WO2004036199A1

WO2004036199A1 - Glass bottle mouth part inspection device

Info

Publication number: WO2004036199A1
Application number: PCT/JP2002/010840
Authority: WO
Inventors: Yasuhiko MORISAKI; Yasusaburo Kodama; Hisao Katogi; Akira Fujiwara
Original assignee: Kirin Techno-System Corporation
Priority date: 2002-10-18
Filing date: 2002-10-18
Publication date: 2004-04-29
Also published as: AU2002344113A1

Abstract

A glass bottle mouth part inspection device capable of detecting any defective formation at an end part or a threaded part of a glass bottle mouth part by picking up the image thereof. The glass bottle mouth part inspection device illuminates a glass bottle (2) and picks up the light from a mouth part (3) so as to detect any defects of the mouth part through the image processing. The device comprises an illumination (7) disposed on an upper part of a top surface of the mouth part of the glass bottle, a plurality of CCD cameras (11-20) disposed around the mouth part of the glass bottle, and an image processor (8) for processing the images obtained by the CCD cameras. The plurality of CCD cameras pick up the light transmitted through the mouth part after the light from the illumination is incident in the mouth.

Description

Glass bottle inspection equipment

TECHNICAL FIELD The present invention relates to a glass bottle inspection device for glass bottles, and in particular, detects a molding defect in a screw hole or a thread portion at a bottle mouth of a glass bottle by imaging.

The present invention relates to an apparatus for inspecting the mouth of a glass bottle.

Background art

When manufacturing glass bottles, cracks such as cracks in the thickness of the bottle

(Cracks) may occur, and these cracks are called bills. The places where the bottle mouth is twisted are limited to some extent.Typically, the bottles are formed near the top of the bottle, the screw is formed at the thread of the bottle, and the neck is formed at the neck of the bottle. There is a neck neck to do. Depending on the direction of the cracks, there are two types of vertical slimes that are drowning in the vertical direction (substantially vertical direction), horizontal slashes that extend in the horizontal direction (substantially horizontal direction), and diagonal slashes that extend diagonally.

Since the above-mentioned squeeze causes damage to the glass bottle, the presence or absence of spiking is detected by imaging the bottle part and the glass bottle with spiking is eliminated as a defective bottle.

2. Description of the Related Art Conventionally, there has been known a glass bottle twist detecting apparatus which automatically detects whether or not there is a blur by imaging a bottle portion of the glass bottle. This squeeze inspection device is composed of a plurality of pairs of transmitters and receivers arranged to surround the periphery of the bottle of a glass bottle, and the plurality of pairs of transmitters and receivers are to be inspected. The glass bottles are adjusted to the optimal position with respect to the bottle opening, and are in a positional relationship. Then, each pair of sender and receiver forms a glass bottle. The reflected light is received and the signal obtained is processed to detect the undulation of the bottle. In this case, the light from the projector enters the bottle, and if there is a bill, it is reflected by the cracked surface of the bill and shines brightly. Sharpness of the bottle part is detected based on whether or not there is a part having brightness equal to or higher than a predetermined value.

The above-mentioned conventional billet inspection apparatus is provided with a plurality of inspection stations for inspecting the beveling of the bottle portion of the glass bottle, and the star wheel for inspection holds the glass bottle and holds the circumference. The glass bottle is conveyed above and indexed (rotated) by multiple inspection stations. At a plurality of inspection stations, the glass bottle is rotated on its own, and each is inspected exclusively for defects such as opening, screwing, and necking.

The above-mentioned conventional squeeze inspection apparatus is a target to be inspected because it has a configuration in which a plurality of inspection stations are provided and a plurality of pairs of light emitters and receivers are arranged for each inspection station. When the type of glass bottle is changed, there is a problem in that the positional relationship between multiple pairs of transmitters and receivers must be readjusted for each inspection station. In other words, there is a problem that the angles and heights of a plurality of pairs of emitters and receivers at each inspection station must be readjusted, and the sensitivity of the receivers must be readjusted.

Further, in a glass bottle having a threaded portion in the bottle portion, the threaded portion has a complicated curved surface, so that there is often the same reflected light as that of the thread in the threaded portion. The processing is performed so that even if there is reflected light from the area where the thread part exists, it is not determined to be pilling. Therefore, even if there is a run in the upper and lower regions of the thread part and the thread part, it is not detected. Furthermore, since the seam of the bottle also has a curved surface that is continuous in the vertical direction, there are many cases where there is reflected light similar to that of Pili. Since the same processing is performed, this bottle Is not detected even if there is a sway in the seam portion and the surrounding area. In addition, it has been difficult to detect defective molding at the thread part and the like that occurs when molding bottles. Disclosure of the invention

The present invention has been made in view of the above circumstances, and does not require readjustment of the arrangement of the light emitter and the light receiver at the time of a type change in which the type of the glass bottle to be inspected is changed. By detecting the bill around the seam of the billiard bottle in the area of the target area, it is possible to accurately detect pills and to detect molding defects in the bottle (especially thread part). The objective is to provide a glass bottle inspection device. .

In order to achieve the above object, one embodiment of the present invention relates to an inspection apparatus that illuminates a glass bottle, captures light from the bottle, and detects a defect in the bottle by image processing. Lighting provided above the top surface of the bottle, a plurality of CCD cameras arranged around the bottle of the glass bottle, and an image processing device for processing images obtained by the CCD camera. The plurality of CCD cameras capture light transmitted from the bottle portion after light from the illumination enters the bottle.

