Can lid detection device and can lid detection method

Abstract

To provide a can lid detection device and a can lid detection method capable of detecting a can body having a can lid with an abnormal tab from a group of can bodies.SOLUTION: A can lid detection device 1 for detecting a can body C having a can lid C1 with an abnormal tab from a group of can bodies includes: storage means 20 for storing in advance reference image data of a can lid C1; and determination means 50 for determining whether or not there is an abnormality in a tab based on the degree of correlation between input image data of a can lid and the reference image data, in the tab of the can lid C1 and a vicinity area, which is in the vicinity of the tab.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a can lid detection device and a can lid detection method. [Background technology]

[0002] Currently, braille is stamped on the lids of cans so that the type of beverage (specifically, whether it is an alcoholic beverage or not) can be identified even by the visually impaired. The lids of cans containing alcoholic beverages are stamped with braille indicating "alcohol" or "sake," while the lids of cans containing non-alcoholic beverages are not stamped with braille. The Braille imprinted on the can lid plays an important role for visually impaired people in identifying the type of beverage inside.

[0003] When sealing a beverage in a can body, the beverage is filled into a can body (which may be a two-piece can in which the can body and bottom lid are integrated, or a three-piece can in which they are separate) equipped with a bottom lid, and then the can lid is placed on top and the ends of the can body and the can lid are seamed. This type of work is usually carried out on a production line, but when switching products (for example, from an alcoholic beverage to a non-alcoholic beverage), there is a risk that the can lid for the later product will be mistakenly seamed onto the product before that timing, or the can lid for the earlier product will be mistakenly seamed onto the product after that timing. There is also a risk that the wrong type of can lid will be set on the production line due to the wrong product type.

[0004] On the production line, cans may be washed with water after the seaming process, and condensation may form on the surface of the can depending on the temperature of the beverage contained in the can and the ambient temperature and humidity. If water droplets adhere to the surface of the can lid due to such cleaning work or condensation, it becomes extremely difficult to determine whether or not the can lid has Braille engraved on it. Therefore, in a situation where water droplets adhere to the surface of a can lid, if a can body with a different type of can lid is mistakenly found as described above, it is impossible for conventional inspection devices to properly detect it, and it was expected that there would be many false detections. Therefore, the applicant proposed the following technique.

[0005] Specifically, as shown in Patent Document 1, this is a can lid detection device that detects can bodies with different types of can lids mixed in a group of can bodies, and is equipped with a memory means that stores information on the position of a characteristic area on the can lid and the positions of multiple predetermined areas defined relative to the position of the characteristic area, a feature detection means that detects the position of the characteristic area of ​​the can lid in image data of the can lid, an area identification means that identifies multiple predetermined areas in the image data based on the position of the characteristic area detected by the feature detection means and the information stored in the memory means, and a judgment means that judges whether a protruding shape exists in all of the multiple predetermined areas identified by the area identification means. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Publication No. 2019-21233 Summary of the Invention [Problem to be solved by the invention]

[0007] According to Patent Document 1, can bodies with different types of can lids that are mixed in a group of can bodies can be detected with extremely high accuracy.

[0008] On the other hand, after examining problems on the production line, the inventors confirmed the occurrence of can lids with abnormal tabs, such as tabs that rotate on the lid, which is a different problem from the "mixing of can bodies with different types of can lids" described in Patent Document 1, and therefore thought that can bodies with can lids with such abnormal tabs should be detected. In addition, since the invention of Patent Document 1 makes its judgment based on the tab, which is a characteristic part of the can lid, the inventors thought that if it were possible to detect the presence of an abnormality in the tab, the possibility of false detection in the invention of Patent Document 1 could be further reduced.

[0009] Therefore, an object of the present invention is to provide a can lid detection device and a can lid detection method that can detect can bodies with can lids having abnormal tabs from a group of can bodies. [Means for solving the problem]

[0010] The above problems can be solved by the following means. (1) Tab is rotating abnormality of A can lid detection device that detects a can body with a certain can lid from a group of can bodies, A normal can lid with the tab not rotating A storage means for storing reference image data in advance, and a tab detection means for detecting a tab based on a degree of correlation between input image data of the can lid and the reference image data in a tab of the can lid and a region adjacent to the tab. is rotating A determination means for determining whether or not there is an abnormality is provided. The vicinity region is a region that includes the tab and extends from the outer edge of the tab to the periphery thereof, and the area of ​​the vicinity region is 1.1 times or more the area of ​​the tab. Can lid detection device. ( 2 The storage means stores information on the position of a tab on a can lid and the positions of a plurality of predetermined areas defined relative to the position of the tab, and the storage means further includes a predetermined area specifying means for specifying the plurality of predetermined areas in the input image data of the can lid based on the position of the tab on the can lid in the input image data of the can lid and the information stored in the storage means, and the determination means determines whether or not a protruding shape exists in all of the plurality of predetermined areas specified by the predetermined area specifying means. 1 to The can lid detection device described. ( 3 )tab is rotating abnormality of A can lid detection method for detecting a can body having a certain can lid from a group of can bodies, comprising: image data of the can lid obtained by capturing an image of the tab of the can lid and a region in the vicinity of the tab; A normal can lid with the tab not rotating Based on the degree of correlation with the reference image data, is rotatingA tab determination process for determining whether or not there is an abnormality. The vicinity region is a region that includes the tab and extends from the outer edge of the tab to the periphery thereof, and the area of ​​the vicinity region is 1.1 times or more the area of ​​the tab. Can lid detection method. ( 4 a predetermined area specifying step of specifying a plurality of predetermined areas in the image data based on the position of the tab in the image data of the can lid obtained by imaging, and information on the position of the tab on the can lid and the positions of a plurality of predetermined areas determined relative to the position of the tab; and a protruding shape determining step of determining whether or not a protruding shape exists in all of the plurality of predetermined areas specified in the predetermined area specifying step. 3 The can lid detection method according to claim 1. [Effects of the Invention]

