JP3241666B2 - Glass container defect inspection equipment - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a defect of a body portion of a glass container having a large opening such as a cup, a pot, a vase, a jar, and the like, that is, a glass piece attached to the inner and outer surfaces of the body portion of the container. The present invention relates to an apparatus for detecting foreign substances and dirt, bubbles inside glass, foreign substances, and color unevenness.

[0002]

2. Description of the Related Art Conventionally, some devices for inspecting defects such as glass fragments adhered to a glass container inspect the mouth and the bottom, but none inspect the body. In the torso inspection, the installation position of the device that inspects the mouth was sometimes lowered and used for the torso inspection, but stable inspection could not be performed, and eventually visual inspection had to be relied on. Did not.

[0003]

Conventionally, inspection of defects in the body of a glass container has had to rely on visual inspection.
The inspector is forced to perform severe work, and it is difficult to completely remove the attached glass pieces by visual inspection, and the glass adhered to the inside and outside body of the glass container generated in the opening process of the glass container The complaint had caused complaints that the consumer was hurt. The present invention
Automatically detects such defects in the body of the glass container,
The purpose is to eliminate defective products without relying on visual inspection by inspectors.

[0004]

According to the present invention, there is provided a light source unit for irradiating inspection light from above a glass container to be inspected, and capturing an image of a body of the glass container provided on a side of the glass container. And an image processing unit for processing an image captured by the camera, wherein the image processing unit converts the image with an edge disturbance light image detection luminance threshold.
A first binarization process for binarizing at a predetermined luminance Vs1 and erasing an image equal to or lower than Vs1, and a set predetermined position is horizontally scanned from an edge of the screen on the screen on which the first binarization processing is performed. An edge disturbance light detection process for detecting a disturbance light image of an edge portion of a glass container, an edge disturbance light mask process for vertically masking a portion where the edge disturbance light is detected in a predetermined width, and an edge disturbance light mask The processed image is converted to a predetermined brightness V which is a glass piece image detection brightness threshold.
a second binarization process for binarizing in s2 and erasing an image equal to or lower than Vs2, a labeling process for numbering remaining images with respect to the image subjected to the second binarization process, and a labeled image And a defect judging process for judging that the number of pixels is within a range of predetermined values SL to SH.

[0005] This device has no fold especially in the trunk,
Suitable for glass containers with smooth curves, irradiates the glass container with inspection light from an upper light source, receives reflected / refracted light with a camera, and removes edge disturbance light from the image. Then, the presence or absence of a defect is determined based on the luminance and area of the remaining image.

If the glass container has a defect such as a piece of glass attached thereto, the light from the light source causes irregular reflection or refraction due to the defect. It has become. Therefore,
If the inspected glass container has a portion with a high luminance, the glass container has a defect.In general, when the glass container is irradiated with the inspection light and the image is captured, the luminance other than the defect is reduced. There is an element that increases, that is, disturbance light, and accurate inspection cannot be performed unless the disturbance light is removed. The first of the disturbance lights is the edge disturbance light. This is because a large portion of luminance is generated around the glass container.

The edge disturbance light is removed in the image processing section as follows. First, a first binarization process of binarizing an image captured by a camera at a predetermined luminance Vs1 and erasing images below Vs1 is performed. In this processing screen, a set predetermined position is scanned in the horizontal direction from the screen edge, and edge disturbance light detection processing for detecting a disturbance light image of the edge portion of the glass container is performed. In this case, the predetermined position is usually near the center in the height direction of the glass container, but an appropriate position is determined according to the shape of the container. When the edge disturbance light is detected, an edge disturbance light masking process for masking the portion in the vertical direction by a predetermined width is performed. Thus, the edge disturbance light is masked and removed.

[0010] A second binarization process for binarizing an image from which edge disturbance light has been removed at a predetermined luminance Vs2 and erasing images below Vs2 is performed. This eliminates disturbance light due to various causes. Further, a labeling process for numbering the remaining images is performed, and a defect determination process for determining that the number of pixels of the labeled image is within a range of a predetermined value SL to SH is performed. The size of the defect is empirically SL-SH
In most cases, small dust or the like is attached when the size is less than SL.
If it is larger, it is some sort of disturbance light.

