JP5504234B2 - Box filling inspection method and box filling inspection apparatus - Google Patents

Description

  The present invention relates to a method and an apparatus for inspecting a boxed state of a boxed product in which a plurality of articles are arranged in an aligned state. In particular, the present invention relates to a final glass bottle formed in a bottle factory. The present invention relates to a packaging inspection method and a packaging inspection apparatus suitable for inspecting whether or not the packaging state is appropriate when packaging a plurality of boxes at a time.

  For example, in a bottle factory, glass bottles formed by a bottle making machine are transferred from a bottle forming process by a bottle making machine to a slow cooling process and cooled, and then sent to an inspection process, as shown in FIG. The glass bottle is inspected for defects. The glass bottles determined to be free from defects in this inspection process are sent to the packaging process, and as shown in FIG. 17, a plurality of glass bottles 1 aligned by the automatic boxing machine 9 remain aligned. It is gripped and packed into the carton 10 on the conveyor 90. In this embodiment, the plurality of glass bottles 1 are packed in an aligned state in which the vertical and horizontal directions are aligned. However, the present invention is not limited to this, and other states such as an aligned state in which the vertical and horizontal directions are aligned are also known. It may be packed in the aligned state. The carton 10 is generally made of corrugated cardboard, and an upper lid 11 is pushed open to open, and a plurality of glass bottles 1 are filled simultaneously through the opening. A partition plate is loaded inside the carton 10 as necessary. Thereafter, the lid 11 of the boxed product 12 is closed and packed, and the packaged product is stacked or shipped in a plurality of stages (see, for example, Patent Document 1).

  In the packaging process, the boxed product 12 is sent to the boxing inspection machine 100 before packing, and it is inspected whether the boxed state is appropriate. Conventionally, a weight checker is used in this type of boxing inspection machine 100. The boxed material 12 is guided onto a belt conveyor 101 that constitutes a weight checker, and the entire weight is measured while being conveyed by the belt conveyor 101 (for example, see Patent Document 2). If the weight measured by the weight checker deviates from the allowable value based on the sum of the weight of the carton 10 and the total weight of the glass bottle 1 (the weight of one bottle x the number of bottled bottles), the packaging is defective. As a result, the package 12 is removed from the conveyance path of the packaging process. It is inspected whether or not a predetermined number of glass bottles 1 are filled in the carton 10 by the inspection by the weight checker, but this is shown in order to inspect the bottle 1 for poor storage (for example, climbing up). However, a detection plate made of synthetic resin is lowered to near the height of the bottle mouth in the carton 10 and it is also detected whether the detection plate does not collide with anything during the lowering.

JP-A-7-156973 JP 2002-116079 A

  However, in the boxed inspection method using a weight checker, in the case of a boxed product packed with small glass bottles, the acceptable value range for pass / fail judgment is larger than the weight of one glass bottle, and one glass bottle is used. Even if it is missing, this cannot be detected. Further, even when a foreign substance having a weight not exceeding the allowable value is mixed, it cannot be detected. In addition, even if a bottle is packed upside down in the carton 10 or any glass bottle is broken in the carton 10, it does not affect the weight. No boxing failure can be detected. Further, when the broken piece of the glass bottle exists on the bottom of the carton 10 or the like, it cannot be detected by the detection plate. In the case of a boxed product in which a partition plate is loaded inside the carton 10, even if the partition plate is bent or deformed, it cannot be detected by a weight checker or a detection plate. In addition, since the weight of the partition plate is small, the lack of the partition plate cannot be detected by the weight checker.

  The present invention has been made paying attention to the above-mentioned problems, and can detect various boxing defects such as missing items and partition plates in a non-contact manner regardless of the type, shape, weight, etc. of the items to be boxed. It is an object of the present invention to provide a boxing inspection method and a boxing inspection apparatus capable of performing the above.