[0011] According to the can lid detection device and can lid detection method of the present invention, can bodies having can lids with abnormal tabs can be appropriately detected from a group of can bodies. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a block diagram showing the configuration of a can lid detection device according to an embodiment of the present invention; [Figure 2] This is a schematic diagram of image data of a can lid in a correct state (a state in which there is no abnormality in the tab) with the tab not rotated at all. [Figure 3] FIG. 10 is a schematic diagram of image data of a can lid in a state where the tab is rotating and there is an abnormality in the tab. [Figure 4] 10 is a schematic diagram of image data of a can lid on which Braille is stamped. FIG. [Figure 5] FIG. 10 is a schematic diagram of image data of a can lid on which no Braille is stamped. [Figure 6] 10 is a schematic diagram of image data of a can lid having no Braille imprinted thereon and having water droplets on its surface; FIG. [Figure 7] 3 is a flowchart of a can lid detection method according to the present embodiment (the operation of the can lid detection device according to the present embodiment). DETAILED DESCRIPTION OF THE INVENTION

[0013] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a can lid detection device and a can lid detection method according to the present invention will be described with reference to the drawings.

[0014] [Configuration of can lid detection device] First, the configuration of the can lid detection device according to this embodiment will be described with reference to FIG. 1 (and also with reference to FIGS. 2 to 6 as appropriate). The can lid detection device 1 is a device that detects can bodies C having can lids C1 with abnormal tabs from a group of can bodies C based on image data of the can lids C1 captured and input by an imaging device 2 such as a camera. The can lid detection device 1 includes an image data input means 10, a memory means 20, a feature detection means 30, a nearby area identification means 40a, a predetermined area identification means 40b, a judgment means 50, and a judgment result output means 60.

[0015] In the following, we will assume that the tabs are properly attached to the can lids of a group of can bodies C that are continuously transported on the production line, and that these can bodies C contain non-alcoholic beverages. In other words, we will explain the case where the "can lids" of the majority of can bodies C in the group of can bodies C have tabs that are properly attached without rotating and are can lids C1b and C1c (Figures 5 and 6) that do not have Braille stamped on them, and the "can lid with an abnormal tab" to be detected is can lid C1r with a rotated tab (Figure 3), and the "different can lid" to be detected is can lid C1a with Braille stamped on it (Figures 2 and 4). Alcoholic beverages are beverages with an alcohol content of 1% or more (beer, happoshu, other brewed alcoholic beverages, liqueurs, etc.), while non-alcoholic beverages are beverages with an alcohol content of less than 1% (soft drinks, non-alcoholic beverages, etc.).

[0016] (Image data input means) The image data input means 10 is a means for inputting image data of the can lid C1 via an imaging device 2 such as a camera. Then, the image data input means 10 outputs the input image data of the can lid C1 to the characteristic part detection means 30.

[0017] (memory means) The storage means 20 is a means for storing information used by the nearby area specifying means 40a, the predetermined area specifying means 40b, and the determination means 50. The information stored in the storage means 20 is read out by the nearby area specifying means 40a, the predetermined area specifying means 40b, and the determination means 50.

[0018] (Storage means: information read by the neighboring area specification means 40a) The information read from the storage means 20 by the nearby area specifying means 40a is information on the position of the characteristic part (tab) on the can lid, and information on the position of the characteristic part and the nearby area (characteristic part + vicinity) that is near the characteristic part.

[0019] Specifically, the information on the "position of the characteristic portion" is information on the location (area) of the tab C2 within the neighboring area R0 relative to the can lid C1 (area R3 surrounding the can lid C1), as shown in Figures 2 to 6. Based on this information, the position of the tab can be identified on the image data of the can lid C1.

