JP6359364B2 - Container inspection method and container inspection apparatus - Google Patents

Description

  The present invention relates to a container inspection method and a container inspection apparatus for inspecting the appearance of a container.

  2. Description of the Related Art Conventionally, a container inspection device that inspects the appearance of a label attached by, for example, wrapping around a container using an image obtained by imaging a container such as a bottle or a bottle with a camera is known. . That is, the container inspection device takes an image of the container with a label on which an image including characters and figures is printed on the outer surface of the container, and uses the image obtained by imaging to tear the label. Inspect appearance such as drowning. As a container inspection apparatus of this type, the reflected light of a container is imaged with a CCD color camera, and the color of the contents (liquid) in the container registered in advance at a position corresponding to the position where the label exists in the captured image. Or the container test | inspection apparatus which determines with a label having a defect when the color of a container is detected is proposed conventionally (for example, refer patent document 1).

JP 2005-119706 A

  However, in the conventional container inspection apparatus, when the label is attached before the contents are filled in the container, it is difficult to detect the color of the contents. Further, when the container is colorless and transparent, it is difficult to detect a label defect by detecting the color of the container. For example, when the container has a substantially cylindrical shape, the color of the contents in the container or the color of the container may change depending on the position of the cylindrical shape and the lighting condition. In such a case, it is difficult to detect the color of the contents (liquid) in the container or the color of the container.

  Such a situation is generally common in a container inspection method and a container inspection apparatus in which a container is imaged by a camera and the appearance of the label is inspected using an image obtained by imaging.

  The present invention has been made in view of such circumstances, and its purpose is to be applied to the outer surface regardless of the color of the contents or the color of the container. An object of the present invention is to provide a container inspection method and a container inspection apparatus capable of performing an appearance inspection of a container on which a label is attached.

Hereinafter, means for solving the above-described problems and the effects thereof will be described.
A container inspection apparatus that solves the above-described problem is a container inspection apparatus that performs an appearance inspection of a container having an inspection target label mounted on an outer surface thereof, and includes at least a part of the inspection target label by imaging the container An imaging unit that generates a partial image of the image, an image specifying unit that specifies a reflected image of the label to be inspected that is reflected in the partial image, and image conversion that converts the reflected image into a plurality of specific color inspection images A container inspection unit that performs pass / fail determination of the inspection target label using an image region of at least one specific color among the plurality of specific colors in the inspection image.

  According to this configuration, for example, even if the color of the contents or the color of the container including the inspection target label varies in the captured partial image, the image area of one specific color converted in the inspection target label is used. As a result, it is possible to determine the quality of the inspection target label. That is, regardless of the color of the contents or the color of the container, and even if the color of the inspection target label varies, it is possible to perform an appearance inspection of the container with the inspection target label attached to the outer surface.

  The container inspection apparatus includes a storage unit that stores a non-defective label image representing a non-defective label attached to the outer surface of the container, and the image specifying unit compares the non-defective label image with the partial image. And specifying a reflected image of the inspection target label reflected in the partial image, and identifying a corresponding image corresponding to the reflected image in the non-defective label image, and the image conversion unit converts the corresponding image to the The container inspection unit uses the image area of the one specific color in the inspection image and the image area of the non-defective image having the same specific color as the one specific color. Thus, it is preferable to determine whether the inspection target label is good or bad.

  According to this configuration, the image area of the specific color of the used inspection image and the image area of the non-defective image of the same specific color as the image area of the inspection image depend on whether the image areas match each other. The quality of the inspection target label can be determined. Therefore, regardless of the color of the contents or the color of the container, and even if there is a variation in the color of the inspection target label, the appearance inspection with high accuracy of the quality determination is performed on the container having the inspection target label mounted on the outer surface. It is possible.

  In the container inspection apparatus, it is preferable that the container inspection unit performs a pass / fail determination of the inspection target label using an outline area indicating an outline of the image area in the image area of the one specific color.

  According to this configuration, since the pass / fail determination is performed using the contour area of the image area of the specific color, the area of the image (number of pixels) used for the determination process is higher than when the pass / fail determination is performed using the entire area of the image area of the specific color. ) And the processing load related to determination can be reduced.

  A container inspection apparatus that solves the above-described problem is a container inspection apparatus that performs an appearance inspection of a container having an inspection object display on an outer surface thereof, and images the container and includes at least a part of the inspection object display. An imaging unit that generates an image, an image specifying unit that specifies a reflected image of the inspection target display reflected in the partial image, and an image converting unit that converts the reflected image into a plurality of specific color inspection images; And a container inspection unit that performs pass / fail determination of the inspection target display using an image region of at least one specific color among the plurality of specific colors in the inspection image.

  According to this configuration, for example, even if the color of the contents or the color of the container including the inspection target display varies in the captured partial image, the image area of one specific color converted in the inspection target display is used. As a result, it is possible to determine whether the inspection target display is good or bad. That is, regardless of the color of the contents or the color of the container, even if the color of the inspection object display varies, it is possible to perform an appearance inspection of the container having the inspection object display on the outer surface.