The method for inspecting a boxed product according to the present invention is for inspecting a boxed state of a boxed product in which a plurality of articles are arranged in a state of being arranged through a partition plate . All of the gray images in each inspection area are set by setting the position where each article is arranged and arranged, the position where the partition plate is installed, and the inspection area for the reference image consisting of the gray image obtained by imaging from the part. From a gray scale image obtained by imaging a boxed product to be inspected from an opening portion of a box, and a preparation step for acquiring information on a density distribution for each inspection region by extracting information on the density distribution of the pixels for the inspection object image comprising an inspection area set respectively in the same position as the reference image, related to the concentration distribution for all the pixels constituting the grayscale image of each inspection area After extracting the examination information, which comprises carrying out an inspection step to determine the appropriateness of boxed states of packing material to be inspected each examination information in comparison with the reference information serving.

  In the above-described configuration of the present invention, the “density distribution” refers to a distribution of density levels (density values) or brightness (brightness values) for all pixels constituting a grayscale image. Also included are images obtained by separating the three colors into three primary colors.

According to the boxed inspection method described above, for example, if there is a missing item in the boxed item to be inspected, or if a defect such as a crack has occurred in the boxed item, the boxed item is in place. Information on the density distribution and reference information for the image to be inspected in each of the inspection areas, such as when there is no object or is upside down, or when a foreign object has entered the box Therefore, it is determined that the boxed state of the boxed item to be inspected is not appropriate. In addition, when there is a missing or deformed partition plate, the information regarding the density distribution for the image to be inspected in any of the inspection areas does not match the reference information. Is not appropriate.
If the boxed items to be inspected do not have the above-mentioned boxing defects, the inspection information and standards regarding the density distribution of the images to be inspected in all the inspection areas, even if there is some variation in the boxed state, such as misalignment of the boxed items Therefore, it is determined that the boxed state of the boxed product to be inspected is appropriate.

In a preferred embodiment of the present invention, the reference information regarding the density distribution for each inspection region acquired in the preparation step is obtained by imaging a plurality of boxed items in an appropriate boxed state from the opening portion of each box. The information on the density distribution of the images in each inspection region extracted from a plurality of reference images is averaged.
According to this embodiment, even if there is a variation in the reference image, the variation is absorbed and reflected in the reference information, so that the inspection is stabilized and the inspection accuracy is improved.

  In the above-described configuration of the present invention, there are several modes for information serving as a reference regarding the concentration distribution acquired in the preparation step and inspection information regarding the concentration distribution extracted in the inspection step. Is a density histogram representing the distribution of the number of pixels for each gradation of density levels for all pixels included in the inspection area as a two-dimensional graph, and the density histogram and the reference image for the inspection target image for each inspection area The degree of similarity with respect to the density histogram is obtained by calculation, and the calculated value is compared with a predetermined threshold value to determine whether the boxed state of the boxed product is appropriate.

According to this embodiment, if there is a packing defect in the boxed item to be inspected, the density histogram for the image to be inspected in any of the inspection areas does not match the density histogram for the reference image. The calculated value indicating the degree of similarity exceeds the threshold value, and it is determined that the boxed state of the boxed product to be inspected is not appropriate.
If there are no boxing defects in the boxed items to be inspected, the density histogram for the image to be inspected in all the inspection areas and the density histogram for the reference image are almost the same, so the calculated value indicating the degree of similarity obtained by the calculation Does not exceed the threshold value, and it is determined that the boxed state of the boxed object to be inspected is appropriate.

The boxing inspection apparatus according to the present invention inspects the boxing state of a boxing that is boxed in a state where a plurality of articles are arranged via a partition plate , and conveys the boxing to be inspected to an inspection position. A mechanism, a camera arranged toward the inspection position so that the entire boxed product is within the field of view above the inspection position, an illumination device for illuminating the inspection position, and a boxed object to be inspected The image processing apparatus includes an image processing apparatus that captures an image to be inspected including a grayscale image obtained by imaging with the camera from an opening portion of the box, and executes image processing for determining whether the boxed state of the boxed product is appropriate. The image processing apparatus includes: an inspection area setting unit that sets an inspection area at a position where the articles are aligned and a position where a partition plate is installed for the inspection target image; The density information extracting means for extracting the inspection information relating to the density distribution for all the pixels constituting the grayscale image in the inspection area, and the boxed product in an appropriate boxed state is obtained by imaging with the camera from the opening of the box. a reference information storage means for storing reference information serving about by said density distribution obtained from the reference image consisting of gray image, each test information extracted by the density information extraction means is stored in said reference information storage unit A discriminating means for discriminating whether the boxed state of the boxed product is appropriate in comparison with the reference information; and an output means for outputting a discrimination result by the discriminating means; It is provided.

