction and at least a part of which is installed under a PTP packaging machine 10 as a device; and a camera 531 which is attached to the horizontal support part 533 and can image an installation surface. The entire surfaces visible when viewed from above at least at the horizontal support part 533 and the camera 531 are inclined surface parts 533a, 531c inclining downward toward a prescribed outer edge side. Thus, a fallen object Rb is less likely to be placed on the horizontal support part 533 or the like, so that the fallen object Rb can be more reliably imaged by the camera 531. As a result, it is possible to more appropriately confirm the presence or absence of the fallen object Rb on the installation surface.

Description

Installation Surface Inspection Device

[0001] The present invention relates to an installation surface inspection device for inspecting whether there is any fallen object on an installation surface of an apparatus.

[0002] Apparatuses having a function of conveying various articles, such as manufacturing apparatuses, conveying apparatuses, and packaging apparatuses, are widely known. For example, in a packaging apparatus for manufacturing blister sheets (blister packaging machine), blister sheets are manufactured by packaging tablets and the like while conveying tablets or capsules as articles.

[0003] Incidentally, in the above-mentioned apparatuses, fallen articles (fallen objects) may enter the installation surface of the apparatus (the surface located below the apparatus). Here, when the articles are pharmaceuticals or the like, the articles must be managed extremely strictly in order to prevent mixing of different types of tablets, etc. Therefore, it is necessary to check whether there is any fallen object on the installation surface.

[0004] However, the work of visually checking the installation surface by an operator to confirm the presence or absence of fallen objects requires labor and time, and there is a possibility that the presence or absence of fallen objects cannot be properly confirmed only by visually checking the installation surface. Accordingly, it has been considered to use an inspection module including: a horizontal support portion extending parallel to the installation surface; a vertical support portion extending vertically downward from the horizontal support portion, the lower surface of which is in contact with the installation surface; and a camera attached to the horizontal support portion and capable of capturing an image of the installation surface (see, for example, Patent Document 1). By installing this inspection module under the apparatus, it becomes possible to confirm the presence or absence of fallen objects based on the image obtained by the camera.

[0005] Japanese Unexamined Patent Publication No. 2020-94897

[0006] However, when the above-mentioned inspection module is used, there is a risk that fallen objects may land on the horizontal support portion or the camera due to bouncing off the installation surface or other reasons. Fallen objects that land on the horizontal support portion or the like are located outside the imaging range of the camera, and as a result, there is a risk that fallen objects may be missed, or it may take a lot of time to find fallen objects.

[0007] The present invention has been made in view of the above circumstances, and its purpose is to provide an installation surface inspection device that can more appropriately confirm whether or not there are any fallen objects on the installation surface of the device.

[0008] Below, we will describe, in separate sections, each means suitable for achieving the above objectives. Furthermore, we will add notes on the effects and benefits specific to each means as needed.

[0009] Means 1. An installation surface inspection device for inspecting the presence or absence of fallen objects, which are said to be the said objects, on the installation surface of a device equipped with the function of transporting a predetermined object, comprising: a horizontal support portion extending horizontally and at least a portion of which is installed below the device; and a camera attached to the horizontal support portion and capable of imaging the installation surface, wherein at least the entire surface of the horizontal support portion and the camera that is visible when viewed from above is an inclined portion that slopes downward toward a predetermined outer edge.

[0010] According to the above-described means 1, the entire surface of the horizontal support and the camera that is visible from above is a sloped surface that inclines downward toward a predetermined outer edge. Therefore, even if a falling object bounces off the installation surface and ends up on the horizontal support or camera, the object is more likely to slide down the sloped surface and fall onto the installation surface. Consequently, it becomes less likely for falling objects to land on the horizontal support, etc., and thus more reliably prevents falling objects from remaining outside the camera's imaging range. This allows the camera to more reliably image falling objects, and consequently, more appropriately confirm the presence or absence of falling objects on the installation surface.

[0011] Furthermore, the camera may comprise a camera body capable of capturing images of the mounting surface, and a cover that surrounds the camera body and whose entire surface, when viewed from above, is shaped like a cone, pyramid, or frustum of a cone or frustum of a pyramid. In this case, the cover will have a sloped surface. The gap formed between the camera body and the cover may be set to a size that prevents falling objects from entering, or it may be sealed off.

[0012] Means 2. The mounting surface inspection device according to Means 1, characterized in that the horizontal support portion is attached to the device so that the entire camera and the components supporting the camera are floating above the mounting surface.

[0013] According to the above-described method 2, the horizontal support is attached to the device, so that the camera and the components supporting the camera are suspended above the mounting surface. Therefore, compared to the case where the legs supporting the horizontal support are erected on the mounting surface, the blind spots in the camera's imaging range can be reduced. This allows for more accurate confirmation of whether or not there are any fallen objects on the mounting surface.

[0014] Means 3. The installation surface inspection device according to Means 2, characterized in that only the base end of the horizontal support portion is attached to the device, and the camera is attached to the side surface of the tip of the horizontal support portion.

[0015] According to the above-described means 3, the horizontal support is a cantilevered structure in which only the base end is attached to the device, and the camera is attached to the side of the tip of the horizontal support. Therefore, the camera can be positioned higher than when the camera is attached to the bottom surface of the horizontal support. This allows for a wider imaging range by the camera, and consequently, makes it possible to more appropriately check for the presence or absence of fallen objects on the installation surface.

[0016] Means 4. An installation surface inspection device according to Means 2, comprising an installation portion interposed between the horizontal support portion and the device for attaching the horizontal support portion to the device, wherein the entire surface of the installation portion that is visible when viewed from above is an installation portion side slope that slopes downward toward a predetermined outer edge.

[0017] According to the above-described means 4, the entire surface of the mounting portion that is visible when viewed from above is a mounting portion side slope that inclines downward toward a predetermined outer edge. Therefore, it is less likely that a fallen object will land on the mounting portion, preventing the camera from capturing an image of the object. This makes it possible to more appropriately check for the presence or absence of fallen objects on the installation surface.

[0018] Means 5. The installation surface inspection device according to Means 1, characterized in that the acute angle among the angles that the inclined surface makes with respect to the horizontal surface is 50° or more.

[0019] According to the above-described method 5, the acute angle (inclination angle) of the inclined surface with respect to the horizontal surface is set to 50° or more. Therefore, it is possible to more reliably prevent falling objects from landing on the horizontal support section, etc.

[0020] Furthermore, it is preferable that the angle of inclination for the mounting side slope portion and the light-shielding plate side slope portion described later be 50° or more.

[0021] Means 6. The installation surface inspection device according to Means 1, characterized in that the horizontal support portion has a shape in which two inclined portions, whose inclination directions are opposite to each other, are connected at the top, and the top portion has a curved surface shape that is convex upward.