According to the present invention, the light from the illumination placed above the glass bottle enters the glass bottle bottle, and some light is transmitted through the bottle. A plurality of CCD cameras arranged around the bottle mouth of the glass bottle capture the light transmitted through the bottle portion of the glass bottle. In this case, if there is a pill inside the bottle, the light that enters the bottle from inside the bottle is reflected by the cracked surface of the pill, and this reflected light passes through the bottle and passes through the CCD camera. It is photographed by. The light reflected on the crack surface of Piri is brighter than the light transmitted through other parts, so the part corresponding to Piri in the image taken with the CCD camera Is a lighter area than other parts. The image processing device determines that this bright area is vibrate.

On the other hand, if there is a molding defect in the thread part or the like, the light from this defective molding part becomes a dark and fuzzy image compared to the image of the thread part formed normally, so this molding defect Defects can be detected.

As described above, according to the present invention, the interior of the bottle is illuminated from the illumination disposed above the bottle of the glass bottle, and the light transmitted from the bottle is transmitted to the periphery of the bottle after entering the bottle. By photographing with a plurality of CCD cameras arranged at different elevation angles, it is possible to detect shaping in the bottle part and molding defects in the thread part. Therefore, when the type of glass bottle to be inspected is changed, the positional relationship between the illumination, which is the light emitter, and the CCD camera, which is the light receiver, does not need to be readjusted. It can be dramatically shortened.

In a preferred aspect of the present invention, the optical axes of the plurality of CCD cameras are on a line extending radially from the bottle portion of the glass bottle. In a preferred aspect of the present invention, the plurality of CCD cameras are installed in a hemisphere centered on a bottle portion of a glass bottle. In this case, it is preferable that the hemisphere is configured to be vertically movable. With such a configuration, it is possible to cope with a type change in which the type of glass bottle is changed only by changing the vertical position of the hemisphere, and the adjustment time at the time of the type change can be drastically reduced.

Another aspect of the present invention is an apparatus arranged at a different inspection position from the glass bottle inspection apparatus according to any one of claims 1 to 4, wherein A first lighting device arranged on the side of the mouth, a plurality of CCD cameras arranged around the bottle portion of the glass bottle, and an image processing device for processing an image obtained by the CCD camera; The plurality of CCD cameras capture light reflected from or transmitted through the bottle portion after light from the illumination enters the bottle portion. According to the present invention, the light from the illumination installed on the side of the glass bottle enters the bottle portion of the glass bottle. A plurality of CCD cameras arranged around the bottle of the glass bottle capture the light reflected or transmitted from the bottle of the glass bottle. In this case, if there is a pill inside the bottle, the light incident on the inside of the bottle from the outer surface of the bottle will be reflected by the cracked surface of the screw, and this reflected light will pass through the bottle. Photographed by CCD camera. The light reflected from the crack surface of the rivet is brighter than the light transmitted through other parts, and therefore, in the image taken by the CCD camera, the part corresponding to pilus is a brighter area than the other parts. The image processing device determines that this bright area is vibrate.

On the other hand, if there is a molding defect in the thread part or the like, the reflected light from this defective molding part becomes a slightly faint image compared to the image of the thread part formed normally. Defective molding defects can be detected.

As described above, according to the present invention, the bottle part is illuminated from the illumination arranged on the side of the bottle part of the glass bottle, and the light reflected or transmitted from the bottle part is placed around the bottle part. Images are taken with a plurality of CCD cameras at different elevation angles to detect the undulation of the bottle. Therefore, even if the crack surface of the vertical pill perfectly matches the direction extending in the radial direction from the axis of the bottle, such a sway can be detected.

One preferred embodiment of the present invention is directed to an angle detecting means for detecting a rotation angle of the glass bottle with respect to a reference position, and a second illumination for the angle detecting means arranged on the side of the bottle opening of the glass bottle. It is further characterized by the following. In this case, it is preferable that the first illumination and the second illumination emit light that does not interfere with each other. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a longitudinal sectional view showing a mouth inspection apparatus according to a first embodiment of the present invention. FIG.

FIG. 2 is a plan view of a hemisphere of the mouth inspection device shown in FIG.

FIG. 3 is a schematic diagram showing the behavior of light from illumination in the first embodiment of the present invention.

FIG. 4 is a schematic diagram showing the relationship between the image processing device of the mouth inspection device and the CCD camera according to the first embodiment of the present invention.

FIG. 5 is a schematic diagram illustrating an example of an image in which angle information and a mold number are written.

FIG. 6 is a schematic diagram showing an image of a glass bottle serving as a sample. FIG. 7 is a frequency distribution showing the distribution of pixel brightness.

FIG. 8 is a schematic diagram showing an image of a non-defective glass bottle.

FIG. 9 is a graph showing the distribution of brightness of pixels in a specific row.

FIG. 10 is a diagram showing the relationship between the brightness distribution of each pixel in the image of the glass bottle to be inspected and the template.

FIG. 11A is a schematic diagram showing a bright template, and FIG. 11B is a schematic diagram showing a dark template.

Fig. 12A shows the bright template image imaged based on each value of the bright template shown in Fig. 11A, and Fig. 12B shows the bright template image based on each value of the dark template shown in Fig. 11B. This shows a template image that has been imaged.

FIG. 13 is a plan view showing a main part of the mouth inspection apparatus according to the second embodiment of the present invention.

FIG. 14 is a sectional view taken along line AA of FIG.

FIG. 15 is a sectional view taken along line BB of FIG.

FIG. 16 is a schematic diagram illustrating the behavior of light from illumination according to the second embodiment of the present invention.

FIG. 17 shows the image processing of the mouth inspection apparatus according to the second embodiment of the present invention. FIG. 2 is a schematic diagram showing a relationship between the processing device and a CCD camera. BEST MODE FOR CARRYING OUT THE INVENTION

Hereinafter, an embodiment of a glass bottle inspection apparatus according to the present invention will be described with reference to FIGS. 1 to 17.