When inspecting the boxed state of the boxed item by the above-described boxed item inspection apparatus, the boxed item to be inspected is guided to the inspection position by the transport mechanism. In a state in which the entire box is within the field of view of the camera, the box is imaged by the camera under illumination by the lighting device. The inspection target image of the boxed product imaged by the camera is taken into the image processing apparatus, and image processing for determining whether the boxed state of the boxed product is appropriate is executed.
The inspection area setting means of the image processing apparatus sets the inspection area at the position where the articles are arranged and arranged with respect to the inspection target image and the position where the partition plate is installed , and the density information extraction means is determined by the inspection area setting means. The inspection information relating to the density distribution for all the pixels constituting the grayscale image in each set inspection area is extracted. And serving as a reference information on the density distribution obtained from the reference image consisting of gray image obtained by imaging by the camera than the opening portion of the packing material boxes proper packing state is stored in the reference information storage means The discriminating unit discriminates whether or not the boxed state of the boxed product is appropriate by comparing each inspection information extracted by the concentration information extracting unit with the reference information stored in the reference information storage unit. The discrimination result by the discrimination means is output by the output means.

  In the above-described configuration of the boxed product inspection apparatus according to the present invention, each unit configuring the image processing apparatus is preferably realized by a programmed computer, but is not limited thereto, and may be realized by a dedicated hardware circuit. it can.

  The boxed product inspection apparatus according to the present invention is suitable for inspecting a boxed product in a state in which a plurality of glass bottles are aligned, but is not limited to this, and a boxed product of various articles is an inspection target. be able to.

The preferable boxed inspection apparatus of the present invention further includes a sensor for detecting that the boxed item to be inspected is introduced into the inspection position, and the camera is inspected when a detection signal is output from the sensor. An image of a boxed object to be inspected passing through the position is captured.
According to the boxed product inspection apparatus of this embodiment, it is possible to automate the inspection, and it is possible to inspect while transporting the boxed product to be inspected, and the inspection efficiency is improved.

  According to this invention, regardless of the type, shape, weight, etc. of the boxed items, various boxing defects such as missing items and partition plates can be detected in a non-contact manner, and even if the boxed state such as misalignment of the boxed items is detected. Even if there is some variation in the packaging, it is possible to reliably detect a packaging failure.

It is explanatory drawing which shows schematic structure of the boxed goods inspection apparatus of this invention. It is a top view which shows the inspection position of a boxing thing. It is explanatory drawing which looked at the structure of the camera and the illuminating device from the downward direction. It is a block diagram which shows the structure of an image processing apparatus. It is explanatory drawing which shows the image of a boxing thing, and the test | inspection area | region set to the image. It is explanatory drawing which shows a part of image of the boxing thing with the lack of a bottle. It is explanatory drawing which shows a part of image of the box stuffing containing the bottle upside down. It is explanatory drawing which shows a part of image of the box stuffing which a part is damaged and contains a bottle. It is explanatory drawing which shows the image of the boxing thing which has a partition plate, and the test | inspection area | region set to the image. It is explanatory drawing which shows a part of image of the box stuffing with the lack of a bottle. It is explanatory drawing which shows a part of image of the boxing thing with the lack of a partition plate. It is explanatory drawing which shows the density | concentration histogram about the test | inspection image of the boxing thing of a proper boxing state. It is explanatory drawing which shows the density | concentration histogram about the test | inspection image of the boxing thing of an improper boxing state. It is a flowchart which shows the flow of the image process in the preparation process of an image processing apparatus. It is a flowchart which shows the flow of the image process in the test | inspection process of an image processing apparatus. It is explanatory drawing which shows a bottle making process. It is explanatory drawing which shows the packing operation | work in a packaging process.