[0020] Furthermore, the information on the "position of the nearby region" is position information on the characteristic part and the nearby region R0 that is near the characteristic part, and is determined relative to the position of the tab C2. Based on the position information on this nearby region R0 or the position information on the characteristic part (tab C2), the image data of the can lid C1 can be rotated within the region R3 to align the orientation, and the nearby region R0 on the image data can be identified. Since the nearby area R0 is an area for determining abnormalities in the tab, such as whether the tab has rotated, it is not sufficient for the area to be approximately the same as the area occupied by the shape of tab C2; it is an area that includes tab C2 but is larger than tab C2. The size of the vicinity region R0 is preferably 1.1 times or more, more preferably 1.3 times or more, 1.5 times or more, or 1.56 times or more of the area of ​​the tab C2 in order to accurately determine whether the tab is abnormal. On the other hand, the upper limit of the vicinity region R0 is not particularly limited, and may be, for example, 4 times or less, 3 times or less, or 2 times or less. Here, the area of ​​the tab C2 specifically refers to the area of ​​the region surrounded by the outer edge of the tab C2, and does not exclude areas such as holes for holding fingers and holes around the rivets. The nearby region R0 is a region on the can lid C1 that includes the tab C2 and extends from the outer edge of the tab C2 to the periphery thereof, and may be, for example, a rectangle as shown in FIG.

[0021] (Storage means: information read by the predetermined area specifying means 40b) The information read from the storage means 20 by the predetermined area specifying means 40b is information on the position of the characteristic part on the can lid and information on the positions of a plurality of predetermined areas defined relative to the position of the characteristic part.

[0022] In detail, the information on the "position of the characteristic part" read from the storage means 20 by the predetermined area identification means 40b is the same as the information on the "position of the characteristic part" read from the storage means 20 by the neighboring area identification means 40a described above.

[0023] The information on the "positions of the plurality of predetermined regions" is the position information of the region R2 including the protruding region of each Braille B stamped on the can lid C1 to indicate that the can C contains an alcoholic beverage, and is determined relative to the position of the characteristic part (the position of the tab C2). Based on this information, the region R2 can be identified on the aligned image data. Each of the multiple predetermined areas R2 may be an area containing Braille B, but from the viewpoint of reducing the occurrence of erroneous detection due to misalignment of the image data of the can lid C1, the area of ​​each predetermined area R2 is preferably 1.5 times or more and 2.0 times or less the area of ​​each Braille B.

[0024] (Storage means: information read by determination means 50) The information read out from the storage means 20 by the determination means 50 is "reference image data" and "information for determining the degree of correlation" between the reference image data and the input image data. In detail, the "reference image data" is image data of a can lid in which the tab has not rotated at all, and is image data of a can lid in which the tab is normally installed as shown in Figure 2. The "information for determining the degree of correlation" is a threshold value of the correlation value between the reference image data and the input image data in the nearby region R0, for example, 85 or more, 87 or more, 90 or more, 93 or more, 93.5 or more, or 95 or more.

[0025] The information read out from the storage means 20 by the determination means 50 is information that serves as a criterion for determining whether or not the protruding shape in each predetermined region R2 on the image data is a Braille character. Specifically, this information is data on the area value A2 (A1≧A2) which is smaller than the area value A1 of each Braille character in the image data. Note that the area value A2 is preferably 0.3×A1 or more and 0.5×A1 or less to eliminate the influence of water droplets with area values ​​smaller than each Braille character. By storing this area value A2 in the memory means 20, the determination means 50 described below can determine that Braille is present if a protruding shape with an area value larger than the area value A2 exists in each specified region R2, and can determine that no Braille is present if a protruding shape with an area value smaller than the area value A2 exists (determining that only small water droplets or the like are present).

[0026] (Feature detection means) The characteristic part detection means 30 is a means for detecting the position of the tab (characteristic part) of the can lid C1 in the image data of the can lid C1. There are no particular limitations on the method for detecting the position of the tab by the characteristic part detection means 30, but examples include a method of identifying the position based on the color shading or brightness in the image data. Then, the characteristic part detection means 30 outputs information on the detected position of the tab to the nearby area specification means 40a and the predetermined area specification means 40b.

[0027] (Means for identifying nearby areas) The nearby area identification means 40a is a means for identifying nearby areas in the input image data based on the position of the tab (feature part) detected by the feature part detection means 30 and the information stored in the storage means 20. 2 and 3, the nearby region specifying means 40a rotates the image data within the region R3 so that the position of the tab on the image data of the can lid C1 matches the position information (information on the "position of the characteristic part") of the storage means 20, thereby aligning the orientation of the image data. Then, the nearby region specifying means 40a specifies the nearby region R0 on the image data based on the information (information on the "position of the nearby region R0") of the storage means 20. Then, the neighboring area specifying means 40 a outputs information on the specified neighboring area to the determining means 50 .

[0028] (Predetermined area identification means) The predetermined area identification means 40b is a means for identifying a plurality of predetermined areas in the image data based on the positions of the tabs (characteristic features) detected by the characteristic feature detection means 30 and the information stored in the storage means 20. 4 to 6, the predetermined area specifying means 40b rotates the image data within the area R3 so that the position of the tab on the image data of the can lid C1 matches the position information (information on the "position of the characteristic part") of the storage means 20, thereby aligning the orientation of the image data. Then, the predetermined area specifying means 40b specifies the plurality of predetermined areas R2 on the image data based on the information (information on the "positions of the plurality of predetermined areas R2") of the storage means 20. Then, the predetermined area specifying means 40b outputs information on the specified plurality of predetermined areas to the determining means 50.

[0029] Although the neighboring area specifying means 40a and the predetermined area specifying means 40b have been described as separate means, they may exist as one means.