The glass bottle to be inspected is held on an inspection starwheel (not shown) and transported along a transport path on the circumference of the starwheel. The glass bottle inspection apparatus according to the present invention is disposed at one station (first inspection station) on the transfer path on the circumference of the star wheel. In this first inspection station, the glass bottle conveyed by the star wheel is indexed (rotated), and the glass bottle roto-inspection device according to the present invention is used to check the screw in the bottle roto or the screw crest. Molding failure is detected.

FIG. 1 is a longitudinal sectional view showing the mouth inspection device according to the first embodiment of the present invention. As shown in FIG. 1, the mouth inspection device supports a hemisphere 4 arranged to cover the bottle holder 3 of a glass bottle 2 placed on a rotatable turntable 1 and a hemisphere 4 It has struts 5. The center O of the hemisphere 4 substantially coincides with the bottle portion 3 of the glass bottle 2 placed on the turntable 1. The hemisphere 4 is attached to a column 5 via a vertically movable sliding member 6, and is configured to be vertically movable with respect to the column 5.

FIG. 2 is a plan view of the hemisphere 4 shown in FIG. As shown in FIGS. 1 and 2, the top of the hemisphere 4, that is, above the top surface of the bottle part 3 of the glass bottle 2 on the turntable 1, is inside the bottle of the glass bottle 2 on the turntable 1. Lighting 7 for irradiating light is installed. A plurality of CCD cameras 10 to 20 are arranged on the hemisphere 4 so as to surround the bottle opening 3 of the glass bottle 2. The optical axis of these CCD cameras 10 to 20 is the center of the hemisphere 4 It is on the line extending radially from O (bottle part 3 of glass bottle 2).

In the present embodiment, a total of 11 CCD cameras are arranged. Of these, one camera 10 is provided with a screw of a bottle section 3 of a glass bottle 2 placed on a turntable 1. It is a camera dedicated to angle detection that detects the rotation angle of the glass bottle 2 with respect to a predetermined reference position by photographing the mountain. As shown in Fig. 1, the angle detection camera 10 is arranged so that the elevation angle of the optical axis is 0 °, and photographs the thread of the bottle part 3 of the glass bottle 2 from the horizontal direction. It has become.

The cameras 11 to 20 other than the camera 10 dedicated to angle detection are inspection CCD cameras for photographing the bottle 3 from various angles and inspecting the bottle 3 for stiffness. In this embodiment, the angle between the projection of the optical axis of each camera on the horizontal plane and the optical axis of the camera 10 dedicated to angle detection is 25 ° (the first detection CCD camera 11 and the second detection CCD camera 11). Operation CCD camera 1 2), 59.5 ° (3rd inspection CCD camera 13), 140 ° (4th inspection CCD camera 14, 5th inspection CCD camera 15), 18 5 ° (6th inspection CCD camera 16), 220 ° (7th inspection CCD camera 17), 260 ° (8th inspection CCD camera 18), 296.5 ° Inspection CCD cameras 11 to 20 are arranged at such positions as to be (ninth inspection CCD camera 19) and 32 ° (10th inspection CCD camera 20).

The elevation angle of the optical axis of the first inspection CCD camera 11 is 30 °, the second inspection CCD camera 12 is 0 °, the fourth inspection CCD camera 14 is 55 °, and the fifth inspection. 15 ° for CCD camera 15, 45 ° for CCD camera 16 for sixth inspection, 20 ° for CCD camera 17 for seventh inspection, 35 ° for CCD camera 18 for eighth inspection The 10th inspection CCD camera 20 is 25. It has become. The third inspection CCD camera 13 and the ninth inspection CCD camera 19 are configured to be able to move up and down on the surface of the hemisphere 4, and the elevation angle of the optical axis can be freely set. You can do it. The number of pixels of each of the CCD cameras 10 to 20 used in the present embodiment is 64 × 64, so that one image can be captured every 0.4 millisecond. . For example, when inspecting 300 glass bottles per minute, the processing time for one glass bottle is 200 milliseconds. If only glass bottles are to be photographed, a maximum of 250 images (2100 / 0.4) can be photographed per glass bottle. During the inspection of the bottle part 3 of the glass bottle 2 by the mouth inspection device, the turntable 1 is rotating, and each CCD camera 10 is rotated while the glass bottle 2 is rotated. The glass bottle 2 is photographed at the same time according to. In this way, by repeatedly taking a picture of the glass bottle 2 while rotating the glass bottle 2, it is possible to take a picture of the bottle portion 3 of the glass bottle 2 over the entire circumference.

Here, light from the illumination 7 disposed above the glass field 2 is incident on the bottle mouth of the glass bottle 2, as shown in FIG. 3, a part of the light L _A inner peripheral surface of the Binro 3 And enters the bottle part 3. If there is a bill C inside the bottle part 3, this light L _A is reflected by the cracked surface of the bill C inside the bottle part 3, and this reflected light L _B passes through the inside of the bottle part 3. Then, the image is captured by the above-described inspection CCD cameras 11 to 20. Light L _B reflected by the crack plane of the pyridinium C is brighter than light which has passed through other parts, therefore, in the image taken by the CCD camera, the portion corresponding to the kink C is bright than the other partial regions Become. The image processing device provided in the mouth detection device detects this bright region from the images obtained by the above-described CCD cameras 10 to 20 and determines that the region is stiff. On the other hand, if there is no internal kink C of Binro portion 3, the part of the light L _A is transmitted through the Binro portion 3 as it enters the Binro portion 3 from the inner circumferential surface of the Binro 3 . In this case, if there is a molding defect in the thread part, etc., the light from this defective molding part will be scattered in a direction that does not enter the corresponding CCD camera, and the image of the thread part that has been formed normally Because the image is slightly darker and fuzzier than Good defects can be detected.