FIG. 1 shows a schematic configuration of a boxed product inspection apparatus according to an embodiment of the present invention.
The illustrated boxed inspection apparatus inspects whether or not the boxed state of the boxed goods 12 packed in a state where a plurality of glass bottles 1 are aligned in the vertical and horizontal directions is appropriate. The boxing inspection apparatus of the present invention is not limited to the glass bottle 1, but is not limited to a glass container other than a bottle, further not limited to a glass container, but a container made of synthetic resin, metal, ceramic, etc. The boxed state can be inspected also for the boxed items packed in a state where the articles are aligned. Moreover, although the carton 10 filled with the glass bottle 1 is made of corrugated cardboard, it is not limited to this and may be made of synthetic resin. Furthermore, the form of the carton 10 is not limited to a rectangular parallelepiped.

  The boxed product inspection apparatus in the illustrated example includes a transport mechanism 5 for transporting a boxed product 12 to be inspected in which a glass bottle 1 formed by a bottle making machine (not shown) is boxed, and a boxed product 12 to be inspected at an inspection position. The illumination device 4 for illuminating the camera, the camera 2 for imaging the boxed item 12 to be inspected from above under the illumination by the lighting device 4, and the inspection target image of the boxed item 12 imaged by the camera 2 by taking in the boxed item 12 The image processing apparatus 3 executes image processing for determining whether or not the boxed state is appropriate.

  As shown in FIGS. 1 and 2, the transport mechanism 5 includes a belt conveyor 50 installed at an inspection position, an upstream roller conveyor 51 for guiding the boxed goods 12 to be inspected to the belt conveyor 50, and an inspected And a roller conveyor 52 on the downstream side for deriving the boxed product 12 from the inspection position. The inspection position for inspecting the boxed object 12 to be inspected is set immediately below the camera 2. The camera 2 is controlled by the image processing apparatus 3 so that the shutter is released when the center of the boxed item 12 to be inspected reaches the inspection position. The camera 2 is installed at a predetermined height position with the objective lens 20 facing downward so that the entire boxed item 12 is within the field of view at the inspection position.

  On one side of the belt conveyor 50, a passage sensor 7 constituted by a photoelectric sensor or the like is installed at a position through which the boxed product 12 passes. The passage sensor 7 outputs a detection signal s to the image processing device 3 when the boxed material 12 is detected. After the detection signal s is output, the package 12 reaches the inspection position that is the center of the visual field of the camera 2 at a timing delayed by a predetermined time T. After inputting the detection signal from the passage sensor 7, the image processing apparatus 3 outputs to the camera 2 a shutter drive signal i for driving the shutter delayed by the time T. In FIG. 2, reference numeral 53 denotes an antireflection sheet made of sponge or the like for preventing the illumination light 4 from being reflected by the illuminating device 4, and is disposed along both sides of the belt conveyor 50.

  As shown in FIG. 3, the illumination device 4 includes four light emitters 41 to 44 arranged so as to surround the camera 2. Each of the light emitters 41 to 44 is configured by arranging a plurality of light emitting diodes 40 in a matrix. In this embodiment, a near-infrared light emitting diode is used as the light emitting diode 40. However, the present invention is not limited to this, and infrared rays, ultraviolet rays, visible rays, X-rays, etc. can be used as long as imaging is possible according to the combination with the camera 2. Light sources such as light emitting diodes that emit light of other wavelengths can be used.

  Although not shown in the figure, an automatic boxing machine that performs an operation of packing the glass bottle 1 into the carton 10 is installed at the upstream position of the roller conveyor 51 on the upstream side. In addition, a reject device 54 (shown in FIG. 4) is provided at the downstream position of the roller conveyor 52 on the downstream side to discharge the boxed material 12 that has been determined to be unpacked to a reject table on the side of the roller conveyor 52. . The reject device 54 operates upon receiving a reject signal r from the image processing device 3. The boxed product 12 determined to be in a proper boxed state is sent to the packing process by the roller conveyor 52, and the lid 11 is closed by an automatic packing machine (not shown) and packed.

  FIG. 4 shows a configuration example of the image processing apparatus 3. The image processing apparatus 3 of the illustrated example is configured by a personal computer, and includes a CPU 30 that is a main body of control and calculation, a hard disk 31 that stores data such as programs and images, and a memory 32 that is used for data read / write operations and the like. Is included. The passage sensor 7 and the reject device 54 are connected to the CPU 30 via a dedicated I / O port 33. The detection signal s output from the passage sensor 7 is taken into the image processing apparatus 3 via the I / O port 33. The reject signal r output from the image processing device 3 is given to the reject device 54 via the I / O port 33.