A glass container having a polygonal shape or the like and a fold not formed of only a smooth plane or curved surface at the body portion generates disturbance light at the fold portion in addition to disturbance light at the edge portion. The fold disturbance light is removed in the image processing unit as follows.

A first binarization process for an image captured by a camera, an edge disturbance light detection process for detecting a disturbance light image of an edge portion of a glass container, an edge disturbance light masking process for removing the edge disturbance light, a predetermined process The process is the same as that of the above-described apparatus until a second binarization process of binarizing the image at the luminance Vs2, erasing images below Vs2, and performing a labeling process of numbering the remaining images is performed. When removing disturbance light at the fold portion, a fold portion in which the aspect ratio of the labeled image is equal to or more than a predetermined value is detected, and a fold portion masking process for masking that portion is performed. As a result, fold disturbance light is removed. Thereafter, as in the case described above, a defect determination process for determining that the number of pixels of the image is within the range of the predetermined value SL to SH is performed to determine whether or not the glass container has a defect.

When four cameras are installed at approximately 90 ° intervals and these cameras are simultaneously operated to capture an image of the glass container, not only can the entire circumference of the body of the glass container be inspected, Since a portion masked to remove disturbance light at an edge portion or a fold portion can be inspected with an image of another camera, there is almost no possibility of missing a defect. If a defect is found in one of the images captured by the four cameras, the glass container has a defect.

If the light source section is composed of four linear light sources combined in a cross-girder shape, the light can be evenly applied to the glass container, and an accurate inspection can be performed. Further, if the interval between the two sets of linear light sources in the parallel direction can be freely changed, the interval between the light sources can be adjusted according to the size of the glass container, and the inspection light can be emitted in a preferable state.

The distance w between the linear light sources forming a pair in the parallel direction
Is set to about ± 10 mm of the diameter (inner diameter) of the glass container to be inspected, the luminance of the defect of the image captured by the camera becomes large, and it becomes difficult to overlook the defect.

This apparatus can perform an inspection while a glass container to be inspected is moving on a straight conveyor. In that case, the light source unit is installed directly above the position where the straight conveyor is inspected, and the camera is also arranged therearound. By performing the inspection on the straight conveyor, it is not necessary to provide a special inspection table, and the apparatus cost is reduced. In addition, it is possible to automatically perform inspections and automatically eliminate defective products by providing a control unit that generates inspection timing and timing for removing defective glass containers and operates a removal device such as a pusher. Becomes In this case, a container sensor for detecting the glass container flowing upstream of the inspection position is provided, and the control unit that has received the detection signal of the container sensor calculates the timing at which the glass container reaches the inspection position, and the timing And sends an inspection signal to the image processing unit. The control unit receives a defect signal sent from the image processing unit when the image processing unit makes a defect determination, generates an exclusion timing at which the inspected glass container reaches a position of an exclusion device such as a pusher, and generates the timing. To send a rejection signal to operate the rejection device to reject defective glass containers.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below in detail with reference to the drawings showing embodiments. 1 is an explanatory view showing an outline of the inspection apparatus of the embodiment, FIG. 2 is a top view of a light source unit, FIG. 3 is an explanatory view showing a relationship between a diameter of a glass container and a luminance ratio, and FIG. 4 is a relationship between a glass container and a camera. FIG. 5 is a block diagram showing an algorithm of an edge part disturbance light removal processing and a defect judgment processing in the image processing unit. FIG. 6 is an explanatory view of the image processing in FIG. 5, and FIG. 7 is a fold part disturbance light. FIG. 8 is a block diagram showing an algorithm of the removal processing and the defect determination processing, and FIG. 8 is an explanatory diagram of the image processing in FIG.