FIG. 4 is a schematic diagram showing a relationship between the image processing apparatus and each of the CCD cameras 10 to 20. As shown in FIG. 4, the image processing apparatus 7 includes operation boards 30 to 40 corresponding to the respective CCD cameras 10 to 20. These operation ports 30 to 40 correspond. They are connected to CCD cameras 10 to 20, respectively.

The angle detection arithmetic board 30 to which the angle detection dedicated camera 10 is connected has a relationship between the height of the screw helix of the glass bottle 2 and the rotation angle of the glass bottle 2 with respect to a predetermined reference position in advance. It is remembered. The calculation port 30 for angle detection detects the height position of the spiral of the screw thread from the image obtained by the camera 10 dedicated to angle detection, and determines the reference position based on the above relationship from the height position of the spiral screw thread. The rotation angle of the glass bottle 2 at the time of shooting with respect to is detected. The detected rotation angle signal of the glass bottle 2 is sent to the operation boards 31 to 40 connected to the respective inspection CCD cameras 11 to 20. In this way, the angle detection dedicated camera 10 and the angle detection calculation board 30 constitute angle detection means for detecting the rotation angle of the glass bottle at the time of shooting with respect to the reference position.

As described above, the rotation angles of the glass bottle 2 sent from the angle detection calculation board 30 are sent to the calculation boards 31 to 40 connected to the inspection CCD cameras 11 to 20, respectively. This rotation angle is written as rotation angle information in each image captured by each of the inspection CCD cameras 11 to 20.

Next, the operation of the glass bottle inspection apparatus configured as described above will be described with reference to FIGS.

As described above, the diffused light from the illumination 7 enters the bottle from above the bottle 3 of the glass bottle 2 placed on the turntable 1. The diffused light that has entered the bottle is diffused radially and passes through the bottle 3. Then, the transmitted light that has passed through the bottle 3 in a radial manner is equivalent to all the C arranged around the bottle 3 Photographed simultaneously by a CD camera (11 CCD cameras) 10 to 20. At this time, as described above, one CCD camera is a camera dedicated to angle detection, and the camera 10 dedicated to angle detection captures the thread of the bottle holder 3 to set the reference position. The angle of rotation of the glass bottle at the time of shooting can be detected. Since the height of the spiral of the thread changes by one pitch per rotation, the relationship between the height of the spiral of the thread and the rotation angle with respect to the reference position is calculated in advance by the dedicated angle detection camera 10. If stored in the board 30, the camera 10 dedicated to angle detection can detect the angle with respect to the reference position at the time of photographing. As the reference position, for example, the starting end, which is the start of the thread, may be set as the reference position (0 °).

With reference to the angle detection dedicated camera 10, the relative position of each detection CCD camera 11 to 20 with respect to the angle detection dedicated force camera 10 is predetermined. The rotation angle detected by the calculation port 30 of the camera 10 can be considered by shifting the reference position relatively. Each of the detection CCD cameras 11 to 20 images the bottle 3 It can also be used as the rotation angle. For this reason, in the present embodiment, each image captured by each of the detection CCD cameras 11 to 20 is transmitted from the angle detection calculation board 30 of the angle detection dedicated camera 10. The rotation angle is written.

While rotating the glass bottle 2 by the turntable 1, the transmitted light transmitted through the bottle 3 is photographed at predetermined time intervals to obtain a large number of images. Then, the angle information at the time of shooting described above is written in all the images.

On the other hand, glass bottle molding machines are equipped with a large number of molds, and many molds are simultaneously molded using these molds. It is known that the properties (thickness, subtle shapes, etc.) of molded glass bottles greatly depend on the mold. In addition, the generation of spiking at the mouth of the glass bottle also depends on the mold. Therefore, some of the images obtained with the inspection device of the present invention contain glass bottles. The information of the mold number which indicates whether the molding has been performed with the mold is also written. The mold number can be detected by a mold number reading device that reads a convex code formed on the bottom of the glass bottle, and the signal from the mold number reading device is a CCD camera for inspection. 20 is input to the operation board 31 to 40, and the mold number is written in each _image.c Also, information related to manufacturing such as the serial number is written in each image. It's swelling. FIG. 5 is a schematic diagram illustrating an example of an image in which the rotation angle information and the mold number obtained as described above are written. The inspection result, for example, the quality of the glass bottle may be written in each image.

Next, each image in which the angle information and the mold number etc. are written is compared with a reference image called a template prepared in advance before inspecting the glass bottle. Check if there is any stiffness. In this case. A reference image (template) is prepared for each angle and for each mold number, and the angle information and the reference image corresponding to the mold number written in the image obtained by each inspection CCD camera are stored. The selected reference image is compared with the image of the glass bottle to be inspected.

Next, a method of creating a reference image (template) will be described. The process of creating a reference image (hereinafter referred to as a template as appropriate) is roughly divided into three processes. That is, a photographing process of photographing a plurality of glass bottles used for template creation with each CCD camera, and excluding images of defective glass bottles from a group of images photographed in the photographing process to obtain non-defective glass bottles. This is an image selection process for selecting images of glass bottles, and an image creation process for creating a template based on the images selected in the image selection process. Hereinafter, each step will be described in order.

(1) Shooting process

The basic data of the imaging process performed in this embodiment is as follows.