[0022] According to the above-described means 6, the horizontal support section has a shape in which two inclined surfaces with opposite directions of inclination are connected at the top, and this top is not angular but curved. Therefore, it is possible to more reliably prevent falling objects from landing on the horizontal support section by preventing them from getting caught on the top. This makes it possible to more appropriately check for the presence or absence of falling objects on the installation surface of the device. In addition, it is possible to more reliably prevent damage to falling objects caused by contact with the top.

[0023] Furthermore, regarding the radius of curvature of the top, a smaller value is preferable in terms of preventing falling objects from getting caught on the top. Therefore, it is preferable to set the radius of curvature of the top to 1 mm or less. However, if the radius of curvature of the top is excessively small, the shape of the top will be almost the same as an angular shape. Therefore, it is preferable to set the radius of curvature of the top to 0.5 mm or more. Means 7. An installation surface inspection device according to Means 1, comprising: an illumination means located below the camera and irradiating a predetermined light onto at least the imaging range of the camera on the installation surface; and a light-shielding plate that covers the illumination means from above to prevent the light irradiated from the illumination means from directly entering the camera, wherein the entire surface of the light-shielding plate that is visible when viewed from above is a light-shielding plate side slope portion that slopes downward toward a predetermined outer edge.

[0024] According to the above means 7, the illumination means can illuminate at least the imaging range of the camera on the installation surface. Therefore, the camera can capture images of fallen objects on the installation surface in a clearer state. In addition, when visually inspecting the installation surface, fallen objects can be found more easily.

[0025] Furthermore, the light-shielding plate prevents light emitted from the illumination means from directly entering the camera. Therefore, it is possible to more reliably prevent problems such as halation in the image data obtained by the camera.

[0026] In addition, the entire surface of the light-shielding plate that is visible when viewed from above is a sloping surface that inclines downward toward a predetermined outer edge. Therefore, it is less likely that falling objects will land on the light-shielding plate, preventing the camera from capturing images of those objects. This allows for more accurate confirmation of the presence or absence of falling objects on the installation surface.

[0027] Means 8. An installation surface inspection device according to Means 1, comprising: an identification means generated by training a neural network having an encoding unit for extracting feature quantities from input image data and a decoding unit for reconstructing image data from the feature quantities, using only image data relating to the installation surface without fallen objects as training data; a reconstructed image data acquisition means capable of acquiring reconstructed image data as reconstructed image data by inputting inspection image data obtained by the camera to the identification means; a comparison means capable of comparing the inspection image data and the reconstructed image data; and a determination means capable of determining the presence or absence of fallen objects on the installation surface based on the comparison results by the comparison means.

[0028] Furthermore, the training data may be image data obtained by actually photographing an installation surface without falling objects (real image data), or it may be a virtually generated image of an installation surface without falling objects (virtual image data). In addition, various types of image data may be used as training data, such as image data under different ambient light conditions (for example, image data obtained during the day or at night, image data obtained in sunny or rainy weather), image data when an operator is working, and image data when the device is operating or stopped, as these are all examples of image data that may have different influences on the condition of the installation surface.

[0029] Furthermore, the above-mentioned "neural network" includes, for example, a convolutional neural network having multiple convolutional layers. The above-mentioned "learning" includes, for example, deep learning. The above-mentioned "discrimination means (generative model)" includes, for example, an autoencoder or a convolutional autoencoder.

[0030] In addition, the "identification means" is generated by training it only with image data relating to installation surfaces without fallen objects. Therefore, when inspection image data relating to installation surfaces with fallen objects is input to the identification means, the reconstructed image data generated will be almost identical to the inspection image data from which the fallen objects have been removed.

[0031] According to the above-described means 8, the inspection image data is compared with the reconstructed image data obtained by inputting the inspection image data into the identification means, and the presence or absence of fallen objects on the installation surface is determined based on the comparison result. Therefore, compared to determining the presence or absence of fallen objects by visually inspecting the image data, the presence or absence of fallen objects can be confirmed more accurately and quickly.

[0032] Furthermore, the conditions of the installation surface (e.g., ambient light) can be matched for both image data being compared. This allows for more accurate verification of the presence or absence of fallen objects on the installation surface.

[0033] Means 9. The installation surface inspection device according to Means 1, characterized in that the inclined surface has a buffering means capable of buffering the impact applied to the falling object when it comes into contact with the falling object.

[0034] According to the above means 9, the cushioning means of the slope can cushion the impact applied to the falling object. Therefore, it is possible to more reliably prevent the falling object from moving significantly (for example, bouncing) due to the impact applied from the slope. As a result, the falling object can be imaged more reliably by the camera, and the presence or absence of falling objects on the installation surface can be confirmed more appropriately.

[0035] Furthermore, by providing a cushioning mechanism, cracking of falling objects due to impact can be more reliably prevented, making it easier to find the fallen objects in their original shape. Consequently, the recovery of fallen objects can be made easier. Preventing cracking of falling objects is particularly effective when the object contains special chemical components.

[0036] Furthermore, the cushioning means can be made of impact-absorbing materials such as soft rubber, felt, urethane, or elastic materials. The cushioning means may also be applied to the aforementioned sloped portion on the mounting side or the sloped portion on the light-shielding plate side.

[0037] Furthermore, the technical aspects related to the above means may be combined as appropriate. For example, at least one of the technical aspects related to means 4 to 9 may be combined with the technical aspects related to means 3.

[0038] This is a perspective view of a PTP sheet. This is a partially enlarged cross-sectional view of a PTP sheet. This is a perspective view of a PTP film. This is a schematic diagram of the configuration of a PTP packaging machine. This is a schematic perspective view showing the inspection unit, etc., when viewed from diagonally above. This is a schematic perspective view showing the inspection unit, etc., when viewed from diagonally below. This is a cross-sectional view of the line J-J in Figure 5. This is a cross-sectional view of the line K-K in Figure 5. This is a schematic cross-sectional view of the horizontal support section. This is a cross-sectional view of the line L-L in Figure 5. This is a block diagram showing the functional configuration of the installation surface inspection device. This is a schematic diagram for explaining the structure of a neural network. This is a flowchart showing the learning process flow of a neural network. This is a flowchart showing the judgment process flow. This is a flowchart showing the judgment process flow. This is a flowchart showing the judgment process flow. This is a schematic cross-sectional view of a horizontal support section having a buffer section in another embodiment. This is a schematic cross-sectional view showing a horizontal support section, etc., with both ends attached to a PTP packaging machine in another embodiment. This is a schematic perspective view showing a horizontal support section, etc., with both ends attached to a PTP packaging machine in another embodiment. This is a schematic perspective view showing a cover in another embodiment. This is a schematic plan view showing a cover in another embodiment.

[0039] The following describes one embodiment with reference to the drawings. First, we will describe a PTP packaging machine as the "device" to which the installation surface inspection device is applied. The PTP packaging machine is a packaging device for manufacturing a PTP sheet 1 (see Figures 1 and 2) which has a container film 3 with a plurality of pockets 2 and a cover film 4 attached to the container film 3 so as to close the pockets 2. The pockets 2 are filled with tablets 5 as "articles". The tablets 5 are circular in shape when viewed from above and are relatively easy to roll.