The container inspection apparatus includes a storage unit that stores a non-defective display image including the inspection target display of the non-defective product of the container,
The image specifying unit specifies the reflected image of the inspection object display reflected in the partial image by comparing the good quality display image with the partial image, and the reflected image in the good quality display image. Identify the corresponding image corresponding to,
The image conversion unit converts the corresponding image into the non-defective images of the plurality of specific colors,
The container inspection unit performs pass / fail determination of the inspection target display using the image region of the one specific color in the inspection image and the image region of the non-defective image of the same specific color as the one specific color. It is preferable.

  According to this configuration, the image area of the specific color of the used inspection image and the image area of the non-defective image of the same specific color as the image area of the inspection image depend on whether the image areas match each other. The quality of the inspection object display can be determined. Therefore, regardless of the color of the contents or the color of the container, and even if there is a variation in the color of the inspection object display, it is possible to perform an appearance inspection with high accuracy of the quality determination for the container having the inspection object display on the outer surface. Is possible.

  In the container inspection apparatus, it is preferable that the container inspection unit determines whether or not the inspection object display is acceptable by using an outline region indicating an outline of the image region in the image region of the one specific color.

  According to this configuration, since the pass / fail determination is performed using the contour area of the image area of the specific color, the area of the image (number of pixels) used for the determination process is higher than when the pass / fail determination is performed using the entire area of the image area of the specific color. ) And the processing load related to determination can be reduced.

In the container inspection apparatus, it is preferable that the plurality of specific colors are colors belonging to each of divided color spaces obtained by dividing a color space indicated by hue, brightness, and saturation.
According to this configuration, the image of the inspection target label or the inspection target display belongs to each of the divided color spaces obtained by dividing the color space indicated by the hue, lightness, and saturation, which makes it easy for a person to recognize the color difference. Since it is converted into a color image, it is possible to appropriately perform the appearance inspection of the inspection object label or the inspection object display as in the case where the person performs the inspection.

  A container inspection method for solving the above-described problem is a container inspection method for inspecting the appearance of a container having an inspection target label mounted on an outer surface thereof, wherein the container includes at least a part of the inspection target label by imaging the container. An imaging step for generating a partial image, an image specifying step for specifying a reflected image of the label to be inspected that is reflected in the partial image, and an image conversion for converting the reflected image into a plurality of specific color inspection images And a container inspection step of determining pass / fail of the label to be inspected using an image region of at least one specific color among the plurality of specific colors in the inspection image.

  According to this method, for example, even if the color of the contents or the color of the container including the inspection target label varies in the captured partial image, the image area of one specific color converted in the inspection target label is used. As a result, it is possible to determine the quality of the inspection target label. That is, regardless of the color of the contents or the color of the container, and even if the color of the inspection target label varies, it is possible to perform an appearance inspection of the container with the inspection target label attached to the outer surface.

  The container inspection method includes a storing step of storing a non-defective label image representing a non-defective label attached to the outer surface of the container, and in the image specifying step, the non-defective label image is compared with the partial image. Identifying the reflected image of the label to be inspected in the partial image and identifying the corresponding image corresponding to the reflected image in the non-defective label image, and in the image converting step, The image is converted into a plurality of non-defective images of specific colors, and the container inspection step uses the image area of the one specific color in the inspection image and the image area of the non-defective image of the same specific color as the one specific color. Thus, it is preferable to determine whether the inspection target label is good or bad.

  According to this method, the quality of the inspection object label is determined by, for example, a difference between the image area of the specific color of the used inspection image and the image area of the non-defective image of the same specific color as the image area of the inspection image. Can be determined. Therefore, regardless of the color of the contents or the color of the container, and even if there is a variation in the color of the label to be inspected, it is possible to perform a visual inspection with a high accuracy of the pass / fail judgment for the container with the label attached to the outer surface. Is possible.

  A container inspection method for solving the above-described problem is a container inspection method for inspecting the appearance of a container having an inspection target display on the outer surface, and images the container and includes at least a part of the inspection target display. An imaging step for generating an image; an image specifying step for specifying a reflected image of the inspection object display reflected in the partial image; and an image converting step for converting the reflected image into a plurality of specific color inspection images; And a container inspection step of performing pass / fail determination of the inspection target display using an image region of at least one specific color among the plurality of specific colors in the inspection image.

  According to this method, for example, even if the color of the contents or the color of the container including the inspection target display varies in the captured partial image, the image area of one specific color converted in the inspection target display is used. As a result, it is possible to determine whether the inspection target display is good or bad. That is, regardless of the color of the contents or the color of the container, even if the color of the inspection object display varies, it is possible to perform an appearance inspection of the container having the inspection object display on the outer surface.

  The container inspection method includes a storage step of storing a non-defective display image including the inspection target display of the non-defective product of the container, and in the image specifying step, by comparing the non-defective display image and the partial image, The reflected image of the inspection object display reflected in the partial image is specified, and the corresponding image corresponding to the reflected image in the non-defective product display image is specified. In the container inspection step, the image area of the one specific color in the inspection image and the image area of the non-defective image of the same specific color as the one specific color are used in the container inspection step. It is preferable to determine whether the inspection target display is good or bad.