① Number of glass bottles used for sample 100 bottles ②Mold number M 1 ~ M 8

③ Angle A1 to A8 (A1: 0 to 45 °, A2: 45 to 90 °, · · ·, A8: 315 to 360 °)

枚数 Number of shots per glass bottle 100

The 100 glass bottles to be sampled are conveyed to the inspection station by the star wheel for inspection, and the first to tenth inspection CCD cameras 11 to 20 provided in the inspection station Photographed at The captured images are sent to the computer 42 (see FIG. 4) connected to the calculation units 30 to 40 of the inspection CCD cameras 11 to 20, and are based on these images. The following steps are performed by the computer 42.

Hereinafter, an example will be described in which a template corresponding to the first inspection CCD camera 11, the mold number M1, and the angle A1 is created.

As described above, 100 images are taken for one glass bottle, so a total of 100,000 images for 100 glass bottles are used for the first inspection. Photographed by CCD camera 11. If three of the 100 glass bottles are molded with mold number M1, the number of images of the glass bottle molded with mold number M1 is 3 x 10 0 sheets = 300 sheets. Therefore, first, 300 images of the glass bottle formed by the mold number M 1 are extracted from the 100,000 images ₍ further, from among the 300 images, In this embodiment, 35 images are photographed at the angle A1. Therefore, the mold number Ml and the angle A1 are used. To create a corresponding template, 35 images are selected. (2) Image selection process

When creating a template that serves as a reference image, if multiple glass bottles that are used as samples contain defective products, a template that includes light based on stiffness is created. Will be. in this way, If a template is created based on an image that contains bright light from a part that should not shine, it is impossible to judge a glass bottle that has stiffness in that part as a defective product. For this reason, as a pre-process for creating a template, work is performed to exclude images of defective glass bottles from multiple images used in the template.

In the image selection step, a frequency distribution indicating the brightness distribution of each pixel constituting the image is created based on the plurality of images selected in the imaging step. FIG. 6 is a schematic diagram showing an image of a glass bottle serving as a sample. FIG. 7 is a frequency distribution showing the distribution of pixel brightness. In FIG. 6, reference numeral 50 indicates a bright part. In FIG. 7, the vertical axis of the frequency distribution represents the number of pixels, and the horizontal axis represents brightness (0 to 255).

As shown in FIG. 6, the image of each of the CCD cameras 11 to 20 is composed of a pixel group of 64 × 64 pixels. The number of pixels can be appropriately adjusted. In this example, one image can be decomposed into 64 × 64 pixels. Then, the pixels in the first row and first column from the decomposed pixel group are plotted on the graph for each image. In this way, when the pixels in the first row and the first column are sequentially plotted on the graph for 35 images, the frequency representing the brightness distribution of the first row and the first column as shown in FIG. 7 is obtained. Distribution can be obtained. This frequency distribution is created from 1st row, 1st column to 64th row, 6th column.

Next, a standard deviation σ indicating the degree of variation in brightness is calculated for each obtained frequency distribution. This standard deviation σ is obtained by a general statistical method. Then, for example, when the brightness of the pixels is distributed within the range of ± 2σ, the detection criterion is set so that the image is judged to be a good glass bottle image. If all sample glass bottles are good, the brightness of all pixels will be distributed around the average value X. Therefore, as shown in Fig. 7, all pixels are within the range of ± 2σ ₍ In this case, no images are rejected and all 35 images are used for template creation.

On the other hand, if there is a pill in a glass bottle, the part 60 of the image that indicates the presence of the stiffness becomes extremely bright (see Fig. 6). Then, in the frequency distribution, as shown in FIG. 7, the number 61 of pixels having the brightness corresponding to this part 60 is plotted in the area to the right of + 2σ. Then, it is determined that an image having such pixels has a sharp image. Similarly, when there is an extremely dark part, the number of pixels having the brightness corresponding to this part is plotted in the area on the left side of 12σ. Then, it is determined that an image having such pixels includes a thread portion or the like having a molding defect. These images are then excluded from the images used to create the template. In the present embodiment, the image selection step is performed by using a statistical method. However, the present invention is not limited to this. Other methods may be used as long as the image to be excluded can be specified. For example, multiple images obtained by the shooting process are displayed on the display for each mold and each angle, and the operator selects images of defective glass bottles by looking at the images on the display. Is also good.

(3) Image creation process

In the image creation step, a reference image serving as a template is created based on the plurality of images selected in the above-described image selection step. This image creation step will be described with reference to FIGS. Figure 8 is a schematic diagram showing an image of a good glass bottle. FIG. 9 is a graph showing a distribution of brightness of pixels in a specific row. FIG. 10 is a diagram showing the relationship between the brightness distribution of each pixel of the image of the glass bottle to be inspected and the template.

The template in the present embodiment is created for each pixel row of the image. First, a certain pixel row is specified. For example, in FIG. 8, assume that the third line is specified. Next, the identified pixel row is placed in the column direction (horizontal direction in Fig. 8). (1, 2, 3 · · · 64 directions), and the brightness of each pixel is shown in a graph. Specifically, each pixel on the specified pixel row is plotted on a graph, with the brightness level of the pixel on the vertical axis and the column number of the pixel on the horizontal axis. As shown in FIG. 8, since the pixels in the third row are located in the dark area 70, the line drawn when each pixel in the third row is plotted on the graph has a brightness near zero. Is a straight line located at

On the other hand, for example, when the pixels in the 10th row are specified, the line drawn when each pixel in this row is plotted is as shown in FIG. 9. That is, in the 10th row, the bright portion 50 formed by the light from the thread portion runs, so that the pixel corresponding to the bright portion 50 has a high degree of brightness.