[0040] As shown in Figure 4, at the upstream end of the PTP packaging machine 10, a raw material 11 is provided, which consists of a strip of container film 3 wound around it, and the container film 3 is fed out from here and transported along a predetermined path. Along the transport path of the container film 3, the PTP packaging machine 10 is equipped with, in order from upstream, a heating device 15, a pocket forming device 16, a filling device 21, an inspection device 22, a sealing device 25, a slit forming device 33, an engraving device 34, and a sheet punching device 37.

[0041] The heating device 15 preheats the container film 3. The pocket portion forming device 16 forms the pocket portions 2 in the container film 3 heated by the heating device 15.

[0042] The filling device 21 fills the tablets 5 into the pocket portions 2 of the continuously conveyed container film 3. The inspection device 22 performs inspections on, for example, whether each pocket portion 2 is reliably filled with the tablet 5, whether there is any abnormality in the tablet 5, and whether there is any foreign matter mixed into the pocket portion 2, and the like.

[0043] The sealing device 25 includes a heating roll 25a having a heating function and a film feed roll 25b that continuously conveys the container film 3, and attaches the cover film 4 to the container film 3 so as to close the pocket portions 2. More specifically, the belt-shaped cover film 4 is guided toward the heating roll 25a side, and the container film 3 and the cover film 4 pass between the two rolls 25a and 25b in a heated and pressure-welded state, whereby the cover film 4 is attached to the container film 3. By attaching the cover film 4 to the container film 3, a belt-shaped PTP film 6 (see FIG. 3) in which the tablets 5 are accommodated in the respective pocket portions 2 is manufactured.

[0044] The slit forming device 33 forms separating slits at predetermined positions of the PTP film 6. The marking device 34 applies a mark to a predetermined position (for example, a tag portion) of the PTP film 6. Note that in FIG. 1, illustration of the separating slits and the marking is omitted.

[0045] The sheet punching device 37 has a function of punching the outer edge of the PTP film 6 into units of one PTP sheet 1, that is, separating the PTP sheet 1 from the PTP film 6. The PTP sheet 1 obtained by the sheet punching device 37 is conveyed by a conveyor 39 and stored in a finished product hopper 40. On the other hand, the scrap 42 remaining after punching in the PTP film 6 is cut into a predetermined size by a cutting device 41 and then stored in a scrap hopper 43.

[0046] In this PTP packaging machine 10, the PTP sheet 1 is manufactured by, for example, packaging the tablets 5 filled in the pocket portions 2 with both films 3 and 4 while conveying the tablets 5 together with the container film 3. That is, the PTP packaging machine 10 has a function of conveying the tablets 5.

[0047] In addition, in the PTP packaging machine 10, the conveying mode of the container film 3 or the PTP film 6 is appropriately changed to continuous conveyance or intermittent conveyance in order to correspond to the processing performed by the various devices described above. In order to prevent slack of the container film 3 or the like caused by such a change in the conveying mode, a tension roll for maintaining the tension of the container film 3 or the like within a predetermined range is provided at a predetermined position in the conveying path of the container film 3 or the like.

[0048] Next, with reference to Fig. 5 and the like, an installation surface inspection device 50 for inspecting whether there is a fallen object Rb that is a fallen tablet 5 on the installation surface 100 of the PTP packaging machine 10 will be described. As shown in Figs. 5 to 11, the installation surface inspection device 50 includes an inspection unit 51 and a control device 52. In Fig. 5 and the like, only the lower part of the PTP packaging machine 10 is shown in a simplified manner. The PTP packaging machine 10 is supported by a plurality of leg portions 10k.

[0049] First, the inspection unit 51 will be described. The inspection unit 51 is a device for acquiring image data for inspection, and a plurality (four in the present embodiment) of inspection units 51 are provided at equal intervals along the longitudinal direction of the installation surface 100 so as to cover almost the entire area of the installation surface 100. Each inspection unit 51 comprises one camera unit 53 and a plurality (four in the present embodiment) of illumination units 54 respectively.

[0050] The camera unit 53 is an integrated unit including a camera 531, and a mounting portion 532 and a horizontal support portion 533 which are components for mounting the camera 531 to the PTP packaging machine 10.

[0051] The camera 531 is a device for capturing an image of the installation surface 100 from directly above. The camera 531 is attached to a side surface of a distal end portion of the horizontal support portion 533, and comprises a camera body 531a and a cover 531b (see Figs. 6 and 8).

[0052] The camera body 531a is sensitive to the wavelength range of light emitted from the illumination device 542 of the illumination unit 54 (described later), and is composed of, for example, a CCD camera or a CMOS camera. When the light emitted from the illumination device 542 illuminates the installation surface 100, the camera 531 takes two-dimensional images of the light reflected from the installation surface 100 and the fallen object Rb. The image data (e.g., luminance image data) obtained by the camera 531 is converted into a digital signal (image signal) inside the camera 531 and then input to the control device 52 in the form of a digital signal. In this embodiment, it is possible to image almost the entire area of ​​the installation surface 100 using multiple cameras 531.

[0053] Furthermore, a 3D camera capable of acquiring distance information to the installation surface 100 and the falling object Rb (for example, a TOF (Time-of-Flight) camera) may be used as the camera 531. Alternatively, a camera capable of acquiring image data including height information may be used as the camera 531.

[0054] The cover 531b surrounds the camera body 531a, and the entire surface visible when viewed from above has the same shape as the side surface of a cone. As a result, the entire surface visible when viewed from above on the camera is sloped downward toward a predetermined outer edge (the outer edge from which the object will fall), forming a camera-side sloped surface 531c configured so that falling objects Rb cannot rest on it. The acute angle α1 (see Figure 8) of the angle that the camera-side sloped surface 531c makes with respect to the horizontal plane is 50° or more.

[0055] Furthermore, in this embodiment, the size of the gap 531d (see Figure 6) formed between the camera body 531a and the cover 531b and opening downwards is set to a size that prevents falling objects Rb from entering the gap 531d. Alternatively, the gap 531d may be closed to prevent falling objects Rb from entering the gap 531d.

[0056] The mounting portion 532 constitutes the mounting part of the camera unit 53 to the PTP packaging machine 10 (for example, the outer frame of the PTP packaging machine 10), and is a plate-like shape extending in the vertical direction. Furthermore, the entire surface of the mounting portion 532 that is visible when viewed from above is inclined downward toward a predetermined outer edge (the outer edge from which the object will fall), and is configured as a mounting portion side sloped surface 532a that prevents falling objects Rb from resting on it. The acute angle α2 (see Figure 8) of the angle that the mounting portion side sloped surface 532a makes with respect to the horizontal plane is 50° or more.

[0057] The horizontal support portion 533 is attached to the PTP packaging machine 10 via the mounting portion 532 and, together with the mounting portion 532, serves to support the camera 531 below the PTP packaging machine 10. The horizontal support portion 533 is rod-shaped and extends horizontally, and at least a portion of it (in this embodiment, most of it excluding the base end) is installed below the PTP packaging machine 10.