[0030] (judgment means) The determining means 50 is a means for determining whether or not there is an abnormality in the tab based on the degree of correlation between the input image data and the reference image data in the nearby region R0 specified by the nearby region specifying means 40a. Specifically, the determination means 50 calculates a correlation value between the input image data and the reference image data in the neighboring region R0. Next, the determination means 50 determines whether the calculated correlation value is equal to or greater than a predetermined value (above a threshold value read from the storage means 20), and determines that "the tab is normal" if the calculated correlation value is equal to or greater than the predetermined value, and determines that "the tab is abnormal" if the calculated correlation value is less than the predetermined value. The determination means 50 may calculate the correlation value of the image data using a general calculation method. Then, the determination means 50 outputs information on the determination result to the determination result output means 60.

[0031] The determining means 50 is a means for determining whether or not a protruding shape exists in all of the plurality of predetermined regions specified by the predetermined region specifying means 40b. 4 to 6, the determination means 50 detects the area value of a protruding shape in each predetermined region R2 of the image data (the area value is 0 if there is no protruding shape). Next, the determination means 50 determines whether or not each area value of a protruding shape in each predetermined region R2 of the image data is equal to or greater than a predetermined value (equal to or greater than the area value A2 read out from the storage means 20). Next, the determination means 50 outputs a result indicating whether or not a protruding shape (a protruding shape with an area value equal to or greater than a predetermined value) exists in all predetermined regions R2. Then, the determination means 50 outputs information on the determination result to the determination result output means 60.

[0032] The method for detecting the area value of the protruding shape in the determination means 50 is not particularly limited, but one example is a method in which, based on the color shade and brightness in the image data, a part whose shade value or brightness value is above (or below) a reference value is determined to be a protruding part, and the area value of that part is calculated.

[0033] (judgment result output means) The determination result output means 60 is a means for outputting the determination result input from the determination means 50 to the outside. This determination result output means 60 outputs the results to, for example, an alarm device that emits sound or light when an abnormality in the tab or a different type of can lid is detected, a device that removes can bodies equipped with can lids having abnormal tabs or different types of can lids from the production line, a monitor that can display all results, etc.

[0034] The image data input means 10, feature detection means 30, nearby area identification means 40a, predetermined area identification means 40b, judgment means 50, and judgment result output means 60 are realized by program execution processing by a CPU (Central Processing Unit), dedicated circuits, etc. The storage means 20 can be configured with a general storage device such as a RAM (Random Access Memory), a ROM (Read Only Memory), an HDD (Hard Disk Drive), or a flash memory.

[0035] [Can lid detection method according to this embodiment (operation of can lid detection device)] Next, a can lid detection method according to this embodiment will be described with reference to Fig. 7 (and Figs. 1 to 6 as appropriate). The operation of the can lid detection device according to this embodiment will also be described.

[0036] The can lid detection method of this embodiment includes an image data input process S1, a feature detection process S2, a feature presence / absence determination process S3, a nearby area identification process S5, a feature abnormality determination process S6, a specified area identification process S8, and a protruding shape presence / absence determination process S9. Each step of the can lid detection method according to this embodiment will be described below.

[0037] (Image data input process) The image data input step S1 is a step of inputting image data of the can lid into the can lid detection device 1. Specifically, in the image data input step S1, image data of a can lid captured by an imaging device 2 such as a camera is input to image data input means .

[0038] (Feature detection process) The characteristic part detection step S2 is a step of detecting the position of the tab (characteristic part) of the can lid in the image data of the can lid. Specifically, in the characteristic part detection step S2, the position of the tab C2 of the can lid C1 is detected in the image data of the can lid C1 as shown in Figures 2 to 6, which is input from the image data input means 10. There are no particular limitations on the method for detecting the position of the tab C2 in the characteristic part detection step S2, but examples include a method of identifying the position based on the color shading or brightness in the image data.

[0039] In order to ensure that the detection of the tab position in the characteristic part detection step S2, the calculation of the correlation value in the characteristic part abnormality determination step S6, and the calculation of the area value in the protrusion shape presence / absence determination step S9 can be performed efficiently, it is preferable to appropriately set the luminous intensity of the annular light source L shown in Figure 1 and the vertical distance between the light source L and the upper end of the can body C being transported on the production line so that the color shading and brightness on the image data can be clearly distinguished.

[0040] (Process for determining whether or not there are characteristic parts) The feature presence / absence determining step S3 is a step of determining whether or not the feature detected in the feature detection step S2 exists. The characteristic part presence / absence determining step S3 is not an essential step, and if there is zero possibility that the tab C2 will come off the can lid C1, this characteristic part presence / absence determining step S3 does not have to be provided.

[0041] In this characteristic part presence / absence determination step S3, if a characteristic part is detected in the characteristic part detection step S2, the step judges "Yes" and proceeds to the next step S5, and if for any reason a characteristic part is not detected in the characteristic part detection step S2, the step judges "No", sends a predetermined instruction to the outside (S4), and proceeds to step S11. The processing of the characteristic portion presence / absence determining step S3 may be performed by the characteristic portion detecting means 30 in FIG. 1 together with the processing of the characteristic portion detecting step S2, or may be performed separately by the determining means 50.