Furthermore, for all images, each pixel in the same row is plotted on the same graph. That is, in the present embodiment, since 35 images are used for creating the template, as shown in FIG. 9, 35 lines (in FIG. 9, only four lines Τ 1 to Τ 4 are used). A brightness distribution map consisting of is created. In this way, a brightness distribution map is created for all rows from 1 to 64.

The maximum area defined by these line groups is the area to be used as a template. That is, as shown in Fig. 10, the line obtained by connecting the points indicating the maximum value (maximum brightness) for each column is defined as a bright template line Τ max, and the minimum value (minimum brightness) for each column is obtained. Let the line obtained by connecting the points indicating）) be the template line T min. The area surrounded by the light template line T max and the 喑 template line T min becomes the reference image (template) to be obtained. That is, a range between the maximum brightness and the minimum brightness is continuously formed in the column direction between the bright template line Tmax and the dark template line Tmin. In this way, 64 templates are created for the mold number Ml and the angle A1. Since such template creation work is performed for mold numbers M1 to M20 and angles A1 to A8, 64 X 20 X 8 = 10 and 24 templates are obtained. Provided for the first inspection CCD camera 11. That is, each of the inspection CCD cameras 11 to 20 has a template for each mold, angle, and pixel row. As described above, the degree of occurrence of waviness and the like of the glass bottle greatly depends on the mold. Can be enhanced.

Next, the reference image (template) obtained by the above-described method is compared with the image obtained from the glass bottle to be inspected, and it is determined whether or not there is a molding defect in a pill or a thread portion in the bottle portion. The determination method will be described.

First, a template created under the same conditions (mold, angle, etc.) as the image to be inspected based on various information such as the angle information and mold number given to the image is used as a comparison target. Selected. Next, the image of the glass bottle to be inspected is compared with the template for each pixel row. Specifically, a line indicating the brightness distribution in the column direction in a specific pixel row is compared with the template. Then, as shown in FIG. 10, a line S1 indicating the degree of brightness of the glass bottle to be inspected is surrounded by a non-defective region (a bright template line Tmax and a template line Tmin) of the template. The glass bottle is judged to be good if it is entirely within the area.

On the other hand, if a part of the line protrudes from the non-defective area of the template, as shown by reference numeral S2, the glass bottle is determined to be defective. If all the pixel rows are determined to be defective in at least one of the rows when compared with the template, the glass bottle has a twist or a thread in the bottle section. Etc., it is determined that a molding defect exists.

Such an inspection is performed at each angle A1 to A8, for example. Even when it is determined that there is no twist or molding failure at the angle A1, it may be determined that there is a twist or molding failure at the angle A2. In the present embodiment, since the detection of the twist or the molding defect is performed at a plurality of angles (A1 to A8), the inspection accuracy of the thread or the molding defect can be improved as compared with the conventional inspection device.

The 64 templates created as described above can be further aggregated into two two-dimensional (planar) bright templates and 喑 templates. FIG. 11A is a schematic diagram showing a bright template, and FIG. 11B is a schematic diagram showing a dark template. Hereinafter, the process of creating a bright template will be described with reference to FIGS. 10 and 11A.

As shown in FIG. 10, the brightness level of the bright template line T max of a specific row can be quantified within the range of 0 to 255 for each column. Then, each numerical value indicating the degree of brightness is plotted on the corresponding row of the table composed of 64 rows × 64 columns shown in FIG. 11A. For example, a numerical value indicating the degree of brightness of the m columns of the n rows of bright template lines T max is plotted in a section located at 11 rows and m columns of the table. Here, in FIG. 11A, the degree of brightness of 0 to 255 is represented by a hexadecimal notation.

When all the brightness levels of the light template lines T max from the 1st line to the 64th line are plotted in the table, one bright template is finally created as shown in Fig. 11A. . The dark template shown in FIG. 11B is also created based on the dark template line T min by the same process as the light template. The range of the maximum brightness and the minimum brightness of each pixel constituting the reference image is determined by the light template and the dark template obtained in this way. That is, in the example shown in FIG. 11A and FIG. 11B, the range of the minimum brightness and the maximum brightness for the pixels located in n rows and m columns is a range from 02 to 08. Figure 12A is an image based on the numerical values of the light template shown in Figure 11A. FIG. 12B shows a dark template image formed based on each numerical value of the dark template shown in FIG. 11B.

In the above description, an example has been described in which the brightness of each pixel is quantified for each row when two two-dimensional (planar) light templates and templates are created. It is also possible to create two-dimensional two-dimensional light and dark templates by quantifying the degree of brightness of each pixel instead of every pixel.

Next, using the two bright templates and the dark templates obtained in the above process, it is determined whether or not the glass bottle to be inspected has a molding defect at the bottle mouth or a screw thread or the like. The method is explained. First, the particular row of the image of the glass bottle to be Ken查object is scanned in the column direction (lateral direction), from one row of _c Next, the row brightness of each pixel is quantified that line has 6 It is determined whether the brightness of each pixel up to 4 columns is within the range of the maximum brightness and the minimum brightness determined by the light template and the dark template (hereinafter referred to as non-defective range). . This process is performed for all rows from 1 to 64. If the brightness of all the pixels constituting the image to be inspected is within the acceptable range, it is determined that the glass bottle does not have a molding defect in the bottle part, a thread part, or the like. . On the other hand, if there is more than the specified number of pixels whose brightness deviates from the acceptable range by more than the allowable value, the image is judged to be an image of a defective glass bottle, and the glass bottle has a slip in the bottle opening or It is determined that there is a molding failure in the thread portion or the like. Note that the allowable value from the non-defective range and the specified value of the number of pixels, which are the criteria for determining a defective product, can be set according to the inspection accuracy to be achieved. For example, if a predetermined number of adjacent pixels in an image have brightness outside the non-defective range, the image may be determined to be an image of a defective glass bottle.