[0058] Furthermore, the horizontal support portion 533 has a cantilever structure in which only its base end is attached to the PTP packaging machine 10 via the mounting portion 532. As a result of the horizontal support portion 533 being attached to the PTP packaging machine 10 via the mounting portion 532, the camera 531 and the mounting portion 532 and horizontal support portion 533 (i.e., the camera unit 53), which are components that support the camera 531, are suspended above the mounting surface 100.

[0059] Furthermore, the entire surface of the horizontal support portion 533 that is visible when viewed from above is configured as a support portion side sloped surface 533a that slopes downward toward a predetermined outer edge (the outer edge toward which the object will fall) so that the falling object Rb does not rest on it. As a result, at least the entire surfaces of the horizontal support portion 533 and the camera 531 that are visible when viewed from above are configured as sloped surfaces 533a and 531c that slope downward toward the predetermined outer edge. In this embodiment, the camera side sloped surface 531c and the support portion side sloped surface 533a each correspond to a "slope".

[0060] Furthermore, the phrase "the surface of the horizontal support portion 533 and the camera 531 that is visible when viewed from above" can be interpreted as the surface that is visible when the horizontal support portion 533 and the camera 531 are viewed together from above. Therefore, for example, if the camera 531 is attached to the lower surface of the horizontal support portion 533, the surface of the camera 531 that is covered by the horizontal support portion 533 does not necessarily have to be a sloped surface.

[0061] Furthermore, the acute angle α3 (see Figure 7) of the angle that the support-side inclined surface 533a makes with respect to the horizontal plane is set to 50° or more.

[0062] In addition, the horizontal support portions 533 for the two inspection units 51 that correspond to the outermost edges of the installation surface 100 have a right-angled triangular cross-section, with only one support portion-side inclined surface portion 533a (see Figure 7).

[0063] On the other hand, the horizontal support portion 533 related to the other inspection unit 51 has an isosceles triangular cross-section, with two support portion side inclined portions 533a, which are in opposite directions of inclination, connected at the apex portion 533b (see Figures 7 and 9). The apex portion 533b has a curved surface that is convex upward, and its radius of curvature r is set to, for example, 0.5 mm or more and 1 mm or less.

[0064] Furthermore, the support-side inclined surface 533a of each inspection unit 51 is shaped like an inclined surface that gradually slopes downward toward the imaging range side of the camera 531 of the inspection unit 51 located next to it. Therefore, the support-side inclined surface 533a of each inspection unit 51 has the function of guiding the tablet 5 that has come into contact with it toward the imaging range side of a camera 531 other than the camera 531 related to it.

[0065] The lighting unit 54 is a unit in which the lighting mounting section 541, the lighting device 542, and the light shielding plate 543 are integrated. In this embodiment, the lighting device 542 constitutes the "irradiation means".

[0066] The lighting mounting section 541 constitutes the mounting portion of the lighting unit 54 to the PTP packaging machine 10 (for example, the outer frame of the PTP packaging machine 10), and is a plate-like shape extending vertically. The entire surface of the lighting mounting section 541 that is visible when viewed from above is inclined downward toward a predetermined outer edge (the outer edge from which objects will fall), and is configured as a lighting-side inclined surface 541a that prevents falling objects Rb from resting on it. The acute angle α4 (see Figure 10) of the angle that the lighting-side inclined surface 541a makes with respect to the horizontal plane is 50° or more.

[0067] Furthermore, since the lighting device 542 and the light-shielding plate 543 are attached to the PTP packaging machine 10 via the lighting mounting section 541, the lighting mounting section 541, the lighting device 542, and the light-shielding plate 543 (i.e., the lighting unit 54) are in a state of being suspended from the mounting surface 100. Combined with the fact that the camera unit 53 is suspended from the mounting surface 100 as described above, the entire inspection unit 51 is in a state of being suspended from the mounting surface 100.

[0068] The lighting device 542 illuminates the imaging range of the camera 531 (camera body 531a) on the installation surface 100 with predetermined light (for example, infrared light) from diagonally above. In this embodiment, the lighting device 542 is provided corresponding to one end of the PTP packaging machine 10 in the width direction and is configured to illuminate the installation surface 100 with predetermined light from one direction.

[0069] The light-shielding plate 543 covers the lighting device 542 from above, preventing light emitted from the lighting device 542 from directly entering the camera 531. The entire surface of the light-shielding plate 543 that is visible when viewed from above is a light-shielding plate side sloped portion 543a that is sloped downward toward a predetermined outer edge (the outer edge from which the object will fall) so that the falling object Rb does not rest on it. The acute angle α5 (see Figure 7) of the angle that the light-shielding plate side sloped portion 543a makes with respect to the horizontal plane is 50° or more. It is preferable that angles α1, α2, α3, α4, and α5 be larger values. Therefore, it is more preferable that angles α1, α2, α3, α4, and α5 each be 60° or more.

[0070] Next, the control device 52 will be described. The control device 52 is composed of a computer system including a CPU that executes predetermined calculation processes, a ROM that stores various programs and fixed value data, a RAM that temporarily stores various data when various calculation processes are executed, and peripheral circuits for these. As shown in Figure 11, the control device 52 functions as various functional units such as the main control unit 52a, lighting control unit 52b, camera control unit 52c, image acquisition unit 52d, reconstructed image data acquisition unit 52e, learning unit 52f, and determination unit 52g, as the CPU operates according to various programs. In this embodiment, the reconstructed image data acquisition unit 52e constitutes the "reconstructed image data acquisition means," and the determination unit 52g constitutes the "comparison means" and the "determination means."

[0071] However, the various functional units described above are realized through the cooperation of various hardware components such as the CPU, ROM, and RAM, and there is no need to clearly distinguish between functions realized in hardware and functions realized in software. Some or all of these functions may be realized by hardware circuits such as ICs.

[0072] Furthermore, the control device 52 is equipped with an input unit 521 consisting of a keyboard, mouse, touch panel, etc., a display unit 522 having a display screen such as a liquid crystal display, and a communication unit 526 capable of sending and receiving various data with the outside. The control device 52 also includes an inspection information storage unit 523, an AI storage unit 524, and an image data storage unit 525, each consisting of an HDD (Hard Disk Drive) or SSD (Solid State Drive), etc.

[0073] First, before describing the various functional units that make up the control device 52, we will explain the input unit 521, the display unit 522, the storage units 523 to 525, and the communication unit 526.

[0074] The input unit 521 is an input means for inputting information to the control device 52. The input unit 521 allows for the modification of various setting information stored in the control device 52.

[0075] The display unit 522 is configured to display various types of information stored in each of the storage units 523 to 525, for example. Therefore, the display unit 522 can display image data obtained by the camera 531, the results of a determination regarding the presence or absence of fallen objects Rb on the installation surface 100, and so on.

[0076] The inspection information storage unit 523 stores the determination result regarding the presence or absence of fallen objects Rb on the installation surface 100, as well as information indicating the location of the fallen objects Rb.