  According to this method, the quality of the inspection object display is determined by, for example, a difference between the image area of the specific color of the used inspection image and the image area of the non-defective image of the same specific color as the image area of the inspection image. Can be determined. Therefore, regardless of the color of the contents or the color of the container, and even if there is a variation in the color of the inspection object display, it is possible to perform an appearance inspection with high accuracy of the quality determination for the container having the inspection object display on the outer surface. Is possible.

The schematic block diagram of the container inspection apparatus of embodiment. The flowchart which shows the inspection process of the external appearance inspection of the container performed with a container inspection apparatus. (A) is explanatory drawing which shows the partial image of the container produced | generated, (b) is a schematic diagram which shows an example of the produced | generated partial image, (c) is a non-defective product corresponding to the reflected image of the inspection object label in the partial image. The schematic diagram which shows the corresponding | compatible image of a label image. The schematic diagram which shows the division | segmentation color space divided | segmented into the several specific color. (A) is a schematic diagram showing a non-defective image converted from a corresponding image to an image region of a specific color, and (b) is a schematic diagram showing an inspection image converted from a partial image to an image region of a specific color. The schematic diagram which shows the external appearance test | inspection using the image area | region of a non-defective image, and the image area | region of an inspection image. (A) is a schematic diagram showing a non-defective image converted from a corresponding image into a specific color contour region, and (b) is a schematic diagram showing an inspection image converted from a partial image into a specific color contour region.

  Hereinafter, an embodiment of a container inspection apparatus will be described with reference to the drawings. The container inspection apparatus according to this embodiment is an apparatus that performs a predetermined appearance inspection on a container that is mounted with a label printed with a display object such as a character or an image wound around the outer surface, using the label as an inspection target label. .

  As shown in FIG. 1, the container inspection apparatus 11 includes, for example, an imaging device 12 such as a digital camera, and a control device 13 having a computer function including a central processing unit (CPU) and a storage element (RAM, ROM). I have. And the control apparatus 13 images the container 20 conveyed by the belt conveyor 14 which moves to one direction as shown by the solid line arrow in FIG. 3A) is generated, and the appearance inspection of the container 20 is performed using the generated partial image BG. In the present embodiment, the container 20 is a resin bottle that can accommodate a liquid therein, and has a cap 21 at the upper end and a round bottle having an outer surface that is cylindrical, and the outer surface is inspected. The target label 30 is attached by winding or the like.

  The container inspection apparatus 11 according to the present embodiment is configured such that the control device 13 operates based on a predetermined computer program supplied by a storage medium MB such as a memory card or an optical disk, thereby causing the storage unit 15, the imaging unit 16, It functions as an image specifying unit 17, an image conversion unit 18, and a container inspection unit 19.

  The storage unit 15 performs a process of storing a non-defective label image YG (see FIG. 3C) representing a non-defective label attached to the outer surface of the container 20. The imaging unit 16 operates the imaging device 12 to image a partial section of the entire circumferential section of the outer surface of the container 20, whereby the partial image BG of the container 20 in which at least a part of the inspection target label 30 is reflected. Process to generate. The image specifying unit 17 specifies the reflected image SG (see FIG. 3B) of the inspection target label 30 included in the partial image BG among the inspection target labels 30, and also the inspection target label 30 in the non-defective label image YG. The corresponding image TG (see FIG. 3C) corresponding to the reflected image SG is identified. The image conversion unit 18 converts the reflected image SG of the inspection target label 30 into a plurality of specific color inspection images IG (see FIG. 5B), and converts the corresponding image TG into a plurality of specific color non-defective images GG ( The process of converting to FIG. 5 (a) is performed. The container inspection unit 19 performs an appearance inspection process using an image region of one specific color in the inspection image IG and an image region of a non-defective image GG having the same specific color as the one specific color.

  In the present embodiment, the control device 13 includes a display unit 13A, an image captured by the image capturing device 12, a partial image BG generated by the image capturing unit 16, or a captured image SG identified by the image identifying unit 17. Alternatively, the result of the appearance inspection of the container 20 performed by the container inspection unit 19 is displayed on the display unit 13A as necessary.

  Furthermore, the container inspection apparatus 11 is provided with an illumination device LT for illuminating the container 20 as necessary. The lighting device LT is more anti-gravity than the imaging device 12 so that the light reflected by the container 20 does not directly enter the imaging device 12 disposed opposite to the container 20 in the horizontal direction intersecting the outer surface of the container 20. It arrange | positions so that the container 20 may be illuminated from diagonally upward which becomes a direction side.

  Next, the operation of the container inspection apparatus 11, that is, the inspection process of the container 20 performed by the container inspection apparatus 11 will be specifically described based on the flowchart shown in FIG. This inspection process is started with the start of conveyance of the container 20 by the belt conveyor 14, for example.

  As shown in FIG. 2, when this inspection process is started, a process for storing a non-defective label image YG is first performed in step S1. Here, the storage unit 15 stores a non-defective label image YG representing the non-defective inspection target label 30 supplied from the storage medium MB. In the present embodiment, the display object 30 is printed on the entire label surface, and the print data of the display object is supplied as image data of the non-defective label image YG.

  In the present embodiment, the stored non-defective label image YG is composed of a plurality of pixels having image data represented by R (red), G (green), and B (blue) gradation values. Further, the non-defective label image YG is an image showing at least the entire inspection target label 30 along the winding direction when the inspection target label 30 is wound along the cylindrical outer surface (cylindrical surface) of the container 20 (FIG. 3). (See (c)).