[0042] (Neighborhood area identification process) The neighboring area identification process S5 is a process of identifying a neighboring area in the image data based on the position of the tab (feature part) identified in the feature part detection process S2 and information on the neighboring area defined relative to the position of the tab area. 2 and 3, in the neighboring region identification step S5, the image data of the can lid C1 is rotated within the region R3 based on the information in the storage means 20 (information on the "position of the characteristic part"), and the orientation of the image data is aligned by aligning the position of the tab on the image data with the position of the tab C2 stored in the storage means 20. Then, based on the information in the storage means 20 (information on the "position of the neighboring region R0"), the neighboring region R0 on the image data is identified.

[0043] (Feature part abnormality determination process) The characteristic portion abnormality determination step S6 is a step of determining whether or not there is an abnormality in the tab based on the degree of correlation between the image data and the reference image data in the nearby region R0 identified in the nearby region identification step S5. More specifically, in the characteristic portion abnormality determination step S6, a correlation value between the image data and the reference image data stored in the storage means 20 is calculated in the vicinity region R0 shown in FIGS. 2, the degree of correlation with the reference image data (image data of the can lid with the tab not rotated at all) in the nearby region R0 is high, and the correlation value is determined to be equal to or greater than a predetermined value. As a result, the characteristic part abnormality determination step S6 in FIG. 7 is determined to be "Yes," and the process proceeds to the next step S8. On the other hand, when image data such as that shown in FIG. 3 is the target of judgment, the degree of correlation with the reference image data (image data of the can lid with the tab not rotated at all) in the vicinity region R0 is low, and as a result, the correlation value is judged to be less than the predetermined value. This is because, in the vicinity region R0 shown in FIG. 3, Braille on the can lid and the uneven shape of the can lid, which are not present in the reference image data, appear, resulting in a low degree of correlation. As a result, the characteristic part abnormality judgment step S6 in FIG. 7 is judged as "No," and a predetermined instruction is sent to the outside (S7), after which the process proceeds to step S11.

[0044] (Predetermined area identification process) The specified area identification process S8 is a process of identifying multiple specified areas in the image data based on the position of the tab (feature) detected in the feature detection process S2, and information on the position of the tab area on the can lid and the positions of multiple specified areas defined relative to the position of the tab area. 4 to 6, in the predetermined region specifying step S8, the image data of the can lid C1 is rotated within the region R3 based on the information from the storage means 20 (information on the "position of the characteristic part"), and the orientation of the image data is aligned by aligning the position of the tab on the image data with the position of the tab C2 stored in the storage means 20. Then, based on the information from the storage means 20 (information on the "positions of the plurality of predetermined regions R2"), the plurality of predetermined regions R2 on the image data are specified. Since the orientation of the image data is aligned in the neighboring region specifying step S5, the process of aligning the orientation of the image data in the predetermined region specifying step S8 may be omitted.

[0045] (Protrusion shape presence / absence determination process) The protruding shape presence / absence determining step S9 is a step of determining whether or not a protruding shape exists in all of the plurality of predetermined regions identified in the predetermined region identifying step S8. More specifically, in the protruding shape presence / absence determination step S9, as shown in FIGS. 4 to 6, the area value of a protruding shape is detected in each predetermined region R2 of the image data (the area value is 0 if there is no protruding shape). Next, it is determined whether the area value of each protruding shape in each predetermined region R2 of the image data is equal to or greater than a predetermined value (equal to or greater than the area value A2 read out from the storage means 20). Next, a result is obtained as to whether or not a protruding shape (a protruding shape with an area value equal to or greater than the predetermined value) exists in all predetermined regions R2. A specific determination method in this protruding shape presence / absence determining step S9 will be described below with reference to FIGS.

[0046] FIG. 4 shows image data in which Braille B is present on the surface of a can lid C1 and no water droplets are present. First, in the protruding shape presence / absence determination step S9, the area value A1 of the Braille, which is the protruding shape in the nine predetermined regions R2, is detected. Next, it is determined that each area value A1 of the Braille B in the nine predetermined regions R2 is equal to or greater than a predetermined value A2 (A1≥A2). That is, it is determined that it corresponds to "there is a protruding shape with an area value equal to or greater than the predetermined value in all nine predetermined regions R2", and it is determined that the lid of the can (the lid for an alcohol beverage can and a different type of can lid) on which the Braille B is imprinted. As a result, in the protruding shape presence / absence determination step S9 of FIG. 7, it is determined "Yes", and after transmitting a predetermined instruction to the outside (S10), the process proceeds to the next step S11.

[0047] FIG. 5 is image data when there is no Braille on the surface of the can lid C1 and there are no water droplets. First, in the protruding shape presence / absence determination step S9, the area value of the protruding shape is detected in the nine predetermined regions R2. Next, it is determined that each area value (0 mm 2 ) of the protruding shape in the nine predetermined regions R2 is not equal to or greater than the predetermined value A2 (0 < A2). That is, it is determined that it does not correspond to "there is a protruding shape with an area value equal to or greater than the predetermined value in all nine predetermined regions R2", and it is determined that the lid of the can (the lid for a beverage other than an alcohol beverage and a desired can lid) on which the Braille is not imprinted. As a result, in the protruding shape presence / absence determination step S9 of FIG. 7, it is determined "No", and without transmitting a predetermined instruction to the outside, the process proceeds to the next step S11.