[0077] The AI ​​memory unit 524 stores the AI ​​(Artificial Intelligence) model 200 as an "identification means".

[0078] The image data storage unit 525 stores various image data, such as image data obtained by the camera 531.

[0079] The communication unit 526 is equipped with a wireless communication interface conforming to communication standards such as wired LAN (Local Area Network) or wireless LAN, and is configured to send and receive various data with the outside. For example, the judgment result made by the judgment unit 52g can be output to the outside (for example, the PTP packaging machine 10) via the communication unit 526.

[0080] Next, the various functional units that constitute the control device 52 will be described in detail.

[0081] The main control unit 52a is a functional unit that controls the entire installation surface inspection device 50 and is configured to send and receive various signals with other functional units such as the lighting control unit 52b and the camera control unit 52c.

[0082] The lighting control unit 52b is a functional unit that drives and controls the lighting device 542, and controls the lighting timing and other parameters based on command signals from the main control unit 52a.

[0083] The camera control unit 52c is a functional unit that drives and controls the camera 531, and controls the imaging timing and other parameters based on command signals from the main control unit 52a.

[0084] In this embodiment, the lighting control unit 52b and the camera control unit 52c perform a process in which, at regular intervals, light is irradiated from the lighting device 542 onto the installation surface 100, and the light reflected from the installation surface 100 is captured by the camera 531. Therefore, multiple inspection image data, which are image data obtained by the camera 531, are acquired within a predetermined period. These inspection image data are sent to the image acquisition unit 52d.

[0085] The image acquisition unit 52d is a functional unit for acquiring image data captured and obtained by the camera 51. The acquired image data is stored in the image data storage unit 525.

[0086] The learning unit 52f is a functional unit that uses training data to train the deep neural network 190 (hereinafter simply referred to as "neural network 190"; see Figure 12) and construct the AI ​​model 200.

[0087] As will be described later, the AI ​​model 200 in this embodiment is a generative model constructed by deep learning a neural network 190 using only image data relating to the installation surface 100 without falling objects Rb as training data, and has the structure of a so-called autoencoder.

[0088] Here, the structure of the neural network 190 will be explained with reference to Figure 12. Figure 12 is a schematic diagram conceptually showing the structure of the neural network 190. As shown in Figure 12, the neural network 190 has the structure of a convolutional auto-encoder (CAE), comprising an encoder unit 191 as an "encoding unit" that extracts feature quantities (latent variables) TA from the input image data GA, and a decoder unit 192 as a "decoding unit" that reconstructs image data GB from the feature quantities TA.

[0089] The structure of the convolutional autoencoder is well known, so a detailed explanation will be omitted. The encoder unit 191 has multiple convolutional layers 193, and in each convolutional layer 193, the result of a convolution operation using multiple filters (kernels) 194 on the input data is output as input data for the next layer. Similarly, the decoder unit 192 has multiple deconvolutional layers 195, and in each deconvolutional layer 195, the result of a deconvolution operation using multiple filters (kernels) 196 on the input data is output as input data for the next layer. Then, in the learning process described below, the weights (parameters) of each filter 194, 196 are updated.

[0090] Here, we will explain the learning process performed by the learning unit 52f when generating the AI ​​model 200. First, prior to the learning process, a large number of image data related to the installation surface 100 without any falling objects Rb are prepared as learning data. In this embodiment, the learning data includes image data actually obtained by the camera 531 when it photographs the installation surface 100, and various types of image data are prepared that have different elements that can affect the state of the installation surface 100, such as image data under different ambient light conditions (for example, image data obtained during the day or at night, image data obtained on a sunny day or on a rainy day), image data when the operator is working, and image data when the PTP packaging machine 10 is in operation or stopped.

[0091] Furthermore, instead of image data obtained by actually photographing the installation surface 100, virtual images of the installation surface 100 without falling objects (virtual image data) may be used as training data. In addition, the amount of data in the training data may be reduced by masking the structural parts of the PTP packaging machine 10 (e.g., the legs 10k) and parts where it is not necessary to check for the presence or absence of falling objects Rb.

[0092] As shown in Figure 13, in the learning process, first, in step S201, an untrained neural network 190 is prepared. For example, a neural network 190 that has been pre-stored in a predetermined storage device is read. Alternatively, the neural network 190 is constructed based on network configuration information (for example, the number of layers in the neural network and the number of nodes in each layer) stored in the storage device.

[0093] Next, in step S202, reconstructed image data is acquired. That is, pre-prepared training data is provided as input data to the input layer of the neural network 190. Then, the reconstructed image data output from the output layer of the neural network 190 is acquired.

[0094] In the following step S203, the training data is compared with the reconstructed image data output by the neural network 190 in step S202, and it is determined whether the error is sufficiently small (whether it is below a predetermined threshold).

[0095] If the error is sufficiently small, step S205 determines whether the learning process termination conditions are met. For example, if a certain number of consecutive affirmative determinations are made in step S203 without going through the process of step S204 described later, or if learning using all of the prepared learning data is repeated a certain number of times, it is determined that the termination conditions are met. If the termination conditions are met, the neural network 190 and its learning information (updated parameters, etc., described later) are stored in the AI ​​memory unit 524 as the AI ​​model 200, and the learning process is terminated.

[0096] On the other hand, if the termination condition is not met in step S205, the process returns to step S202 and the neural network 190 is trained again.

[0097] Furthermore, if the error is not sufficiently small in step S203, the network update process (training of the neural network 190) is performed in step S204, and then the process returns to step S202, and the above series of processes is repeated.

[0098] In the network update process of step S204, the weights (parameters) of each filter 194, 196 in the neural network 190 are updated to more appropriate values, using a known learning algorithm such as backpropagation, so that the loss function representing the difference between the training data and the reconstructed image data is minimized. For example, BCE (Binary Cross-entropy) can be used as the loss function.

[0099] By repeatedly performing steps S202 to S204, the neural network 190 minimizes the error between the training data and the reconstructed image data, resulting in the output of more accurate reconstructed image data.

[0100] The final AI model 200 is image data obtained by the camera 531 (inspection image data), and when image data relating to the installation surface 100 without fallen objects Rb is input, it generates reconstructed image data that is almost identical to the input image data. On the other hand, when image data obtained by the camera 531 (inspection image data), relating to the installation surface 100 with fallen objects Rb is input, the AI ​​model 200 generates reconstructed image data that is almost identical to the input image data, with the noise portion (the portion corresponding to the fallen objects Rb) removed. In other words, when there are fallen objects Rb on the installation surface 100, the reconstructed image data relating to the installation surface 100 is a virtual image data relating to the installation surface 100, assuming that there are no fallen objects Rb.

[0101] Furthermore, when the AI ​​model 200 receives image data (inspection image data) obtained by the camera 531, which pertains to the installation surface 100 containing foreign objects other than the fallen object Rb (for example, film fragments or dust), it generates reconstructed image data that closely matches the input image data, with the noise portion (the portion corresponding to the foreign object) removed.