  Next, in step S2, a process of capturing the container 20 and generating a partial image BG is performed. The container inspection apparatus 11 is provided with a detection sensor (not shown) that detects the position of the container 20 conveyed by the belt conveyor 14, and when the detection sensor detects the container 20 that has moved to a predetermined imaging position, the imaging unit 16 Operates the imaging device 12. By the operation of the imaging device 12, an image of a partial section of one round (entire circumference) along the outer surface of the container 20 is taken.

  Then, the process which produces | generates partial image BG from the image of the one part area of the imaged container 20 is performed. In the present embodiment, this processing is performed by the imaging unit 16, and in the captured image of the partial section, a rectangular image that is reflected at a predetermined size at a predetermined position with respect to a predetermined reference position of the container 20. A partial image BG is generated by cutting out KG and correcting the cut-out rectangular image KG.

  That is, in an image of a partial section of the container 20 acquired by the imaging device 12, a feature point (for example, the center of the cap 21) is set as a reference position, and is a predetermined position and a predetermined size with respect to the reference position. Cut out a rectangular image. Then, the cut-out rectangular image is developed into a flat image (also referred to as “panorama development”) and corrected to generate a partial image BG. In the present embodiment, the generated partial image BG is an image configured by a plurality of pixels having image data represented by R (red), G (green), and B (blue) gradation values. .

  Next, in step S3, a process for specifying the reflected image SG of the inspection target label 30 reflected in the partial image BG is performed, and in the next step S4, the reflected image SG in the non-defective label image YG is displayed. Processing for specifying the corresponding corresponding image TG is performed. The specifying process in step S3 and step S4 is performed by the image specifying unit 17 comparing the non-defective label image YG with the partial image BG.

  In the present embodiment, as an example, comparison is performed using a known image matching process such as template matching or pattern matching, and the position of the partial image BG in the non-defective label image YG is specified. Note that, in the image matching processing, the image specifying unit 17 determines the image size (upper and lower lengths) of the inspection target label 30 in the partial image BG and the size of the label portion image in the non-defective label image YG. At least one of the partial image BG and the non-defective label image YG is corrected by enlarging or reducing in advance so that the height (upper and lower lengths) matches.

Here, the processing from step S1 to step S4 will be described with reference to FIG.
First, as shown on the left side in FIG. 3A, in the image of the container 20 captured by the imaging device 12 having a predetermined angle of view, the angle in the circumferential direction (left-right direction) intersecting the cylindrical axis of the outer surface of the container 20 At both ends of the Ga cylindrical surface, in principle, the image pickup device 12 is difficult to focus on and tends to be blurred images. Therefore, with respect to the left-right direction of the image of the container 20, an image corresponding to an angle Gc that is inside the both ends of the cylindrical surface and narrower than the angle Ga, that is, one of one round (entire circumference) along the outer surface of the container 20. The image of the partial section is used for the appearance inspection of the container 20. Incidentally, in the present embodiment, the angle Gc is set to 100 °, which is 50 ° left and right with the cap 21 as the center. Of course, the angle Gc may be 120 °, for example, and is preferably as large as possible within a range where an image suitable for the appearance inspection of the container 20 is obtained.

  Further, as shown in the center of FIG. 3A, in the vertical direction of the image of the container 20, the dimension La is downward from the cap 21 serving as the reference position so that all the images of the inspection target label 30 are included. An image portion corresponding to a dimension Lb that is longer than the upper and lower lengths of the inspection target label 30 positioned below the position separated from the length is used for appearance inspection. That is, as shown by a portion surrounded by a broken line in the center of FIG. 3A, a rectangular image KG of a predetermined size is cut out as an image used for appearance inspection from the image of the container 20 imaged by the imaging device 12. It is.

  Next, since the cut-out rectangular image KG is a cylindrical image in the circumferential direction (left-right direction) of the outer surface of the container 20, as illustrated on the right side in FIG. . That is, the rectangular image KG is an image having a size of Lc / 2 on the left and right sides and a size of Lb on the upper and lower sides around the cap 21 by the correction by the panoramic development, and each gradation value of R, G, B. Is generated as a partial image BG composed of a plurality of pixels having the image data indicated by. In the container inspection device 11, the generated partial image BG is used for appearance inspection of the container 20. Note that panoramic development of the rectangular image KG is possible by using a well-known method, and therefore description of panoramic development is omitted here.

  FIG. 3B shows an example of the partial image BG generated from the image of the container 20 imaged by the imaging device 12. In FIG. 3B, a rectangular image KG surrounded by a broken line in the captured image of the container 20 shown on the left side of the drawing is cut out, and the cut-out rectangular image KG is panorama developed, and an upper and lower dimension Lb shown on the right side of the drawing. And a flat partial image BG having left and right dimensions Lc. The partial image BG includes a reflected image SG that is a reflected portion in which at least a part of the inspection target label 30 is reflected.