[0048] FIG. 6 is image data when there is no Braille on the surface of the can lid C1 but there is a water droplet W. First, in the protruding shape presence / absence determination step S9, the area value of the protruding shape is detected in the nine predetermined regions R2. Next, each area value (0 mm 2) determines that it does not become equal to or greater than a predetermined value A2 (0 < A2). That is, it determines that it does not correspond to "there is a protruding shape with an area value equal to or greater than a predetermined value in all nine predetermined regions R2", and determines that it is a lid of a can (a lid for a beverage other than an alcoholic beverage and a desired lid of a can) on which Braille is not engraved. As a result, in the protruding shape presence / absence determination step S9 in FIG. 7, it is determined as "No", and without transmitting a predetermined instruction to the outside, the process proceeds to the next step S6.

[0049] The method for detecting the area value of the protruding shape in the protruding shape presence / absence determination step S9 is not particularly limited. For example, based on the light and shade or brightness and darkness of colors in the image data, it is determined that a portion where the light and shade value or brightness and darkness value is equal to or greater than (or equal to or less than) a reference value is a protruding portion, and the area value of the said portion is calculated.

[0050] (Steps after the protruding shape presence / absence determination step) When a predetermined instruction is transmitted to the outside (S4, S7, S10), for example, an alarm device that notifies by emitting sound or light that an abnormality of the tab or a different type of can lid (a can lid for an alcoholic beverage) has been detected operates, or a device that excludes a can body equipped with an abnormal tab or a different type of can lid from the production line operates. Note that, regardless of the determination result obtained, it may be configured to display all the results on a monitor.

[0051] In step S11, if there is the next image data, the process returns to step S1 which is the image data input step, and the image data is input. On the other hand, if there is no next image data, the flow ends.

[0052] [Effects of the can lid detection device and the can lid detection method according to the present embodiment] Next, the effects of the can lid detection device and the can lid detection method according to the present embodiment will be described with reference to the drawings.

[0053] According to the can lid detection device 1 and can lid detection method of this embodiment, the presence or absence of an abnormality in the tab C2 of the can lid C1 and the nearby area R0 near the tab C2 is determined based on the degree of correlation between the input image data of the can lid and the reference image data, thereby making it possible to properly detect a can lid C1r (Figure 3) in an abnormal state in which the tab C2 is rotating. Furthermore, according to the can lid detection device 1 and can lid detection method of this embodiment, it is possible to properly detect can lids C1r in an abnormal state in which the tab C2 is rotating in advance, thereby further reducing the possibility of erroneous detection when detecting different types of can lids C1 based on multiple specified areas R2 defined relative to the position of the tab C2. Specifically, in the case of a can lid C1r in an abnormal state where the tab C2 is rotating, the multiple specified areas R2 defined relative to the position of the tab C2 will differ from the desired areas, resulting in a false detection.However, by detecting in advance the can lid C1r in an abnormal state that could lead to such a situation, the possibility of false detection of a different type of can lid can be further reduced.

[0054] According to the can lid detection device 1 and can lid detection method of this embodiment, the can lid C1 is identified based on the plurality of predetermined regions R2, and therefore can be appropriately identified between a can lid C1a (FIG. 4) in which a protruding shape exists in all of the plurality of predetermined regions R2 and can lids C1b and C1c (FIGS. 5 and 6) in which a protruding shape does not exist in some or all of the plurality of predetermined regions R2. Therefore, the can lid detection device 1 and can lid detection method of this embodiment can appropriately identify a can lid C1a for alcoholic beverages (can lid C1a stamped with Braille) and can lids C1b and C1c for non-alcoholic beverages (can lids C1b and C1c not stamped with Braille), even in a situation where water droplets W adhere to the surface of the can lid C1. Therefore, according to the can lid detection device 1 and can lid detection method of this embodiment, the possibility of erroneous detection can be reduced when detecting can bodies C equipped with different types of can lids C1a mixed in a group of can bodies C.

[0055] [Modifications of the can lid detection device and can lid detection method according to the present embodiment] In the can lid detection device and can lid detection method of this embodiment, a configuration has been described in which a neighborhood identification means 40a that identifies a neighborhood area in image data or a neighborhood identification process S5 is included, but this neighborhood identification means 40a or neighborhood identification means S5 may also be omitted. Specifically, position data of a nearby area defined relative to the position of the tab is specified on the reference image data stored in the storage means 20. Then, when the determination means 50 or the characteristic part abnormality determination step S6 calculates the correlation value between the image data and the reference image data, the correlation value in the nearby area may be calculated based on the position data of the nearby area stored in the storage means 20.

[0056] 4 shows the braille on the can lid C1 indicating "alcohol," but it is not limited to this braille and may be various braille indicating the type of food or drink contained in the can. In this case, the plurality of predetermined areas R2 naturally corresponds to the positions and number of the braille.