The inspection apparatus shown in FIG. 1 includes a light source unit 1, four cameras 2, 3, 4, 5, an image processing unit 6, a control unit 7, a container interval adjusting unit 8, a conveyor speed sensing unit 9, and a container sensor 10. And the elimination device 11. The glass container to be inspected flows on the conveyor 12 from left to right in FIG. 1, and the interval between the containers is adjusted to a predetermined interval by the container interval adjusting unit 8 on the upstream side. The container sensor 10 detects the glass container G flowing at the adjusted interval, and sends a detection signal to the control unit 7. The control unit 7 having received the detection signal calculates the timing at which the glass container reaches the inspection position based on the detection signal and the speed signal from the conveyor moving speed detector 9, and outputs the inspection signal to the image processing unit 6 at that timing. Send to Upon receiving the inspection signal, the image processing unit 6 captures the images of the four cameras 2, 3, 4, and 5, performs predetermined image processing, and, when it determines that there is a defect, sends the defect signal to the control unit 7. send. Upon receiving the defect signal, the control unit 7 sets the glass container to an elimination device 11 such as a pusher.
Is generated, an exclusion signal is sent to the exclusion device 11 at that timing, and this is activated.
Eliminate defective glass containers.

As shown in FIG. 2, the light source unit 1 is a combination of four straight light sources, ie, straight tube fluorescent lamps 14, and is provided right above the inspection position. The interval w between the pair of straight tube fluorescent lamps parallel to each other can be freely changed. FIG. 3 shows the results of experiments on various glass containers with respect to the interval w between the straight fluorescent lamps and the luminance ratio of the image captured by the camera. In the figure, the horizontal axis represents the value obtained by subtracting the interval w (mm) between the straight fluorescent lamps from the diameter (inner diameter) of the glass container, and the vertical axis represents the luminance ratio. As described above, it has been discovered that the brightness ratio is increased by setting the interval w between the straight fluorescent lamps to about ± 10 mm of the diameter (inner diameter) of the glass container to be inspected. Therefore, by setting the interval w between the fluorescent lamps in this range, the luminance of the defect of the image captured by the camera increases, and the defect is hardly overlooked. A blackout curtain 13 is provided around the four cameras to reduce disturbance light as much as possible.

Next, image processing for a glass container (such as a cylindrical container) having a smooth curve without a fold in the body will be described with reference to FIGS. Prior to the inspection, appropriate parameter values are set in the glass container to be inspected by initial setting. The parameters are the mask width x of the vertical mask, Vs1 (the threshold value of the edge disturbance image detection luminance), Vs2 (the glass piece image detection luminance threshold), and SL (the number of pixels (= area) of the image after labeling).
Threshold, images below this threshold are erased), SH
(Threshold of the number of pixels of the image after labeling, and erasing an image having the threshold or more).

As shown in FIG. 6 (a), an image captured by a camera has an edge portion disturbance light, other disturbance light, and a defective portion such as an attached glass piece having a higher luminance than other portions. ,It is shining now. The area inside the inspection range 15 of this image is to be inspected. This image is subjected to a first binarization process for binarizing the image at a predetermined luminance Vs1 and deleting an image below Vs1. As a result, an image of disturbance light having a luminance of Vs1 or less is removed. In this processing screen (b), a horizontal scan line 16 at a set predetermined position (usually set at the center of the screen) is horizontally scanned from both ends of the screen, and an edge disturbance which detects a disturbance light image of the edge of the glass container is detected. Light detection processing is performed. As a result, part A and B
Edge disturbance light is detected in the portion. When the edge disturbance light is detected, an edge disturbance light masking process for vertically masking a predetermined width x around the portions A and B is performed.
Thus, the edge disturbance light is masked and removed. An image from which edge disturbance light has been removed is converted to a predetermined luminance Vs2.
, A second binarization process for erasing an image having a luminance equal to or lower than Vs2 is performed. Thus, disturbance light due to various causes is eliminated, and the image shown in FIG. Further, a labeling process for numbering (R1, R2, R3) the remaining images is performed. Next, the number of pixels (area) of the labeled image
Is determined to be within a range of a predetermined threshold value SL to SH. The size of the defect is empirically SL
In most cases, the size is a predetermined size in the range from SH to SH. When the size is less than SL, small dust or the like is attached. When the size is larger than SH, some disturbance light is generated.
In (c), R1 is erased as SL or less, R3 is erased as SH or more, and R2 is recognized as a defect in the range of SL to SH.