The camera 2 is connected to the CPU 30 via the USB port 34. The camera 2 receives a shutter drive signal i output from the USB port 34 of the image processing apparatus 3 and performs an imaging operation. The grayscale image data D obtained by the imaging is taken into the image processing apparatus 3 from the USB port 34. The camera 2 is a CCD camera, for example, and the grayscale image data D obtained by imaging the boxed product 12 is stored in the hard disk 31 of the image processing device 3 in units of one pixel.
A display 6 is connected to the CPU 30. The display 6 displays an image captured by the camera 2, a density histogram (to be described later), an inspection result, and the like.

FIG. 5 shows an inspection object image 12p obtained by imaging the boxed object 12 to be inspected by the camera 2 from the opening. According to this inspection object image 12p, the inside of the carton 10 is packed with 20 cylindrical glass bottles 1 aligned vertically and horizontally with the bottle mouths facing up. In the inspection target image 12p of FIG. 5, 1p is an image of the glass bottle 1, 10p is an image of the carton 10, and 11p is an image of the lid 11 of the carton.
When the inspection object image 12p of the boxed product 12 is taken into the image processing apparatus 3 from the camera 2, the image processing apparatus 3 positions the image bottles on the image corresponding to the positions where the glass bottles are aligned with respect to the inspection object image 12p ( By setting rectangular inspection areas W at a total of 20 locations, extracting inspection information relating to the density distribution of the inspection target images in each inspection area W, and comparing each inspection information with the reference information The suitability of the boxed state is determined.
The image processing apparatus 3 extracts the vertical and horizontal contour lines of the carton image 10p when the image captured from the camera 2 is tilted, and rotates the image so that each contour line is horizontal and vertical. The inspection area is set after the image rotation processing.

By the image processing by the image processing device 3, the glass bottle is missing (see FIG. 6), the glass bottle is upside down (see FIG. 7), and the glass bottle is broken (see FIG. 8). ) And the like can be detected.
In the inspection target image 12p shown in FIG. 6, X indicates that the glass bottle image 1p is missing. Since the density distribution of the inspection target image in the inspection region W set in the portion X is clearly different depending on the presence or absence of the glass bottle image 1p, it is possible to determine the lack of the glass bottle.
In addition, in the inspection target image 12p shown in FIG. 7, the image 1p of the central glass bottle is an inverted state with the bottle bottom facing up. The density distribution of the image to be inspected in the inspection area W set in this portion is apparent depending on whether the glass bottle image 1p is an image showing the bottle mouth or an image showing the bottle bottom. Since it is different, it is possible to determine that the glass bottle is upside down.
Further, in the inspection object image 12p shown in FIG. 8, the image 1p of the central glass bottle is a state in which the side surface portion of the glass bottle is broken. Since the density distribution of the image to be inspected in the inspection area W set in this portion is clearly different depending on whether or not the glass bottle image 1p is broken, it is possible to determine whether the glass bottle is broken.

FIG. 9 shows another inspection object image 12p obtained by imaging the inspection object box 12 with the camera 2 from the opening. According to this inspection object image 12p, 20 cylindrical glass bottles 1 are boxed in the carton 10 in a state where they are vertically and horizontally aligned through vertical and horizontal partition plates with the bottle mouth facing upward. 9, 1p is an image of the glass bottle 1, 10p is an image of the carton 10, 11p is an image of the carton lid 11, and 13p and 14p are images of vertical and horizontal partition plates.
When the inspection object image 12p of the boxed product 12 is taken into the image processing apparatus 3 from the camera 2, the image processing apparatus 3 positions the image bottles on the image corresponding to the positions where the glass bottles are aligned with respect to the inspection object image 12p ( A rectangular inspection region W1 is formed in a total of 20 places), and a vertically long rectangular inspection region W2 is formed at positions (16 places in total) on the image corresponding to the positions where the vertical partition plates between the adjacent glass bottles are installed. The inspection area W3 having a horizontally long rectangular shape is arranged at positions on the image corresponding to the positions where the horizontal partition plates between the adjacent glass bottles are installed (a total of 15 locations), the vertical partition plate and the horizontal partition plate. A square rectangular inspection region W4 is set at each of the positions on the image corresponding to the intersection positions with each other (a total of 12 locations), and inspection information relating to the density distribution of the inspection target images in each of the inspection regions W1 to W4 is obtained. The Extracted, to determine the appropriateness of boxed states boxed product by each inspection information comparing the reference information.