[0057] In Figures 2 to 6, the tip of tab C2 is positioned so that the orientation of the image data is aligned to face left, but this is not particularly limited, and for example, the tip of tab C2 may be positioned so that the orientation of the image data is aligned to face right. The shape of the vicinity region R0 shown in FIGS. 2 and 3 and the shapes of the plurality of predetermined regions R2 shown in FIGS. 4 to 6 are not limited to a rectangle, but may be a square, a circle, an ellipse, or the like.

[0058] In the nearby area identification means 40a and predetermined area identification means 40b shown in Figure 1, and the nearby area identification step S5 and predetermined area identification step S8 shown in Figure 7, a configuration has been described in which the orientation of the image data is aligned by rotating the image data of the can lid C1 within the area R3 and aligning the position of the tab C2 on the image data with the position information of the storage means 20 (information on the "position of the characteristic part", i.e., information on the "position of the tab"), but this method is not limited to this. For example, as shown in Figure 6, a memory means 20 may store an area R1 that specifies where the tab C2 is located relative to the can lid C1 (area R3 surrounding the can lid C1), and the image data of the can lid C1 may be rotated within area R3 to align the tab C on the image data with area R1, thereby aligning the orientation of the image data. When region R1 is used, regions R0 and R2 may be defined relative to region R1. As shown in FIG. 6, the region R1 is a rectangular region having one side equal to the length of the tab C2 in the longitudinal direction and the other side equal to the length of the tab C2 in the lateral direction.

[0059] Although the explanation was given assuming that the group of can bodies C continuously transported on the production line shown in Figure 1 contains non-alcoholic beverages, the method can also be applied to cases where the group of can bodies C contains alcoholic beverages. In other words, the method can also be applied to cases where the "can lids" of the majority of can bodies C in the group of can bodies C are can lids C1a stamped with Braille (Figures 2 and 4), and the "different can lids" to be detected are can lids C1b and C1c (Figures 5 and 6) that do not have Braille stamped on them.

[0060] In the predetermined region specifying means 40b shown in FIG. 1 and the protruding shape presence / absence determining step S9 shown in FIG. 5, a configuration has been described in which it is determined that Braille is present when the area value of a protruding shape in a plurality of predetermined regions R2 is equal to or greater than a predetermined value (greater than or equal to the area value A2 read from the storage means 20: A2≦area value A1 of each Braille character). However, the area value A2 may not be set. For example, if water droplets with an area value smaller than that of Braille are rarely generated, it may be determined that Braille is present when even a slight protruding shape is detected in each predetermined region R2. In this case, the storage means 20 does not need to store the information (area value A2) read by the determining means 50.

[0061] It should be noted that the present embodiment has been described in detail to clearly explain the present invention, and is not necessarily limited to having all of the described configurations. Furthermore, some of the configurations of the present embodiment can be added to, deleted from, or replaced with other configurations. Furthermore, the mechanisms and configurations described above are those that are considered necessary for the explanation, and do not necessarily represent all mechanisms and configurations of the product. [Example]

[0062] Next, the can lid detection device and can lid detection method according to the present invention will be described by way of examples that satisfy the requirements of the present invention and comparative examples that do not.

[0063] [Example 1: Preliminary test] A preliminary test was conducted to confirm the relationship between the rotation angle of the tab on the can lid and the ease of opening the lid (lid opening property).

[0064] (Sample preparation) As a sample, a commercially available 350 mL can (with Braille indicating alcohol on the can lid) was prepared, and the tab on the can lid was rotated to the rotation angle shown in Table 1.

[0065] (Openability evaluation test) One man and three women lifted the tab of each sample with their fingers to evaluate the ease of opening. For the evaluation of ease of opening, a control sample with a tab rotation angle of 0° (a sample of the can body in a normal state with the tab not rotated) was used as the standard, and a score of 3 was given for cases where the opening was as easy as the control sample, a score of 2 for cases where the opening was slightly more difficult than the control sample, and a score of 1 for cases where the opening was more difficult than the control sample.

[0066] Table 1 shows the results of the lid opening evaluation test. The "tab rotation angle" in Table 1 refers to the clockwise rotation angle of the tab tip on the can lid, with the normal state where the tab is not rotated being set at 0°.

[0067] [Table 1]

[0068] (Review of results) According to the results in Table 1, Samples 1-1, 1-2, 1-5, and 1-6 had scores (average value) of 2.5 or more in the evaluation of the ease of opening, and thus obtained favorable results. In other words, it was confirmed that sufficient opening properties can be ensured if the rotation angle of the tab is in the range of 0 to 10°, 350 to 360° (if the clockwise rotation angle of the tab tip is defined as a positive angle and the counterclockwise rotation angle of the tab tip is defined as a negative angle, then the range is -10 to +10°).

[0069] [Example 2: Main Test] (Sample preparation) As a sample, a commercially available 350 mL can (with Braille indicating alcohol on the can lid) was prepared, and the tab on the can lid was rotated to the rotation angles shown in Tables 2 and 3. The samples shown in Table 2 were prepared without any water droplets attached, while the samples shown in Table 3 were prepared by spraying water onto the surface of the can lid to create a state where water droplets were attached, in order to simulate the condition after cleaning work on a production line.