  Therefore, as shown in FIG. 3C, the image specifying unit 17 compares the stored non-defective label image YG with the generated partial image BG. Image matching processing is performed by comparing the image data of the partial image BG for each pixel. As a result, the reflected image SG of the inspection target label 30 reflected in the partial image BG is specified and corresponds to the reflected image SG in the non-defective label image YG as indicated by a thick broken line in FIG. The corresponding image TG is specified.

  Incidentally, in the present embodiment, the identified corresponding image TG has an upper and lower dimension Ld and a left and right dimension Lc that are the same size as the reflected image SG of the inspection target label 30, and is the same as the reflected image SG. The image has the number of pixels. The dimension Le along the outer surface (cylindrical surface) of the non-defective label image YG is a length corresponding to 360 ° around the outer surface of the container 20, and the dimension Lc along the outer surface (cylindrical surface) of the corresponding image TG is 3 The length is 1/6 (100 °). Further, the upper and lower dimensions of the non-defective label image YG are set to the same dimension Lb as that of the partial image BG. If the dimension Le along the outer surface (cylindrical surface) of the stored good product label image YG has a length of 360 ° around the outer surface of the container 20, the position of the partial image BG in the non-defective product label image YG is determined by matching processing. However, it may be a length obtained by adding the same length as the dimension Lc along the outer surface (cylindrical surface) of the corresponding image TG. By doing so, in the matching process for specifying the position of the partial image BG in the non-defective label image YG, the partial image BG can be used without being divided, and the matching process is facilitated.

  Returning to FIG. 2, in the following step S5, the reflected image SG is converted into a plurality of specific color inspection images IG. In the next step S6, the corresponding image TG is converted into a plurality of non-defective images of specific colors. A process of converting to GG is performed. Here, the image conversion unit 18 converts the image data of the reflected image SG and the image data of the corresponding image TG into image data of a plurality of different specific colors specified in a predetermined color space.

A plurality of different specific colors to be converted will be described with reference to FIG.
As shown in FIG. 4, in the present embodiment, a plurality of specific colors are HSV indicated by hue (H), saturation (S), and lightness (V) each having a gradation value from 0 to 255. The color space belongs to any one of divided color spaces obtained by dividing the color space into eight spaces.

  That is, regardless of the hue and lightness gradation values, the color space having a saturation (S) gradation value of 0 to 60 is white, and the lightness is independent of the hue and saturation gradation values. A color space having a gradation value (V) of 0 to 60 is black. Further, the other space is divided by the hue gradation value, the color space having the hue (H) gradation value from 20 to 62 is yellow, and the color space having the value from 62 to 107 is green. , 107 to 149, the color space is light blue. Further, the color space having a hue (H) gradation value from 149 to 192 is blue, the color space having a value from 192 to 234 is pink, and from 234 to 25 after 255. The color space having the value of is red.

  In step S5 and step S6, the image conversion unit 18 first converts the reflected image SG and the corresponding image TG from image data indicated by R, G, and B gradation values using a known conversion formula. The image data is converted into image data indicated by H, S, and V gradation values. Next, based on the converted gradation values of H, S, and V, the image conversion unit 18 converts the pixels of the reflected image SG and the pixels of the corresponding image TG to one specific color among a plurality of specific colors. The image data is converted into image data having one gradation value.

  For example, in the HSV color space shown in FIG. 4, pixels having tone values of hue (H) of 120, saturation (S) of 100, and lightness (V) of 100 are all light blue pixels, and the image The data is converted into one gradation value having a hue (H) of 128, a saturation (S), and a lightness (V) of 255 each. Such conversion of the image data is performed on all the pixels of the reflected image SG and the corresponding image TG, and converted into the inspection image IG and the non-defective image GG, respectively.

  FIGS. 5A and 5B show an example of a non-defective image GG and an inspection image IG converted into a plurality of specific color image data. In the present embodiment, it is assumed that white and blue triangular figures, black straight lines, and yellow letters are printed as display objects on a green background.

  As shown in FIG. 5A, the non-defective image GG converted from the corresponding image TG includes a label image (non-defective label) on which a display object is printed in each color, a green image region GR, and a white image region. The image is converted into an image composed of WH, a blue image region BL, a yellow image region YL, and a black image region KR.

  On the other hand, as shown in FIG. 5B, in this embodiment, the inspection target label 30 is broken by the first peeling portion H <b> 1 and the second peeling portion H <b> 2 due to tearing, and the upper end portion of the inspection target label 30 is broken. It is assumed that the third peeling portion H3 exists. For this reason, in the inspection image IG converted from the reflected image SG, the color of the container 20 is converted into a specific color in the region corresponding to the peeled portion. Incidentally, in this embodiment, it shall convert into light blue. As a result, the reflected image SG has the first peeling in addition to the green image region GR, the white image region WH, the blue image region BL, the yellow image region YL, and the black image region KR. The portion H1, the second peeling portion H2, and the third peeling portion H3 are converted into a light blue image region MI.

  Returning to FIG. 2, in the subsequent step S <b> 7, processing for comparing the image area of the inspection image IG with the image area of the non-defective image GG is performed. Here, the container inspection unit 19 compares the image data of the target pixel of the inspection image IG with the pixel data of the pixel of the non-defective image GG located at the same position as the target pixel. Next, in step S8, a process for determining whether or not the image areas of the same specific color match is performed. Here, the container inspection unit 19 determines whether or not the image regions of the same specific color match depending on whether or not there is a pixel in the inspection image IG that does not match the compared image data.