[0102] Furthermore, it is not necessary to perform the above learning process each time the control device 52 is manufactured. The neural network 190 and its learning information (updated parameters, etc.) may be acquired in advance and stored as the AI ​​model 200 in the AI ​​memory unit 524 of the control device 52.

[0103] The reconstructed image data acquisition unit 52e inputs the image data obtained by the camera 531 into the AI ​​model 200 and acquires the reconstructed image data as reconstructed image data. More specifically, the reconstructed image data acquisition unit 52e provides the inspection image data obtained by the camera 531 as input data to the input layer of the neural network 190. The reconstructed image data acquisition unit 52e then acquires the image data output from the output layer of the neural network 190 as reconstructed image data.

[0104] The determination unit 52g is a functional unit that performs a determination process regarding the presence or absence of fallen objects Rb on the installation surface 100. The determination unit 52g compares the inspection image data obtained by the camera 531 with the reconstructed image data reconstructed by inputting the inspection image data into the AI ​​model 200, and determines the presence or absence of fallen objects Rb based on the comparison result.

[0105] Referring to Figures 14-16, the determination process performed by the determination unit 52g will be explained in more detail. The determination process is performed when a predetermined number or more of inspection image data are newly acquired.

[0106] In the judgment process, first, in step S301, the reconstructed image data acquisition unit 52e inputs the inspection image data obtained by the camera 531 into the AI ​​model 200 to acquire the reconstructed image data.

[0107] Next, in step S302, the inspection image data obtained by the camera 531 is compared with the reconstructed image data obtained by the reconstructed image data acquisition unit 52e, and the difference in brightness for each pixel of the two sets of data is calculated. Subsequently, pixels whose difference is not within a predetermined tolerance range are identified as defective pixels.

[0108] Next, in step S303, the area of ​​the connected component of the defective pixels (defective area) is calculated.

[0109] Next, in step S304, it is determined whether the largest area among the calculated areas is greater than or equal to a predetermined threshold. In other words, it is determined whether the defective area is outside the acceptable range. If the largest area is greater than or equal to the threshold, it is determined in step S305 that there is a fallen object Rb or foreign object on the installation surface 100, and that an "abnormal portion" exists in the inspection image data. Then, the portion corresponding to the connected component of the defective pixels in the inspection image data is identified as the "abnormal portion".

[0110] On the other hand, if the maximum area falls below the threshold, in step S306, it is determined that there are no "abnormal parts" in the inspection image data. Note that the above determination method is just one example; for example, if even one defective pixel is present, it may be determined that there are "abnormal parts".

[0111] Then, after determining whether or not there are any "abnormal parts," in step S307, it is determined whether or not the processing in steps S301 to S306 (i.e., the processing to determine whether or not there are any abnormal parts) has been performed on all of the newly acquired inspection image data. If the above processing has been performed on all of the new inspection image data, the process proceeds to step S308. On the other hand, if there is any new inspection image data that has not been processed, the process returns to step S301 and the above processing is performed on the inspection image data for which the processing has not been completed.

[0112] After performing the above processing on all newly acquired inspection image data, step S308 determines whether there are multiple inspection image data in which the presence of "abnormal parts" is detected (i.e., whether the processing in step S305 has been performed multiple times).

[0113] If there are multiple inspection image data in which the presence of an "abnormal part" is confirmed (step S308: Yes), in step S309, it is determined whether or not there has been a change in the position of the "abnormal part" among the multiple inspection image data. If the position of the "abnormal part" has not changed, it is estimated that the "abnormal part" is a fallen object Rb, and in step S310, it is determined that there is a fallen object Rb on the installation surface 100. Furthermore, in step S311, information indicating the position of the fallen object Rb is obtained (for example, information to identify the camera 531 that captured the image of the fallen object Rb among the multiple cameras 531, or the coordinates of the "abnormal part" in the inspection image data).

[0114] Then, in step S312, the determination result that there is a fallen object Rb on the installation surface 100 and information indicating the position of the fallen object Rb are stored in the inspection information storage unit 523. Note that, as information indicating the position of the fallen object Rb, inspection image data with the "abnormal part" highlighted, or inspection image data in which the "abnormal part" is replaced with a predetermined mark, may also be stored in the inspection information storage unit 523.

[0115] On the other hand, if the position of the "abnormal part" has changed (step S308; No), it is presumed that the "abnormal part" is a foreign object that moves relatively easily (for example, a piece of film or dust), and in step S313, it is determined that there was a foreign object on the installation surface 100, but no fallen object Rb. Then, in step S312, the determination result is stored in the inspection information storage unit 523. At this time, information indicating that a foreign object was present on the installation surface 100 and information indicating the position of the foreign object may also be stored in the inspection information storage unit 523 along with the determination result.

[0116] Furthermore, if a negative determination is made in step S308, step S314 determines whether there is only one inspection image data in which the presence of the "abnormal part" is confirmed. If there is only one inspection image data in which the presence of the "abnormal part" is confirmed (step S314; Yes), it is presumed that the "abnormal part" is either a fallen object Rb or a foreign object, and step S315 determines that there is a fallen object Rb or a foreign object on the installation surface 100. Furthermore, in step S316, the determination result is stored in the inspection information storage unit 523. Note that at this time, information indicating the location of the fallen object Rb or foreign object may also be stored in the inspection information storage unit 523 along with the determination result.

[0117] On the other hand, if a negative determination is made in step S314, that is, if no "abnormal part" is found in any of the inspection image data, then in step S317, it is determined that there are no fallen objects Rb on the installation surface 100. Then, in step S316, the determination result that there are no fallen objects Rb on the installation surface 100 is stored in the inspection information storage unit 523. At this time, information indicating that no foreign matter was found on the installation surface 100 may also be stored in the inspection information storage unit 523 along with the determination result.

[0118] As described in detail above, according to this embodiment, the entire surface of the horizontal support portion 533 and the camera 531 that is visible when viewed from above is a sloped portion 533a, 531c that inclines downward toward a predetermined outer edge. Therefore, even if a falling object Rb bounces off the installation surface 100 and ends up on the horizontal support portion 533 or the camera 531, the falling object Rb is more likely to slide down the sloped portion 533a, 531c and fall onto the installation surface 100. Consequently, it becomes less likely for the falling object Rb to land on the horizontal support portion 533, etc., and thus it is possible to more reliably prevent the falling object Rb from remaining outside the imaging range of the camera 531. As a result, the camera 531 can more reliably image the falling object Rb, and consequently, the presence or absence of the falling object Rb on the installation surface 100 can be confirmed more appropriately. In particular, in this embodiment, since the acute angles α3 and α1 of the angles made between the inclined surfaces 533a and 531c with respect to the horizontal plane are 50° or more, it is possible to more reliably prevent falling objects Rb from resting on the horizontal support portion 533, etc.