The image processing by the image processing device 3 detects defective packing such as missing or climbing glass bottles (see FIG. 10) and missing partition plates (see FIG. 11). be able to.
In the inspection target image 12p shown in FIG. 10, X indicates that the image 1p of the glass bottle is missing, and Y indicates that the glass bottle has been climbed and the glass bottle below it has been hidden. The density distribution of the inspection target image in the inspection region W1 set in the X and Y portions is obviously different depending on the presence or absence of the glass bottle image 1p and the overlap of the glass bottle images 1p. It is possible to determine missing or climbing.
Further, in the inspection object image 12p shown in FIG. 11, Z indicates that the image 14p of the horizontal partition plate is missing. Since the density distribution of the inspection target image in the inspection region W3 set in the Z portion is clearly different depending on the presence or absence of the partition plate image 14p, it is possible to determine the absence of the partition plate.

FIG. 12 and FIG. 13 show density histograms h1 and h2 extracted as information regarding density distribution from the images in each inspection region W or W1 to W4 for the reference image and the inspection target image. The density histograms h1 and h2 shown in the drawing are two-dimensional graphs showing the distribution of the number of pixels for each gradation of the density level for all the pixels included in the inspection area. 12 and 13, the horizontal axis indicates the density level of 0 to 255 gradations, and the vertical axis indicates the number of pixels having each density level.
12 and 13, h2 indicates a density histogram extracted from one of the plurality of inspection areas set in the reference image in the preparation process, and h1 is set as an inspection target image in the inspection process. A density histogram extracted from one of the plurality of inspection areas (the same inspection area as the reference image) is shown.

  In the preparation step, the inspection areas W or W1 to W4 are respectively set for the reference images obtained by capturing the boxed goods 12 in an appropriate boxed state with the camera 2 from the opening of the carton 10, This is a step of extracting a density histogram for the reference image. In this embodiment, a plurality of boxed articles 12 in an appropriate boxed state are prepared, and density histograms for images in each inspection region are extracted for a plurality of reference images obtained by imaging each boxed article 12. Thereafter, the density histogram of the reference image is obtained by averaging the density histograms for a plurality of reference images for each inspection region, that is, by obtaining the average value of the number of pixels for each gradation of the density level. .

In the inspection step, the inspection regions W or W1 to W4 are set for the inspection target images obtained by imaging the boxed articles 12 to be inspected with the camera 2 from the opening portion of the carton 10, and the inspection regions W This is a step of extracting a density histogram for the image to be inspected. In the inspection process, the degree of similarity between the density histogram h1 of the inspection target image and the density histogram h2 of the reference image is determined for each inspection region.
In the specific example shown in FIG. 12, the density histogram h1 of the inspection target image and the density histogram h2 of the reference image almost overlap each other and the degree of similarity between them is high, so that the boxed state of the boxed item 12 to be inspected is determined to be appropriate. To do.
In the specific example shown in FIG. 13, since the density histogram h1 of the inspection target image and the density histogram h2 of the reference image do not overlap and the degree of similarity between them is low, it is determined that the boxed state of the boxed item 12 to be tested is not appropriate. To do.

FIG. 14 shows a control procedure of the image processing apparatus 3 in a preparation process executed to acquire the density histogram h2 for the reference image. In the figure, “ST” is an abbreviation for “STEP” and indicates each procedure in the flow of control.
In the preparation process, a plurality of boxed articles 12 in an appropriate boxed state are prepared, and in ST1 of the figure, it is determined whether or not the boxed articles 12 are introduced onto the belt conveyor 50 at the inspection position by the passage sensor 7. Yes. When the passage sensor 7 detects the boxed material 12, the determination in ST1 is “YES”, and the detection signal s output by the passage sensor 7 is taken into the image processing device 3. The image processing device 3 outputs a shutter drive signal i to the camera 2 with a predetermined timing delay, the camera 2 receives the shutter drive signal i, opens the shutter, and images the boxed material 12 from the opening of the carton 10, and the reference image Is acquired (ST2). The grayscale image data D of the reference image is taken into the image processing device 3 and stored in the hard disk 31 (ST3).