Next, image processing for a glass container having a fold in its body (a polygonal container or the like) will be described with reference to FIGS.
It will be described based on. In the case of a glass container having a fold, the disturbance light at the fold portion is generated in addition to the disturbance light at the edge portion. Since the removal of the edge disturbance light is the same as that described with reference to FIGS. 5 and 6, for simplicity, FIGS. 7 and 8 show that there is no edge disturbance light. In this case, the inspection parameters to be initially set are, in addition to the above case, the X / Y ratio (the ratio of the height to the width of the image labeled beyond Vs2, and the portrait longer than this set value is recognized as a fold. Must be set).

In FIG. 8, in the image (d) captured by the camera, disturbance light such as disturbance light at the fold portion and a defective portion such as a piece of glass adhered have a higher luminance than other portions and shine. I have. The area inside the inspection range 15 of this image is to be inspected. The image is binarized at a predetermined luminance Vs2, and a second binarization process for deleting an image having a luminance equal to or lower than Vs2 is performed. Thus, disturbance light due to various causes is eliminated, and the image shown in FIG. Further, the remaining images are numbered (R1, R2, R3, R4,
R5) Perform labeling processing. Next, a fold portion where the X / Y ratio (aspect ratio) of the labeled image is equal to or more than a predetermined value is detected, and a fold portion mask process for masking the portion in the vertical direction is performed. In this case, R4 and R5 are recognized as folds, and are erased by being masked in the vertical direction at the widths of Y1 and Y2.
Next, with respect to the image (f) from which the fold portion has been deleted, the number of pixels (area) of the labeled image is set to a predetermined threshold SL.
欠 陥 SH is determined. The size of the defect is almost empirically determined to be a predetermined size in the range of SL to SH. When the size is smaller than SL, small dust or the like is attached. This is the disturbance light. In (c),
R1 is erased as SL or less, R3 is erased as SH or more, and R2 is recognized as a defect in the range of SL to SH.

The above apparatus was used to inspect the glass container for a certain period of time, and 62 of them were excluded. The breakdown of the defects was as follows.
%, Foam = 15%, color stripe = 15%, press mark = 6%,
Foreign matter = 2%, dust = 20%. Further, a visual inspection was performed on a glass container determined to be free from defects by this inspection, but no defects were found.

[0023]

The inspection apparatus of the present invention automatically detects a defect in the body of a glass container such as adhesion of a glass piece, and eliminates defective products with high accuracy without relying on a visual inspection by an inspector. This makes it possible to prevent the inspector from having to perform any severe work and to prevent the consumer from being injured by the attached glass piece due to the omission of the inspection. In addition to the adhesion of glass pieces, defects such as bubbles and contamination of foreign matter can be eliminated, so that the glass container also serves as an appearance inspection.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram illustrating an outline of an inspection apparatus according to an embodiment.

FIG. 2 is a top view of a light source unit.

FIG. 3 is an explanatory diagram showing a relationship between a diameter of a glass container and a luminance ratio.

FIG. 4 is a top view showing a relationship between a glass container and a camera.

FIG. 5 is a block diagram illustrating an algorithm of a process of removing disturbance light in an edge portion and a process of determining a defect in an image processing unit.