[0119] Furthermore, because the horizontal support section 533 is attached to the PTP packaging machine 10, the entire camera 531 and the components supporting the camera 531 (horizontal support section 533 and mounting section 532) are suspended above the installation surface 100. Therefore, compared to the case where the legs supporting the horizontal support section 533 are erected on the installation surface 100, the blind spots in the imaging range of the camera 531 can be reduced. This makes it possible to more appropriately check for the presence or absence of fallen objects Rb on the installation surface 100.

[0120] In addition, the horizontal support section 533 has a cantilever structure in which only its base end is attached to the PTP packaging machine 10, and the camera 531 is attached to the side of the tip of the horizontal support section 533. Therefore, the camera 531 can be positioned higher than when the camera 531 is attached to the lower surface of the horizontal support section 533. This allows the imaging range of the camera 531 to be wider, and consequently, it becomes possible to more appropriately check for the presence or absence of fallen objects Rb on the installation surface 100.

[0121] Furthermore, the entire surface of the mounting portion 532 that is visible when viewed from above is a mounting portion side sloped surface 532a that slopes downward toward a predetermined outer edge. Therefore, it is less likely that a falling object Rb will land on the mounting portion 532, preventing the camera 531 from capturing an image of the falling object Rb. This allows for more accurate confirmation of the presence or absence of falling objects Rb on the installation surface 100. In particular, in this embodiment, the acute angle α2 of the angle that the mounting portion side sloped surface 532a makes with respect to the horizontal plane is 50° or more, so it is possible to more reliably prevent falling objects Rb from landing on the mounting portion 532.

[0122] Furthermore, since the lighting mounting portion 541 has a lighting-side inclined surface portion 541a, it is possible to effectively suppress the falling object Rb from resting on the lighting mounting portion 541. In particular, in this embodiment, the acute angle α4 of the angle that the lighting-side inclined surface portion 541a makes with respect to the horizontal plane is set to 50° or more, so it is possible to more reliably prevent the falling object Rb from resting on the lighting mounting portion 541.

[0123] In addition, the top portion 533b of the horizontal support portion 533 is not angular but curved. Therefore, the falling object Rb can be caught on the top portion 533b, more reliably preventing the falling object Rb from resting on the horizontal support portion 533. This makes it possible to more appropriately check for the presence or absence of falling objects Rb on the installation surface 100. Furthermore, it is possible to more reliably prevent damage to the falling object Rb caused by contact with the top portion 533b.

[0124] In addition, in this embodiment, the lighting device 542 can illuminate at least the imaging range of the camera 531 on the installation surface 100. Therefore, the camera 531 can image the fallen object Rb on the installation surface 100 in a clearer state. Furthermore, when visually inspecting the installation surface 100, the fallen object Rb can be found more easily.

[0125] Furthermore, the light-shielding plate 543 prevents light emitted from the illumination device 542 from directly entering the camera 531. This makes it possible to more reliably prevent problems such as halation in the image data obtained by the camera 531.

[0126] In addition, the entire surface of the light-shielding plate 543 that is visible when viewed from above is a light-shielding plate side sloped portion 543a that slopes downward toward a predetermined outer edge. Therefore, it is less likely that a falling object Rb will land on the light-shielding plate 543, making it impossible for the camera 531 to image the falling object Rb. This allows for a more appropriate confirmation of the presence or absence of falling objects Rb on the installation surface 100. In particular, in this embodiment, the acute angle α5 of the angle that the light-shielding plate side sloped portion 543a makes with respect to the horizontal plane is set to 50° or more, so that it is possible to more reliably prevent falling objects Rb from landing on the light-shielding plate 543.

[0127] Furthermore, the determination unit 52g compares the inspection image data obtained by the camera 531 with the reconstructed image data reconstructed by inputting the inspection image data into the AI ​​model 200, and determines the presence or absence of fallen objects Rb on the installation surface 100 based on the comparison result. Therefore, compared to determining the presence or absence of fallen objects Rb by visually inspecting the image data, the presence or absence of fallen objects Rb can be confirmed more accurately and quickly.

[0128] Furthermore, the conditions of the installation surface 100 (e.g., ambient light) can be matched in both image data being compared. This allows for more accurate confirmation of the presence or absence of fallen objects Rb on the installation surface 100.

[0129] In addition, in this embodiment, the determination unit 52g does not simply determine the presence or absence of a fallen object Rb based on whether or not there is an "abnormal part" in the inspection image data, but rather determines the presence or absence of a fallen object Rb based on whether or not there is a change in the position of the "abnormal part". Therefore, the presence or absence of a fallen object Rb on the installation surface 100 can be determined with even greater accuracy.

[0130] Furthermore, the embodiment is not limited to the description above, and may be implemented as follows, for example. Of course, other applications and modifications not exemplified below are also possible.

[0131] (a) In the above embodiment, the lighting device 542 is provided corresponding to one end of the PTP packaging machine 10 in the width direction and is configured to irradiate the installation surface 100 with light from one direction. Alternatively, the lighting device 542 may be provided corresponding to both ends of the PTP packaging machine 10 in the width direction, so that predetermined light is irradiated onto the installation surface 100 from two different directions. Alternatively, predetermined light may be irradiated onto the installation surface 100 from all four sides of the installation surface 100.

[0132] (b) As shown in Figure 17, the support-side inclined surface 533a may be provided with a cushioning portion 535 capable of cushioning the impact applied to the falling object Rb when it comes into contact with the falling object Rb. The cushioning portion 535 corresponds to a "cushioning means". The cushioning portion 535 can be made of a material capable of absorbing impact, such as soft rubber, felt, urethane, or an elastic member.

[0133] By providing the buffer portion 535, the impact applied to the falling object Rb from the support-side slope portion 533a can be buffered. Therefore, it is possible to more reliably prevent the falling object Rb from moving significantly (for example, bouncing) due to the impact applied from the support-side slope portion 533a. As a result, the camera 531 can more reliably image the falling object Rb, and the presence or absence of the falling object Rb on the installation surface 100 can be confirmed more appropriately.

[0134] Furthermore, by providing the buffer portion 535, damage (such as cracking) to the fallen object Rb due to impact can be more reliably prevented, making it easier to find the fallen object Rb in its original shape. Therefore, the recovery of the fallen object Rb can be made easier. In addition, preventing the falling object Rb from cracking is particularly effective when the tablet 5 (fallen object Rb) contains special drug components.

[0135] Furthermore, the buffer portion 535 may also be applied to the camera-side slope portion 531c, the mounting portion-side slope portion 532a, the lighting-side slope portion 541a, and the light-shielding plate-side slope portion 543a as described above.

[0136] (c) In the above embodiment, the horizontal support portion 533 has a cantilever structure in which only its base end is attached to the PTP packaging machine 10. In contrast, as shown in Figures 18 and 19, mounting portions 532 may be provided at both ends of the horizontal support portion 533, and the horizontal support portion 533 may be attached to the PTP packaging machine 10 via the mounting portions 532, thereby giving the horizontal support portion 533 a double-support structure. In this case, the camera 531 may be attached to the lower surface of the horizontal support portion 533.