With reference to FIG. 6, an example of the process in step S7 and step S8 will be described.
As shown in FIG. 6, in this embodiment, in step S7, the container inspection unit 19 selects a pixel located at one of the four corners of the inspection image IG as the first pixel of interest P, and thereafter sequentially. While selecting adjacent pixels as the target pixel P, the image data of each pixel is compared. Incidentally, in FIG. 6, in the X and Y coordinate system in which the pixel at the lower left corner of the inspection image IG is the origin 0, the pixel located at the origin is selected as the first pixel of interest P, and thereafter the pixels adjacent to the right side and the upper side are selected. The target pixel P is sequentially selected along the X axis and the Y axis, for example, in the order of raster scanning. Then, in the X and Y coordinate system in which the origin 0 is the pixel at the lower left corner of the non-defective image GG and the image data of the target pixel P that sequentially moves in the inspection image IG, the same coordinates (X, Y) as the coordinates (X, Y) of the target pixel P The image data of the corresponding pixel Pg located at X, Y) is compared.

  In step S8, as shown by coordinates (X1, Y1) in FIG. 6, when the target pixel P is located in the first peeled portion H1, the same coordinates (X1, Y1) as the target pixel P in the non-defective image GG. The image data of the corresponding pixel Pg located in () is different from the image data of the pixel of interest P. That is, since it is rare that the color of the container 20 matches the color of the non-defective label, the specific color converted from the color of the container 20 is usually different from the specific color converted from the color of the non-defective label. The compared image data do not match. Therefore, the container inspection unit 19 determines that each pixel in the first peeling portion H1 in the inspection image IG is a pixel that does not match the compared image data.

  Similarly, as shown by coordinates (X2, Y2) in FIG. 6, when the target pixel P is located in the second peeling portion H2, the position is at the same coordinates (X2, Y2) as the target pixel P in the non-defective image GG. The image data of the corresponding pixel Pg is different. Further, as shown by coordinates (X3, Y3) in FIG. 6, when the target pixel P is located in the third peeled portion H3, it is located at the same coordinates (X3, Y3) as the target pixel P in the non-defective image GG. The image data of the corresponding pixel Pg is different. Therefore, the container inspection unit 19 determines that the pixels in the second peeling portion H2 and the pixels in the third peeling portion H3 in the inspection image IG are pixels whose compared image data do not match.

  Returning to FIG. 2, as a result of the determination process in step S8, when it is determined that the image areas of the same specific color match (step S8, YES), the process proceeds to step S9 and the inspection target label 30 is determined to be good. This process is performed, and the inspection process of the container 20 ends. That is, in the inspection image IG, when the first peeling portion H1, the second peeling portion H2, and the third peeling portion H3 are not present, there is no pixel in which the compared image data does not match. Therefore, it is determined that the image areas of the same specific color match, and it is determined that the inspection target label 30 is good. In addition to the peeling of the label, the label 30 to be inspected can be detected not only for the label tearing but also for the label curling and the contamination due to the different color adhesion of the label.

  On the other hand, as a result of the determination process in step S8, if it is determined that the image areas of the same specific color do not match (NO in step S8), the process proceeds to step S10 and the process determines that the inspection target label 30 is defective. Is done. That is, when there is at least one of the first peeling portion H1, the second peeling portion H2, and the third peeling portion H3 in the inspection image IG, there is a pixel that does not match the compared image data. The container inspection unit 19 determines that the inspection target label 30 is defective.

According to the above embodiment, the following effects can be obtained.
(1) Depending on whether or not the image areas of the specific color of the used inspection image IG coincide with the image area of the non-defective image GG of the same specific color as the image area of the inspection image IG The quality of the inspection target label 30 can be determined. Therefore, regardless of the color of the contents or the color of the container 20, it is possible to perform an appearance inspection with high accuracy of quality determination for the container 20 having the inspection target label 30 attached to the outer surface. In addition, since inspection is performed by converting to a specific color, even when an imaging device having a small number of pixels is used, high-precision appearance inspection is possible.

  (2) The image of the label 30 to be inspected is converted into an image of a specific color belonging to each of divided color spaces obtained by dividing the HSV color space indicated by hue, brightness, and saturation, which makes it easy for humans to recognize the difference in color. Therefore, the appearance inspection of the inspection object label 30 can be appropriately performed as in the case where a person performs the inspection. In the partial image BG, for example, even if there is a variation in hue, saturation, or lightness of the imaged label 30 to be inspected depending on the lighting condition of the lighting device LT, it is easily converted into one specific color. The quality of the container 20 can be determined by comparing the image data.

In addition, you may change the said embodiment into another embodiment as follows.
In the above-described embodiment, it is preferable that the container inspection unit 19 performs the pass / fail determination of the inspection target label 30 by using an outline area indicating the outline of the image area among the image areas of one specific color. This modification will be described with reference to the drawings.