In the embodiment shown in FIGS. 1 to 12A and 12B, The inside of the bottle is illuminated, and the vignetting or molding defect is detected from the transmitted light that has passed through the bottle. According to this configuration, the lateral vibration extending in the horizontal direction and the oblique vibration extending obliquely can be completely detected. In addition, most of the vertical runout extending in the vertical direction can be detected.However, if the crack surface of the vertical runout completely matches the direction extending in the radial direction from the axis of the bottle, the bottle must be removed. Since transmitted light travels in parallel to the crack plane, there is a possibility that longitudinal vibration cannot be detected. For this reason, in the second embodiment, a second inspection station is provided in the middle of the transport path on the circumference of the star wheel, and the second inspection station detects a vertical wobble by reflected light at the second inspection station. (B) The inspection device is located. In the second embodiment, the inspection apparatus using the transmitted light shown in FIGS. 1 to 12A and 12B is of course installed in the first inspection station.

Next, a description will be given of a glass bottle inspection apparatus for detecting vertical warpage with reference to FIGS. 13 to 16. FIG.

Fig. 13 is a plan view showing the main part of the mouth detection device for detecting vertical runout, Fig. 14 is a cross-sectional view taken along line A-A in Fig. 13, and Fig. 15 is a cross-sectional view taken along line B-B in Fig. 13. It is a figure. As shown in FIGS. 13 to 15, this mouth detection device includes a hemisphere 104 arranged so as to cover the bottle portion 3 of the glass bottle 2. The center 0 of the hemisphere 104 substantially coincides with the bottle part 3 of the glass bottle 2. On the side of the hemisphere 104, that is, on the side of the bottle 'part 3 of the glass bottle 2, a first illumination 107a for irradiating light to the bottle part 3 of the glass bottle 2 is installed. I have. A plurality of CCD cameras 110 to 119 are arranged on the hemisphere 104 so as to surround the bottle part 3 of the glass bottle 2. The optical axis of these CCD cameras 110 to 119 is on a line extending radially from the center O of the hemisphere 104 (the bottle part 3 of the glass bottle 2).

In the present embodiment, a total of 10 CCD cameras are arranged, and one of these cameras 110 is provided for the bottle section 3 of the glass bottle 2. It is a camera dedicated to angle detection that detects the screw thread and detects the rotation angle of the glass bottle 2. As shown in Fig. 15, the dedicated angle detection camera 110 is positioned so that the elevation angle of the optical axis is 0 °, and photographs the thread of the bottle part 3 of the glass bottle 2 from the horizontal direction. To do so. On the side of the hemisphere 104 facing the camera 110 dedicated to angle detection, a second light 107 b is arranged, and the second light 107 b causes the bottle 2 It illuminates the threads of part 3. The light emitted from the second illumination 107b is infrared light, and does not interfere with the light emitted from the first illumination 107a. Further, the camera 110 dedicated to angle detection receives only infrared light emitted from the second illumination 107 b.

The cameras 111 to 119 other than the angle detection camera 110 are inspection CCD cameras for photographing the bottle 3 from various angles and inspecting the bottle 3 for stiffness. In the present embodiment, the angle between the projection of the optical axis of each camera on the horizontal plane and the optical axis of the camera 110 dedicated to angle detection is 90 ° (the first detection CCD camera 111). ), 130 ° (2nd inspection CCD camera 1 1 2), 150 ° (3rd inspection CCD camera 1 1 3), 180 ° (4th inspection CCD camera 1 1 4) ), 220 ° (5th inspection CCD camera 115, 6th inspection CCD camera 1 16), 260 ° (7th inspection CCD camera 1 17), 300 ° (Eighth inspection CCD camera 118), 311 ° (Ninth inspection CCD camera 111) CCD inspection cameras 111 to 119 are arranged respectively. It has been done.

The elevation angle of the optical axis of the first inspection CCD camera 111 is 40 °, the second inspection CCD camera 112 is 35 °, the third inspection CCD camera 113 is 0 °, 4 Inspection CCD camera 1 14 is 50 °, 5th inspection CCD camera 1 15 is 40 °, 6th inspection CCD camera 1 16 is 10 °, 7th inspection CCD camera 1 17 is 35 °, 8th inspection CCD camera 1 1 8 is 35 °, and the ninth inspection CCD camera 119 is 0 °.

Here, as shown in FIG. 1 6, the light L _c from the first illumination 1 0 7 a, and enters from the outer circumferential surface of the Binro portion 3 of the glass bottle 2 to Binro unit 3. Inside last place of Binro portion 3 (vertical Billiton) If a C is, of the light L _c is kink C inside the Binro portion 3 reflected by the crack plane, the reflected light L _D is the Binro 3 The image is transmitted through the inside and photographed by the above-described detection CCD cameras 11 1 to 11 19. The light L _D reflected from the crack surface of the billiary C is brighter than the light transmitted through other parts, and therefore, in the image taken by the CCD camera 11 1 to 1 19, the part corresponding to the billiary C Is a lighter area than other parts. The image processing device provided in the mouth detection device detects this bright region from the images obtained by the above-described CCD cameras 11 1 to 11 9 and determines that the region is vibrant. On the other hand, when there is no kink C inside the Binro unit 3, the light L _c from the first illumination 1 0 7 a it enters the Binro portion 3 from the outer peripheral surface of Binro 3 Binro Either passes through part 3 or reflects on the outer surface of bottle part 3. In this case, if there is a molding defect in the thread part, etc., the light from this defective part will be scattered in a direction not incident on the corresponding CCD camera, and the image of the thread part formed normally Since the image is slightly darker and more blurry than this, it is possible to detect this defective molding defect. The configuration of the image processing device is the same as the image processing device of the mouth inspection device in the first embodiment described above, and thus the description is omitted here. Next, the operation of the glass bottle inspection apparatus configured as described above will be described with reference to FIGS.