[0070] (Experimental content of this test) Using an apparatus with a configuration similar to the can lid detection device of this embodiment (Figure 1), the correlation values ​​(correlation values ​​in the nearby region R0 shown in Figure 2) between the image data of the can lids of the samples shown in Tables 2 and 3 and the reference image data were calculated three times for each sample. The area of ​​the nearby region R0 is 560 mm 2 (approx. 28mm x approx. 20mm) and the tab area (approx. 360mm 2 ) was approximately 1.56 times larger than the reference image data. The reference image data was image data of a can lid in a normal state where the tab was not rotated at all (tab rotation angle 0°). In addition, in the can lid detection device of this embodiment, the vertical distance between the upper end of the can body C shown in Figure 1 and the light source L (model: CA-DC50E, lighting brightness (volume value): 511) was approximately 15 mm.

[0071] The following camera was used as the imaging device 2 shown in FIG. Inspection machine camera model: CA-H048CX, 470,000 pixel 16x speed color camera Effective pixels: 784 (H) x 596 (V) (progressive) Shutter speed: 1 / 20000

[0072] Tables 2 and 3 show the correlation values ​​obtained in this study. As in Table 1, the "tab rotation angle" in Tables 2 and 3 refers to the clockwise rotation angle of the tab tip on the can lid, with the normal state in which the tab is not rotated being taken as 0°.

[0073] [Table 2]

[0074] [Table 3]

[0075] (Review of results) According to the results in Table 2, when the present invention is applied to a production line where water droplets are not expected to adhere to can lids, it has been confirmed that even slight rotation of the tab (rotation exceeding the range of -5 to +5 degrees, where the clockwise rotation angle of the tab tip is defined as a positive angle and the counterclockwise rotation angle of the tab tip is defined as a negative angle) can be detected by setting the correlation value threshold to 90 or more, 93 or more, 95 or more, etc.

[0076] According to the results in Table 3, when the present invention is applied to a production line where water droplets are expected to adhere to can lids, it has been confirmed that even slight rotation of the tab (rotation exceeding the range of -5 to +5 degrees, where the clockwise rotation angle of the tab tip is defined as a positive angle and the counterclockwise rotation angle of the tab tip is defined as a negative angle) can be detected by setting the correlation value threshold to 93.5 or more. [Explanation of symbols]

[0077] 1 Can lid detection device 2. Imaging device, camera 10 Image data input means 20 Memory means 30 Feature detection means 40a Neighborhood area specifying means 40b Specified area identification means 50 Judgment means 60 Judgment result output means L light source C Can body C1 Can lid C2 Features, tab B Braille W water drop R0 neighborhood R1 area R2 Multiple predetermined areas S1 Image data input process S2 Feature detection process S3: Process for determining whether or not there are characteristic parts S5 Neighborhood area identification process S6 Feature abnormality determination process S8 Specified area identification process S9 Protrusion shape presence / absence determination process

Claims

1. A can lid detection device that detects can bodies having can lids with an abnormality in which the tab is rotating from a group of can bodies, a storage means for storing in advance reference image data of a normal can lid with an unrotated tab; a determination means for determining whether or not there is an abnormality, such as the tab rotating, in the tab of the can lid and a region adjacent to the tab, based on the degree of correlation between the input image data of the can lid and the reference image data; Equipped with A can lid detection device in which the vicinity region is a region that includes the tab and extends from the outer edge of the tab to the periphery, and the area of ​​the vicinity region is 1.1 times or more the area of ​​the tab.

2. the storage means stores information on the position of the tab on the can lid and the positions of a plurality of predetermined areas defined relative to the position of the tab; The method further includes a predetermined area specifying means for specifying a plurality of predetermined areas in the image data based on the position of the tab of the can lid in the input image data of the can lid and the information stored in the storage means, 2. The can lid detection device according to claim 1, wherein the determining means determines whether or not a protruding shape exists in each of the plurality of predetermined regions specified by the predetermined region specifying means.

3. A can lid detection method for detecting can bodies having can lids with an abnormality in which the tab is rotating from a group of can bodies, comprising: a tab determination step of determining whether or not the tab is rotated based on the degree of correlation between image data of the can lid captured in the tab of the can lid and a nearby area near the tab and reference image data of a normal can lid whose tab is not rotated; Including, A can lid detection method, wherein the vicinity region is a region that includes the tab and extends from the outer edge of the tab to the periphery, and the area of ​​the vicinity region is 1.1 times or more the area of ​​the tab.

4. a predetermined area specifying step of specifying a plurality of predetermined areas in the image data based on the position of the tab in the image data of the can lid obtained by capturing an image, and information on the position of the tab on the can lid and the positions of a plurality of predetermined areas defined relative to the position of the tab; a protruding shape determination step of determining whether or not a protruding shape exists in each of the plurality of predetermined regions identified in the predetermined region identification step; The can lid detection method according to claim 3, further comprising:

Images