  As shown by thick lines in FIGS. 7A and 7B, in this modification, the container inspection unit 19 is centered on a boundary line between two adjacent specific color image areas, and a predetermined distance on both sides thereof. An image area composed of a plurality of pixels existing inside is defined as an outline area of each specific color image area. As a result, in the non-defective image GG, the thick line portion indicated by reference numerals SE1 to SE4 in FIG. 7A is the outline region, and in the inspection image IG, the thick line portion indicated by reference signs SE1 to SE7 in FIG. It is said. The container inspection unit 19 performs the inspection processing from step S7 onward using the inspection image IG and the non-defective image GG respectively indicated by the contour region.

  For example, the container inspection unit 19 sequentially selects each pixel located in the contour region of the inspection image IG indicated by reference numeral SE1 in FIG. 7B as the target pixel P, and the non-defective image corresponding to the position of the target pixel P Each image data is compared with the corresponding pixel Pg of GG. If there is a difference between the contour regions as a result of the comparison, there will be a difference between the image data of the pixel of interest P and the image data of the corresponding pixel Pg. 30 is determined to be defective. On the other hand, if there is no difference between the contour regions as a result of the comparison, there is no difference between the image data of the pixel of interest P and the image data of the corresponding pixel Pg. It is determined that the target label 30 is good.

  Incidentally, in the present embodiment, in the inspection image IG, when a pixel located in the contour region indicated by reference numerals SE5, SE6, and SE7 is selected as the pixel of interest P, the pixel of interest is different from the contour region of the non-defective image GG. The image data of P and the image data of the corresponding pixel Pg of the non-defective image GG do not match, and there is a difference.

According to this modification, in addition to the effects (1) and (2) of the above embodiment, the following effects are obtained.
(3) Since the pass / fail determination is performed using the contour area of the image area of the specific color, the image area (number of pixels) used for the determination process is smaller than when the pass / fail determination is performed using the entire area of the image area of the specific color. Thus, it is possible to reduce the processing load related to the determination.

  In the above-described embodiment, the container inspection unit 19 uses the image area of at least one specific color among the plurality of specific colors in the inspection image IG without using the non-defective label image YG. You may make it perform determination.

  For example, in the case where the label 30 is peeled off or torn within one specific color image area in the inspection target label 30, the one specific color image area is used in the inspection image IG without using the non-defective image GG. Thus, it is possible to determine whether the inspection target label 30 is good or bad.

  As an example of this, as shown in FIG. 5B, when the substantially peeled second peeled portion H2 is present in the blue triangular display in the reflected image SG, the blue color in the converted inspection image IG A light blue image area MI other than blue is present inside the image area BL. Therefore, in the process of sequentially selecting the target pixel P by raster scan, the container inspection unit 19 confirms that the image data of the selected target pixel P has changed from blue to light blue different from blue and then to blue again. By detecting, it detects that the light blue image area | region MI different from blue exists in the blue image area | region BL. As a result, the container inspection unit 19 can determine that the inspection target label 30 is peeled off or torn.

  Alternatively, when the display object 30 is an image in which all the display objects are configured by straight lines, the contour shape of the image area of each specific color is configured by a continuous straight line in the inspection image IG. Therefore, in the contour shape of one specific color image area, when there is a cut in one straight line, the cut indicates peeling or tearing of the inspection target label 30. Therefore, in such a case, it is possible to determine pass / fail of the inspection target label 30 by using the contour shape of the image area of the specific color in the inspection image IG without using the non-defective image GG.

  As an example of this, as shown in FIG. 5B, when the first peeling portion H1 is present between the green display and the triangular white display in the reflected image SG, the converted inspection image IG In the straight line portion on the lower side of the white image area WH, there is a cut where the straight line is interrupted at least partially. Therefore, in the process of sequentially selecting each pixel corresponding to the underline of the white image region WH as the target pixel P, the container inspection unit 19 changes the image data of the selected target pixel P from white to light blue in the middle. When it is detected that the color has returned to white again, the container inspection unit 19 can determine that the inspection target label 30 is peeled off or torn.

According to this modification, in addition to the effect (2) of the above embodiment, the following effect is obtained.
(4) By using an image area of one specific color in the inspection target label 30, for example, the color of the content in the partial image BG imaged due to the lighting condition of the lighting device LT or individual differences regarding the imaging performance of the imaging device 12, etc. Or even if the color of the container 20 varies including the inspection target label 30, it is possible to perform the pass / fail determination. That is, regardless of the color of the contents or the color of the container 20, it is possible to perform the appearance inspection of the container having the inspection target label 30 mounted on the outer surface without using the non-defective label image YG.

  Of course, when the pass / fail determination of the inspection target label 30 is performed using only the inspection image IG without using the non-defective label image YG, the container inspection apparatus 11 stores the non-defective label image YG and the image specifying unit 15. 17 may not be provided. The image conversion unit 18 may convert only the reflected image SG into a plurality of specific color inspection images IG.

  In the above embodiment, each time the process of comparing the image data of the target pixel P and the corresponding pixel Pg is performed, the image data of each pixel of the captured image SG and the image data of each pixel of the corresponding image TG are respectively set to a specific color. You may make it convert into image data. In this case, prior to the processing in step S5, the container inspection unit 19 may sequentially select the target pixel P in the captured image SG, and select the corresponding pixel Pg in the corresponding image TG each time the target pixel P is selected. .