  In the next ST4, the CPU 30 of the image processing apparatus 3 sets the inspection region W or W1 to W4 at positions on the image corresponding to the positions where the glass bottles are aligned and the positions where the partition plates are installed with respect to the reference image. After that, a density histogram is generated for the inspection target image in the first inspection region (ST5). Similarly, when density histograms are generated for the images to be inspected in all the inspection regions, the determination in ST6 is “YES”, and the same boxed product 12 in the second proper boxed state is introduced to the inspection position. A density histogram generation process is executed. When density histograms are generated for all of the prepared boxed items 12 in the proper boxed state, the determination in ST7 is “YES”, and a plurality of reference images obtained by imaging each boxed item 12 are inspected regions. Each density histogram is averaged to obtain a density histogram h2 of the reference image, which is stored in the hard disk 31 (ST8).

FIG. 15 shows a control procedure of the image processing apparatus 3 in the inspection process executed to extract the density histogram h1 of the inspection target image and determine whether the boxed state of the inspection target boxed item 12 is appropriate.
In ST1 of the figure, it is determined by the passage sensor 7 whether or not the inspection object box 12 has been introduced onto the belt conveyor 50 at the inspection position. When the passage sensor 7 detects the boxed material 12 to be inspected, the determination in ST1 is “YES”, and the detection signal s output from the passage sensor 7 is taken into the image processing device 3. The image processing apparatus 3 outputs a shutter drive signal i to the camera 2 with a predetermined timing delay, the camera 2 receives the shutter drive signal i, opens the shutter, images the boxed material 12 from the opening of the carton 10, and is inspected. An image is acquired (ST2). The grayscale image data D of the inspection target image is taken into the image processing device 3 and stored in the hard disk 31 (ST3).

  In the next ST4, the CPU 30 of the image processing apparatus 3 prepares the inspection target image at a position on the image corresponding to the position where the glass bottle is aligned and the position where the partition plate is installed (the same position as the reference image). After setting the inspection area W or W1 to W4 set in step 1, a density histogram h1 is generated for the inspection target image in the first inspection area (ST5). Next, the CPU 30 calculates the degree of similarity between the density histogram h1 of the image to be inspected and the corresponding density histogram h2 of the reference image acquired in the preparation step by executing the normalized cross-correlation calculation of the following equation.

  In the above equation, i is a gradation of a density level and takes a range of 0 to 255. T (i) is the number of pixels having a density level of i (where i = 0 to 255) with respect to the density histogram h2 of the reference image, and I (i) is about the density histogram h1 of the inspection target image (where i = 0 to 255) is the number of pixels having a density level. R represents a value indicating the degree of similarity calculated by Equation 1, that is, a correlation coefficient. The correlation coefficient R takes a value of 1 or less, and the closer to “1”, the higher the degree of similarity between the density histogram h1 of the inspection target image and the density histogram h2 of the reference image. Although the correlation coefficient R can be used to determine the degree of similarity, in this embodiment, the degree of similarity is determined based on the value E calculated by the following equation (ST6).

  The smaller the value E, the higher the similarity between the density histogram h1 of the inspection target image and the density histogram h2 of the reference image, and the lower the value E, the lower the similarity between the density histogram h1 of the inspection target image and the density histogram h2 of the reference image. .

  Returning to FIG. 15, in the next ST 7, the CPU 30 determines whether or not the value E is equal to or greater than a predetermined threshold value, and stores the determination result in the memory 32. Next, returning from ST8 to ST5, a density histogram h1 is generated for the inspection target image in the second inspection region, and the same procedure as described above is executed (ST6, 7). When the steps ST5 to ST7 are executed for all the inspection objects in the inspection region, the determination in ST8 is “YES”, the CPU 30 refers to the stored contents of the memory 32, and the value E is equal to or greater than the threshold value. It is determined whether there is at least one inspection area (ST9).