FIG. 6 is an explanatory diagram of the image processing in FIG. 5;

FIG. 7 is a block diagram illustrating an algorithm of a process of removing disturbance light from a fold and a process of determining a defect;

FIG. 8 is an explanatory diagram of the image processing in FIG. 7;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Light source part 2 Camera 3 Camera 4 Camera 5 Camera 6 Image processing part 7 Control part 8 Container spacing adjustment part 9 Conveyor speed detector 10 Container sensor 11 Elimination device 12 Conveyor 13 Dark curtain 14 Straight fluorescent lamp 15 Inspection range 16 Scan line

──────────────────────────────────────────────────続 き Continued on the front page (58) Fields investigated (Int.Cl. ⁷ , DB name) G01N 21/88 G01N 21/90 G01B 11/30 G06T 1/00 G06T 7/00

Claims (6)

(57) [Claims] 1. A light source unit for irradiating inspection light from above a glass container to be inspected, a camera provided on a side of the glass container and capable of capturing an image of a body of the glass container, and a camera. An image processing unit that processes the captured image, wherein the image processing unit binarizes the image at a predetermined luminance Vs1 that is an edge part disturbance light image detection luminance threshold, and binarizes the image to Vs1 or less. A first binarization process for erasing an image and, on a screen on which the first binarization process has been performed, a set predetermined position is horizontally scanned from an edge of the screen to detect a disturbance light image at an edge portion of the glass container. An image obtained by performing an edge disturbance light detection process, an edge disturbance light mask process for vertically masking a portion where the edge disturbance light is detected in a predetermined width, and a glass piece image obtained by performing the edge disturbance light mask process
A second binarization process for binarizing the image at a predetermined luminance Vs2 which is a detection luminance threshold value and erasing an image equal to or lower than Vs2, and an image remaining after the image subjected to the second binarization process. A defect inspection apparatus for a glass container, comprising: a labeling process for numbering; and a defect determination process for determining that the number of pixels of the labeled image is within a predetermined range of SL to SH. 2. A light source unit for irradiating inspection light from above a glass container to be inspected, a camera provided on a side of the glass container and capable of capturing an image of a body of the glass container, An image processing unit that processes the captured image, wherein the image processing unit binarizes the image at a predetermined luminance Vs1 that is an edge part disturbance light image detection luminance threshold, and binarizes the image to Vs1 or less. A first binarization process for erasing an image, and in a screen on which the first binarization process has been performed, a predetermined range set from an edge of the screen is horizontally scanned to detect a disturbance light image at an edge portion of the glass container. The image obtained by performing the edge disturbance light detection process, the edge disturbance light mask process for vertically masking the portion where the edge disturbance light is detected by a predetermined width, and the glass fragment image
A second binarization process for binarizing the image at a predetermined luminance Vs2 which is a detection luminance threshold value and erasing an image equal to or lower than Vs2, and an image remaining after the image subjected to the second binarization process. Labeling processing for numbering, fold part mask processing for detecting a fold part where the aspect ratio of the labeled image is a predetermined value or more, and masking that part in the vertical direction, and pixels of the image subjected to the fold part mask processing The number is a given value S
A defect determination process for determining that the defect is in the range of L to SH. 3. The defect of the glass container according to claim 1, wherein four cameras are installed at approximately 90 ° intervals, and these cameras are simultaneously operated to capture an image of the glass container. Inspection device 4. The inspection apparatus according to claim 1, 2 or 3, wherein the light sources are four linear light sources combined in a grid pattern, and the interval between two sets of linear light sources in parallel directions is freely changeable. Inspection apparatus for glass containers 5. The glass container according to claim 4, wherein the distance between the two sets of linear light sources in the parallel direction is equal to the diameter (inner diameter) of the glass container to be inspected plus or minus 10 mm. Defect inspection equipment 6. The inspection apparatus according to claim 1, wherein a light source unit and a camera are arranged so as to perform inspection when a glass container to be inspected is moving on a straight conveyor. It receives a signal from a container sensor on the upstream side of the conveyor that detects a glass container, generates an inspection timing, sends an inspection signal to the image processing unit, and is sent from the image processing unit when the image processing unit determines a defect. A defect inspection apparatus for a glass container, comprising a control unit that receives a defect signal, generates an exclusion timing, and operates an elimination device such as a pusher.