[0137] (d) In the above embodiment, the top 533b of the horizontal support portion 533 is curved in a convex shape towards the upward, but the top 533b may be angular.

[0138] (e) In the above embodiment, the cover 531b of the camera 531 is configured such that the entire surface visible when viewed from above has the same shape as the side surface of a cone, but the shape of the cover 531b is not limited to this. For example, as shown in Figures 20 and 21, the cover 531b may be configured such that the entire surface visible when viewed from above has the same shape as the side surface of a pyramidal pyramid. Also, when the camera 531 is attached to the lower surface of the horizontal support portion 533, the cover 531b may be configured such that the entire surface visible when viewed from above has the same shape as the side surface of a frustum of a cone or a frustum of a pyramidal pyramid.

[0139] (f) In the above embodiment, the determination unit 52g estimates whether the "abnormal part" is a fallen object Rb or a foreign object based on the change in the position of the "abnormal part". However, without performing such an estimation, the "abnormal part" may simply be uniformly estimated to be a fallen object Rb.

[0140] (g) In the above embodiment, the presence or absence of a falling object Rb is determined by the determination unit 52g, but the presence or absence of a falling object Rb may also be determined by visually checking the image data obtained by the camera 531 by an operator or the like. Alternatively, the presence or absence of a falling object Rb may be determined without using the AI ​​model 200.

[0141] (h) The learning unit 52f may have a function to perform additional training of the AI ​​model 200 using image data obtained by the camera 531.

[0142] (i) The configuration of the AI ​​model 200 (neural network 190) as a "discrimination means" and its learning method are not limited to the above embodiment. For example, when performing the learning process of the neural network 190 or the acquisition process of reconstructed image data, the configuration may include normalization or other processing on various data as needed. Also, the structure of the neural network 190 is not limited to that shown in Figure 12, and for example, a pooling layer may be provided after the convolutional layer 193. Of course, the number of layers of the neural network 190, the number of nodes in each layer, and the connection structure of each node may also be different configurations.

[0143] Furthermore, in the above embodiment, the AI ​​model 200 (neural network 190) is a generative model having the structure of a convolutional autoencoder (CAE), but it is not limited to this, and may be a generative model having the structure of a different type of autoencoder, such as a variational autoencoder (VAE).

[0144] Furthermore, in the above embodiment, the neural network 190 is trained using the backpropagation method, but the system is not limited to this, and various other learning algorithms may be used for training.

[0145] In addition, the neural network 190 may be composed of a dedicated AI processing circuit, such as a so-called AI chip. In that case, only learning information such as parameters may be stored in the AI ​​memory unit 524, and the dedicated AI processing circuit may read this and set it in the neural network 190 to configure the AI ​​model 200.

[0146] (j) In the above embodiment, the article is given as a tablet 5, but the article is not limited to a tablet 5. Therefore, the article may be, for example, an electronic component.

[0147] Furthermore, in the above embodiment, the tablet 5 as an article has a circular shape in plan view, but the shape of the article may be other than a circular shape in plan view. However, articles with a circular cross-section, such as tablets with a circular shape in plan view, cylindrical capsules, or spherical pills, tend to roll easily on inclined surfaces (inclined surfaces 533a, 531c, etc.), and the effect of providing inclined surfaces 533c, 531, etc. is more pronounced. Therefore, the installation surface inspection device 50 is particularly suitable for use with devices that have a function of transporting articles with a circular cross-section.

[0148] (k) In the above embodiment, the installation surface inspection device 50 is applied to the PTP packaging machine 10, but the device to which the installation surface inspection device 50 is applied can be any device that has the function of transporting a predetermined article (i.e., a device to which the article may fall during transport), and is not limited to the PTP packaging machine 10.

[0149] 5... Tablet (article), 10... PTP packaging machine (device), 50... Installation surface inspection device, 52e... Reconstructed image data acquisition unit (reconstructed image data acquisition means), 52g... Judgment unit (comparison means, judgment means), 191... Encoder unit (encoding unit), 192... Decoder unit (decoding unit), 200... AI model (identification means), 531... Camera, 531c... Camera side sloped part (slope), 532... Mounting part, 532a... Mounting part side sloped part, 533... Horizontal support part, 533a... Support part side sloped part (slope), 533b... Top, 535... Cushioning part (cushioning means), 542... Lighting device (irradiation means), 543... Light shielding plate, 543a... Light shielding plate side sloped part, Rb... Falling object.

Claims

1. An installation surface inspection device for inspecting the presence or absence of fallen objects, which are said to be the said objects, on the installation surface of a device equipped with the function of transporting a predetermined object, comprising: a horizontal support portion extending horizontally and at least a portion of which is installed below the device; and a camera attached to the horizontal support portion and capable of imaging the installation surface, wherein at least the entire surface of the horizontal support portion and the camera that is visible when viewed from above is an inclined portion that slopes downward toward a predetermined outer edge.

2. The installation surface inspection device according to claim 1, characterized in that the horizontal support portion is attached to the device so that the entire camera and the components supporting the camera are floating above the installation surface.

3. The installation surface inspection device according to claim 2, characterized in that only the base end of the horizontal support portion is attached to the device, and the camera is attached to the side surface of the tip of the horizontal support portion.

4. The installation surface inspection device according to claim 2, wherein the device is interposed between the horizontal support portion and the device and comprises a mounting portion for attaching the horizontal support portion to the device, and the entire surface of the mounting portion that is visible when viewed from above is a mounting portion side slope that slopes downward toward a predetermined outer edge.

5. The installation surface inspection device according to claim 1, characterized in that the acute angle among the angles that the inclined surface makes with respect to the horizontal surface is 50° or more.

6. The installation surface inspection device according to claim 1, characterized in that the horizontal support portion has a shape in which two inclined portions, which are oriented in opposite directions, are connected at the top, and the top portion has a curved surface shape that is convex upward.

7. The installation surface inspection device according to claim 1, comprising: an illumination means located below the camera and irradiating a predetermined light onto at least the imaging range of the camera on the installation surface; and a light-shielding plate that covers the illumination means from above to prevent the light irradiated from the illumination means from directly entering the camera, wherein the entire surface of the light-shielding plate that is visible when viewed from above is a light-shielding plate side slope portion that slopes downward toward a predetermined outer edge.

8. An installation surface inspection device according to claim 1, comprising: an identification means generated by training a neural network having an encoding unit for extracting feature quantities from input image data and a decoding unit for reconstructing image data from the feature quantities, using only image data relating to the installation surface without fallen objects as training data; a reconstructed image data acquisition means capable of acquiring reconstructed image data as reconstructed image data by inputting inspection image data obtained by the camera to the identification means; a comparison means capable of comparing the inspection image data and the reconstructed image data; and a determination means capable of determining the presence or absence of fallen objects on the installation surface based on the comparison results by the comparison means.

9. The installation surface inspection device according to claim 1, characterized in that the inclined surface has a buffering means capable of buffering the impact applied to the falling object when it comes into contact with the falling object.

Images