Infrared light from the second illumination 107 b enters the bottle 3 from the side of the bottle 3 of the glass bottle 2 placed on the turntable 1 and passes through the bottle 3 . The infrared light transmitted through the bottle part 3 is photographed by the angle detection camera 110 provided opposite to the second illumination 107 b. As in the first embodiment described above, the camera 110 dedicated to angle detection takes an image of the thread of the bottle part 3 so that the glass bottle at the time of photography with respect to the reference position at the time of photography is taken. W 200

23 The rotation angle can be detected.

On the other hand, the diffused light from the first illumination 107 a enters the bottle part 3 of the glass bottle 2 placed on the turntable 1. The inspection CCD cameras 1 1 1 to 1 1 9 capture the light reflected from the bottle 3 of the glass bottle 2. In this case, if there is a pill inside the bottle part 3, the light incident on the inside of the bottle part from the outer peripheral surface of the bottle part 3 is reflected by the cracked surface of the bottle, and this reflected light passes through the bottle part 3. The image is transmitted through the CCD camera 111 to 119.

Based on the camera 110 dedicated to angle detection, since the relative positions of the CCD cameras 11 1 to 11 for detection with respect to the camera 110 dedicated to angle detection are predetermined, the camera dedicated to angle detection is used. The rotation angle detected by the camera 110 can be considered by shifting the reference position relative to each other, so that each of the detection CCD cameras 111 to 119 captures the bottle 3 It can also be used as a rotation angle. For this reason, in this embodiment, the rotation angle detected by the angle detection dedicated camera 110 is written in each image captured by each of the detection CCD cameras 11 to 11. ing. Then, as in the first embodiment, a comparison is made between the reference image (template) and the image obtained from the glass bottle to be inspected to determine whether or not the bottle B has any warpage.

Here, as shown in FIG. 17, the operation boards 30 to 40 of the image processing device 8 of the mouth inspection device at the first inspection station and the mouth detection device of the mouth inspection device at the second inspection station are provided. The above-described reference image may be created by connecting the arithmetic boards 130 to 139 of the image processing apparatus 108 to the host computer 142 by, for example, the Ethernet 141. That is, images taken by the CCD cameras 10 to 20 and 110 to 119 of each mouth inspection device are sent to the host computer 142, and the host computer is operated based on these images. The reference image can also be created by the computer 142.

So far, one embodiment of the present invention has been described. It is needless to say that the present invention is not limited to the embodiment, and may be implemented in various different forms within the scope of the technical idea.

As described above, according to the present invention, the interior of the bottle is illuminated from the illumination disposed above the bottle of the glass bottle, and the light transmitted from the bottle is illuminated after entering the bottle. By photographing with multiple CCD cameras with different elevation angles arranged around the camera, it is possible to detect the undulation of the bottle and to detect the molding failure of the bottle (especially the thread). Therefore, when changing the type of glass bottle to be inspected, the positional relationship between the illumination, which is the projector, and the CCD camera, which is the receiver, does not need to be adjusted, and the adjustment time when changing the type Can be dramatically reduced.

Further, according to the present invention, even if the bottle is a glass bottle having a thread portion in the bottle portion, the screw around the joint portion of the screw bottle in the upper and lower regions of the thread portion is to be inspected. By doing so, it is possible to accurately detect piri. Industrial potential

INDUSTRIAL APPLICABILITY The present invention can be suitably used for a glass bottle inspection device capable of detecting a molding defect at a bottle portion or a thread portion at a bottle mouth portion of a glass bottle by imaging.

Claims

The scope of the claims

1. An inspection device that illuminates a glass bottle, captures light from the bottle, and detects defects in the bottle by image processing.

Lighting placed above the top of the bottle of the glass bottle,

A plurality of CCD cameras arranged around the bottle portion of the glass bottle; and an image processing device for processing an image obtained by the CCD camera, wherein the plurality of CCD cameras allow light from illumination to enter the bottle mouth. B) Inspection of glass bottles characterized by taking light transmitted through the bottle

2. The glass bottle part according to claim 1, wherein the optical axes of the plurality of CCD cameras are on a line extending radially from the bottle part of the glass bottle.

3. The mouth inspection apparatus for a glass bottle according to claim 1, wherein the plurality of CCD cameras are installed in a hemisphere centered on a mouth of the glass bottle.

4. The glass bottle inspection apparatus according to claim 3, wherein the hemisphere is configured to be vertically movable.

5.A device arranged at a different inspection position from the glass bottle inspection device according to any one of claims 1 to 4,

A first light located on the side of the bottle of the glass bottle;

A plurality of CCD cameras arranged around the bottle opening of the glass bottle,

An image processing device for processing an image obtained by the CCD camera, An apparatus for inspecting a mouth of a glass bottle, wherein the plurality of CCD cameras photograph light reflected or transmitted from the mouth of the bottle after light from the illumination enters the bottle.

6. An angle detecting means for detecting a rotation angle of the glass bottle with respect to a reference position, and a second illumination for the angle detecting means arranged on a side of a bottle of the glass bottle. 6. The glass bottle inspection device according to claim 5, wherein:

7. The glass bottle inspection apparatus according to claim 6, wherein the first illumination and the second illumination emit light that does not interfere with each other.