If the determination in ST9 is "YES", the CPU 30 determines that the boxed state of the boxed item 12 is not appropriate, outputs a reject signal r to the reject device 54 as a determination output of "inappropriate", and displays 6 Display a determination result of “inappropriate” (ST10). If the value E is smaller than the threshold value for the inspection target images in all the inspection areas, the determination in ST9 is “NO”, and the CPU 30 determines that the boxed state of the boxed item 12 is appropriate, A determination result indicating “appropriate” is displayed on the display 6 (ST11).
In this embodiment, all the images to be inspected are stored in the hard disk 31 of the image processing apparatus 3 with the imaging time as a tag, and this allows the images to be confirmed over the past, thereby ensuring traceability. ing.

DESCRIPTION OF SYMBOLS 1 Glass bottle 10 Carton 2 Camera 3 Image processing apparatus 4 Illumination apparatus 5 Conveyance mechanism 7 Pass sensor W, W1-W4 Inspection area

Claims (6)

A boxing inspection method for inspecting a boxed state of a boxed item in which a plurality of articles are arranged in a state of being arranged through a partition plate, and images an image of the boxed item in an appropriate boxed state from an opening portion of the box With respect to the reference image composed of the obtained gray images , inspection areas are respectively set at positions where the articles are arranged and arranged and positions where the partition plates are installed, and all the pixels constituting the gray image in each inspection area are set. Preparatory process for acquiring information that serves as a reference for the density distribution for each inspection region by extracting information on the density distribution, and an inspection target image comprising a grayscale image obtained by imaging the boxed product to be inspected from the opening of the box the respectively set an inspection area at the same position as the reference image, extracts the examination information on the concentration distribution about the all pixels constituting the grayscale image of the inspection area for After, boxed product inspection method which comprises carrying out an inspection step to determine the appropriateness of boxed states of packing material to be inspected each examination information in comparison with the reference information serving. 2. The boxed product inspection method according to claim 1, wherein the reference information regarding the concentration distribution for each inspection region acquired in the preparation step is a plurality of boxed products in an appropriate boxed state. A boxing inspection method in which information on density distributions of images in each inspection region extracted from a plurality of reference images obtained by imaging from a portion is averaged . The reference information on the density distribution acquired in the preparation process and the inspection information on the density distribution extracted in the inspection process are the distribution of the number of pixels for each gradation of the density level for all the pixels included in the inspection area. Is a density histogram represented by a two-dimensional graph, and the degree of similarity between the density histogram for the image to be inspected and the density histogram for the reference image is obtained for each inspection region by calculation, and the calculated value is set to a predetermined threshold value. The method for inspecting a boxed product according to claim 1 or 2, wherein the suitability of the boxed state of the boxed product is determined by comparing with each of the items. An inspection device for a boxed product for inspecting the boxed state of a boxed product in which a plurality of articles are arranged in a state of being aligned via a partition plate, and a conveyance mechanism for guiding the boxed product to be inspected to an inspection position, and an inspection A camera arranged toward the inspection position so that the entire boxed product is within the field of view above the position, an illuminating device for illuminating the inspection position, and an opening portion of the boxed object to be inspected And an image processing apparatus that executes an image process for determining whether or not a boxed state of a boxed product is appropriate by taking an inspection object image formed of a grayscale image obtained by imaging with the camera from the camera. Are set by the inspection area setting means for setting the inspection area at the position where the articles are aligned and the position where the partition plate is installed, and the inspection area setting means, respectively. In addition, density information extraction means for extracting inspection information relating to the density distribution for all pixels constituting the grayscale image in each inspection area, and a boxed product in an appropriate boxed state is imaged by the camera from the opening of the box Reference information storage means for storing reference information regarding the density distribution acquired from the reference image made up of the obtained grayscale image, and each inspection information extracted by the density information extraction means is stored in the reference information storage means. A box inspection apparatus comprising: a determination unit that determines whether or not the boxed state of the boxed product is appropriate in comparison with the reference information that has been made; and an output unit that outputs a determination result of the determination unit. The boxing inspection apparatus according to claim 4, wherein the boxing is a boxed container in which a plurality of glass bottles are aligned . 6. The boxing inspection apparatus according to claim 4 or 5, further comprising a sensor for detecting that the boxing object to be inspected has been introduced into the inspection position, and the camera outputs a detection signal from the sensor. A box inspection apparatus that images the inspection target box passing through the inspection position .

Images