  In the above embodiment, the plurality of specific colors do not necessarily have to belong to each of the divided color spaces obtained by dividing the HSV color space indicated by hue, lightness, and saturation. For example, it may be an R, G, B color space, or a CMYK color space of cyan (C), magenta (M), yellow (Y), and black (K). Further, the number of divided color spaces is not limited to eight. For example, the number of divided color spaces may be more than eight to improve detection accuracy.

  In the above embodiment, the imaging device 12 may not be configured to have only one. That is, the container inspection apparatus 11 may include a plurality of imaging devices 12 and generate the partial images BG by operating the respective imaging devices 12 on the containers 20 conveyed by the belt conveyor 14. At this time, the appearance inspection of the container 20 may be performed using each of the plurality of generated partial images BG, or the container may be formed using one partial image BG obtained by synthesizing the plurality of partial images BG. Twenty visual inspections may be performed.

  In the above embodiment, the display object may not be the same size as the inspection target label 30 but may be printed inside the inspection target label 30. At this time, when the inspection target label 30 is colored instead of being colorless (transparent), the appearance inspection processing may be performed using the entire inspection target label 30 as a display object. Of course, when the inspection target label 30 is colorless, the appearance inspection of the container 20 may be performed using only the image portion of the printed display object, not the entire inspection target label 30, as the inspection target label 30.

  In the above embodiment, the inspection target label 30 may be wound around the entire outer surface of the container 20. At this time, the inspection target label 30 may be a shrink label that is contracted by heating or the like and wound around the entire outer surface of the container 20.

  In the above embodiment, the container 20 is not limited to a bottle having a cylindrical shape on the outer surface (round bottle), but may be a square bottle having an outer surface that is prismatic. The container 20 is not limited to a bottle, and may be a can or a bottle. The material of the container 20 is not limited to resin, but may be paper or metal.

  In the above embodiment, the object of the appearance inspection is not limited to the label. For example, a print image printed directly on the outer surface of the container 20 or a print image printed on the transfer sheet is transferred to the outer surface of the container 20. It may be an inspection object display such as a transferred image. Further, the inspection target display may not be an image formed by printing, for example, a display formed by applying paint on the outer surface of the container 20 or processing the outer surface by cutting or the like. Good. In this way, when the inspection object display is other than the label, it is possible to detect a lack of display or a blur.

  DESCRIPTION OF SYMBOLS 11 ... Container inspection apparatus, 12 ... Imaging equipment, 13 ... Control apparatus, 15 ... Memory | storage part, 16 ... Imaging part, 17 ... Image specific | specification part, 18 ... Image conversion part, 19 ... Container inspection part, 20 ... Container, 30 ... Inspection target label, BG: partial image, BL, GR, KR, MI, WH, YL: image region, GG: non-defective image, IG: inspection image, SG: embedded image, TG: compatible image, YG: non-defective label image.

Claims (4)

A container inspection device for inspecting the appearance of a container having a label to be inspected on its outer surface,
A storage unit for storing a non-defective label image representing the whole of the non-defective label attached to the outer surface of the container;
An imaging unit that images the container and generates a partial image of the container including at least a part of the label to be inspected;
By specifying the reflected image of the inspection target label reflected in the partial image, comparing the non-defective label image with the partial image, and specifying the position of the partial image in the non-defective label image, An image specifying unit for specifying a reflected image of the label to be inspected reflected in a partial image, and for specifying a corresponding image corresponding to the reflected image in the non-defective label image ;
An image conversion unit that converts the reflected image into a plurality of specific color inspection images and converts the corresponding image into the plurality of specific color non-defective images ;
Using the image area of at least one specific color among the plurality of specific colors in the inspection image and the image area of the non-defective image having the same specific color as the one specific color, the pass / fail determination of the inspection target label A container inspection unit for performing
A container inspection apparatus comprising: The container inspection apparatus according to claim 1, wherein the container inspection unit performs the pass / fail determination of the inspection target label using an outline region indicating an outline of the image region among the image regions of the one specific color. Wherein the plurality of specific colors, hue, lightness, container inspection apparatus according to claim 1 or 2 which is a color belonging to each of the divided color space obtained by dividing the color space indicated by the color saturation in a plurality. A container inspection method for inspecting the appearance of a container having a label to be inspected on its outer surface,
A storage step of storing a non-defective label image representing the whole of the inspection target label of the non-defective product mounted on the outer surface of the container;
An imaging step of imaging the container and generating a partial image of the container including at least a part of the inspection target label;
By specifying the reflected image of the inspection target label reflected in the partial image, comparing the non-defective label image with the partial image, and specifying the position of the partial image in the non-defective label image, An image specifying step of specifying a reflected image of the inspection object label reflected in the partial image and identifying a corresponding image corresponding to the reflected image in the non-defective label image ;
An image conversion step of converting the reflected image into a plurality of specific color inspection images and converting the corresponding image into the plurality of specific color non-defective images ;
Using the image area of at least one specific color among the plurality of specific colors in the inspection image and the image area of the non-defective image having the same specific color as the one specific color, the pass / fail determination of the inspection target label Performing a container inspection step;
A container inspection method comprising: