JPH05157707A - Appearance inspection method of bottle top - Google Patents

Abstract

PURPOSE:To enable a defect such as cutout, crack, and contamination at a top with screw of a light-transmission bottle to be detected, for example, even if the defect is generated in parallel with a flight. CONSTITUTION:A top with a screw of a transmission bottle 1 which rotates around an axis line is irradiated from a side-outer direction for picking up 4 an image of a portion on an inner-periphery surface from a skew outer direction, a shaded image signal of one periphery of an inner-periphery surface of the top whose image is picked up is latched 5, it is subjected to differential processing in periphery direction and axis-line direction of the top, and then binary-coding processing is performed based on a specified threshold for obtaining a logical sum 6. After the logical sum is obtained, it is subjected to a thin-line processing 7, a part which exists separately from other within the thin-lined part is labeled 8, it is judged 11 that a first defect exists based on excess of the number of labels which are obtained by this labeling over a reference number, and then it is judged 11 that there is a second defect based on presence 9 of recessed and projection parts exceeding an allowable limitation at the thin-lined part excluding an edge part of a screw after the thin-line processing.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a bottle mouth appearance inspection method for detecting defects such as cracks, cracks, and stains in a threaded mouth portion of a light-transmitting bottle and performing a bottle mouth appearance inspection.

[0002]

2. Description of the Related Art A conventional bottle mouth appearance inspection apparatus will be described with reference to the block diagram of FIG. 7 and a schematic diagram relating to each processing step relating to the processing of FIG. In FIG. 7, 1 is a bottle as an object, 2 is a turntable, and the bottle 1 is rotated around an axis, and 3 is an illuminating unit, which irradiates the mouth of the bottle 1 from the outside. An image pickup unit 4 picks up an image of the inner peripheral surface portion of the irradiated mouth of the bottle 1 from an obliquely outer side. An image processing unit 20 processes a signal from the image pickup unit 4 to obtain a feature amount related to the appearance. Reference numeral 21 denotes a determination unit, which determines pass / fail based on the feature amount from the image processing unit 20. As the constituents of the image processing unit 20, reference numeral 15 is an original image capturing means, which captures the imaged grayscale image signal for one round of the inner peripheral surface of the mouth portion, and 16 is captured by a differential / binarizing means. The grayscale image signal is differentiated in the circumferential direction of the mouth to obtain an absolute value, and then binarized based on a threshold value. A projecting means 17 projects the binarized image signal in the circumferential direction of the bottle mouth (in the axial direction).
Reference numeral 18 denotes a feature amount calculation means, which determines the feature amount from the pattern when projected, such as the distribution width in the circumferential direction or the area of the projected pattern,
Find the height of the projection pattern.

FIG. 8 shows the operation of the conventional visual inspection apparatus.
Will be mainly described with reference to FIG. Figure 8
(a) is a grayscale image (original image) of one round of the inner peripheral surface of the mouth captured by the original image capturing means 15. In this original image, A is a screw end portion and B is a defect as a defect. Same figure
(b) is an image obtained by differentiating the original image in the circumferential direction by the differentiating / binarizing means 16 to obtain an absolute value and binarizing it based on a predetermined threshold value. In this binarized image, A1 relates to the screw end portion A and B1 relates to the chip B. Since the threads other than the ends of the screws are substantially parallel to the horizontal axis, they are zero when the original image is differentiated in the circumferential direction, and disappear as an image. In addition, the double line with a narrow interval for both A1 and B1 is when transitioning from the bright part to the dark part,
On the contrary, this is to cope with the case where the dark portion is switched to the bright portion. FIG. 6C shows the respective projection patterns A2 and B2 when the image signal binarized by the projection means 17 is projected in the circumferential direction of the bottle mouth portion (in the axial direction). Next, the feature amount calculating means 18 obtains the feature amount from each projected pattern A2, B2, for example, each distribution width Wa, Wb in the circumferential direction, each area Sa, Sb, and each height Ha, Hb. Be done. Next, the determination unit 21 compares each distribution width Wa, Wb as the calculated feature amount with each distribution width Wao, Wbo of the corresponding reference screw portion, and if the distribution width Wa, Wb is within the allowable range, it is a non-defective product, the allowable range. If it deviates from the above, it is judged as a defective product.

[0004]

In the conventional appearance inspection apparatus, in FIG. 8 (a), defects such as chips, cracks, stains, etc.
If it occurs parallel to the thread of the threaded mouth, there is a problem that it does not appear as an image when differentiated in the circumferential direction of the bottle mouth, that is, it is not detected.

The object of the present invention is to solve the above problems of the prior art and to eliminate defects such as chips, cracks, and dirt even when they are parallel to the threads of the threaded mouth. An object is to provide a detectable bottle mouth appearance inspection method.

[0006]

A bottle mouth appearance inspection method according to a first aspect of the present invention is to illuminate a threaded mouth portion of a translucent bottle which is rotated around an axis from the outside and to illuminate it. Imaging the inner peripheral surface portion of the mouth from the outside obliquely;
Capturing a captured grayscale image signal for one round of the inner peripheral surface of the mouth; differentiating the captured grayscale image signal with respect to the circumferential direction of the mouth and the axial direction, and then predetermined. A step of performing a binarization process based on a threshold value; a step of taking a logical sum of the binarized items and then performing a thinning process, or a process of taking a logical sum after the thinning process; Later, in the thinned part,
A step of performing a labeling process for labeling a portion existing separately from the others; a step of determining that there is a first defect based on the number of labels obtained by this labeling process exceeding a reference number; And a second defect is determined based on the presence of the uneven portion exceeding the allowable limit, in the thinned portion excluding the end portion of the screw, after the hardening treatment. ..

The bottle mouth appearance inspection method according to a second aspect is the method according to the first aspect, wherein the labeling process of the thinned portion related to the screw is performed by expanding the thinned line for one round of the inner peripheral surface of the mouth portion. It is performed based on the judgment that parts having the same axial direction coordinate at each end portion in the direction of development are continuous.

A bottle mouth appearance inspection method according to a third aspect is the method according to the first or second aspect, wherein the uneven portion exceeding the allowable limit is a first pixel arbitrarily selected in the thinned portion, and It is obtained based on the fact that the angle formed by the line segment connecting the second and third pixels, which are separated from each other by a predetermined number of pixels on each side, is less than or equal to the allowable value.

[0009]

[Operation] In the bottle mouth appearance inspection method according to claim 1, (1)
The screwed mouth of the translucent bottle that is rotated around the axis is illuminated from the outside, and the inner peripheral surface of the illuminated mouth is imaged obliquely from outside, and (2) the imaged mouth is imaged. The grayscale image signal for one round of the inner peripheral surface of is captured, and (3) the captured grayscale image signal is differentiated in the circumferential direction of the mouth and in the axial direction, and then binarized based on a predetermined threshold value. Process and take the logical sum of these, and (4) for this binarized one, take the logical sum and then perform the thinning process, or take the logical sum after the thinning process, and (5) Labeling is applied to the existing part that is isolated from the others, and (6) It is isolated and isolated from the screw part because the number of labels obtained by this labeling process exceeds the standard number. It is judged that there is the first defect of the shape,
(7) After the thinning process, find the uneven part above the allowable limit in the thinned part excluding the end part of the screw, and based on the existence of the uneven part above the allowable limit, it is not isolated from the screw part but uneven It is determined that there is a terrible shape of the second defect.

Particularly, in the bottle mouth appearance inspection method according to the second aspect, in the thinned portion which is developed for one round of the inner peripheral surface of the mouth portion,
Labeling processing of the thinned portion related to the screw is performed on the basis that it is determined that those having the same axial direction coordinate at each end portion in the direction of the development are continuous.

In particular, in the bottle mouth appearance inspection method according to the third aspect, the first pixel arbitrarily selected in the thinned portion and the second and third pixels spaced from each other by a predetermined number of pixels on the basis of the first pixel. Since the angle formed by the line segment connecting each pixel is equal to or less than the allowable value, the uneven portion exceeding the allowable limit is obtained.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An appearance inspection apparatus as an application example of the bottle mouth appearance inspection method according to the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram of a visual inspection apparatus. This application example is different from the conventional example in the internal configuration of the image processing unit and the determination reference of the determination unit. In FIG. 1, a bottle 1 as an object, a turntable 2, an illumination unit 3, and an imaging unit 4 are the same as in the conventional example. In the image processing unit 10, the original image capturing means 5 captures a grayscale image signal for one round of the inner peripheral surface of the mouth, like the original image capturing means 15 in the conventional example. The differentiating / binarizing / OR means 6 differentiates the captured grayscale image signal in the circumferential direction of the bottle mouth portion and the axial direction, and then binarizes it based on a predetermined threshold value. Take the logical sum of each. Thinning means 7
Performs thinning processing of the logical sum. After the thinning process, the labeling unit 8 performs a labeling process of labeling a part of the thinned part which is isolated from the others. After the thinning process, the unevenness detecting means 9 obtains an uneven part that is equal to or more than the allowable limit, that is, an abnormal uneven part in the thinned part excluding the end part of the screw. Next, the determination unit 11
Is based on the fact that the number of labels obtained by the labeling process exceeds the reference number, and on the other hand, it is determined that there is a first defect. Based on the presence, it is determined that there is a second defect.

Regarding the operation of the visual inspection apparatus, refer to the flowcharts of the front and rear stages of FIGS. 2 and 3 and the schematic diagrams of the images in the respective processing steps of the processing by the visual inspection apparatus of FIGS. 4 (a) to 4 (c). While explaining. In FIG. 2, in step S1, a grayscale image signal for one round of the inner surface of the bottle mouth is captured. A schematic diagram of the corresponding image is shown in FIG. In the figure, F and G are cracks and chips as defects, but of course they may be stains. In step S2, the peripheral direction of the bottle mouth is differentiated (to be precise, its absolute value is taken), and in step S3, 2
Value processing is performed. Similarly, in step S4, differentiation processing is performed in the axial direction of the bottle mouth (correctly, its absolute value is taken), and binarization processing is performed in subsequent step S5. In step S6, the logical sum of the values obtained in steps S3 and S5 is calculated, and the thinning process is performed in step S7.
It is also possible to perform a thinning process after each of the binarization processes of steps S3 and S5 and then take a logical sum of these processes. Note that the calculation of the differential processing is not easy if it is attempted to be rigorous, so a differential operator is used for the approximate calculation, and a differential mask, which is a matrix of third vertical and horizontal cells, is usually applied. In any case, the schematic diagram of the processed image when the above process is completed is shown in FIG.
The image relating to each of the defects F and G substantially represents the actual condition. In this figure, the image of the screw portion and each of the chips F and G becomes a double line with a narrow interval, which is the reverse of the transition from the bright portion to the dark portion in the differential processing, as described in the conventional example. This is because it corresponds to the transition from the dark portion to the bright portion.

Turning to FIG. 3, the labeling process is performed in step S8. The labeling process is a process of labeling (numbering) a part that is isolated from the others and exists in isolation. This is a defect that appears as an isolated part on the image,
It is performed in order to distinguish from a screw portion that appears as an image in isolation. The labeling process will be described with reference to FIGS. 4B and 4C and the explanatory diagram relating to the labeling of the screw portion in the process of the visual inspection apparatus of FIG. The method of this labeling process is claimed in claim 2.
Corresponding to. First, as shown in FIG. 4 (b), temporary label numbers (1) to (11) are assigned to each apparently isolated line image as an image.
Attach. In the left column of the explanatory diagram of FIG. 5, temporary label numbers are
In order from (1) to (11), in the middle column, Y lines at the left and right ends of the screen of the line image corresponding to each temporary label number are displayed.
(Vertical direction) Enter the coordinates. In the right column, the book label numbers ~, which are finally decided under the coordinate data judgment as described below, are added.

In FIG. 5, the Y-coordinates at the left end and the right end of the line image of the temporary label number (1) are 15 and 15, respectively (see FIG. 4B, the same applies hereinafter), and the temporary label number ( The coordinates of the line image in 2) are 20 and 20, respectively. However, the line image of the temporary label number (3) does not have the left end part, and 60,8 at the right end part.
There are two, 0. Similarly, the line image of the temporary label number (4) does not have the left end portion, and the right end portion is 65,75. Since the line image of the temporary label number (5) is located in the middle part, there are no coordinates at the left and right ends. Similarly, the left and right coordinates can be obtained for the line images with the respective temporary label numbers (6) to (11). So, for the judgment of coordinate data,
It is determined that "the line images of two temporary label numbers having the same left coordinate and right coordinate are continuous". In other words, in other words
It is determined that "the line images of two temporary label numbers having different left and right coordinates are independent".

For example, in FIG. 5, (a) since there is no temporary label number having the same coordinates as the left and right coordinates of each temporary label number (1) and (2), each temporary label number (1) The line image of (2) is judged to be a single image.
Therefore, this label number is attached to each item (see Fig. 4 (c), the same applies to any of the following items). (b) Temporary label number (3) right 60 and temporary label number (6) left 6
Since 0 is the same, it is determined that the temporary label number (3) and the temporary label number (6) are continuous. Next, since the right 80 of the temporary label number (3) and the left 80 of the temporary label number (9) are the same, the temporary label number (3) and the temporary label number (9) are continuous. to decide. The right 95 of the temporary label number (6) and the left 95 of the temporary label number (10) are the same, so the temporary label number
It is determined that (6) and the temporary label number (10) are continuous.
From the above, it can be said comprehensively that the line image of each label number is abbreviated by that number, and the chain relationship is schematically shown as (9) ─ (3) ─ (6) ─ (10 ). Therefore, a common main label number is given to the line images of each of these temporary label numbers as being essentially continuous. (c) The detailed description of the line images of the temporary label numbers (4), (7), (8), and (11) is omitted, but the same judgment is made for the coordinate data, and they have a chain relationship, (8) ─ (4) ─ (7) ─ (1
1) Therefore, a common main label number is given to the line images of each of these temporary label numbers as being essentially continuous. (d) The line image of the temporary label number (5) is given as a single label image because each of the left and right coordinates does not exist.

Above, the explanation of step S8 in FIG. 3 is finished, and the process moves to the next step. In step S9, it is determined whether the number of labels is the same as the reference value. This reference value is set as the number of labels in a state where there is no defect, and is 4 as the sum of 2 for one round of the convex portion near the upper end of the bottle mouth and 2 for the screw portion. In this case, since the actual number of labels is 5, the determination in step S9 is NO and step S9
At 12, it is determined that there is a first defect. If the determination in step S9 is yes, the process moves to step S10. In step S10, unevenness detection is performed based on the image subjected to the thinning processing in step S7 in FIG. 2, and based on this, unevenness determination is performed in step S11. FIG. 6 is an explanatory diagram related to the determination of the uneven portion in the process of the appearance inspection device. In FIG. 6, a certain pixel P0 on the thinned portion is selected, and a predetermined number of pixels on each side, starting from this pixel P0, in this case, pixels P1 and P2 separated by 5 pixels are selected. Then, the angle formed by the line segment connecting each pixel Po, P1 and the line segment connecting each pixel Po, P2 is obtained, and this angle is used as the unevenness value representing the degree of unevenness. When the unevenness value is θ and the allowable value α is less than or equal to the allowable value α, the determination is bad, and when the allowable value α is exceeded, the determination is good. By the way, as the unevenness value, in addition to the above determination method, the length of the line segment connecting the pixels P1 and P2, or the perpendicular length from the pixel Po to the line segment connecting the pixels P1 and P2 can be used. .. Further, since the unevenness of the screw end portion naturally exceeds the allowable value, the end portion of the screw is excluded from the detection of the unevenness portion. In other words, the pixels Po, P1, P2 are not selected at the end of the screw. Returning to FIG. 3 again, in step S11, if the unevenness value is less than or equal to the allowable value, that is, YES, it is determined as a non-defective product in step S14, and if the unevenness value exceeds the allowable value, that is, if NO, there is a second defect in step S13. The judgment ends and the process ends.

[0018]

According to the bottle mouth appearance inspection method according to any one of claims 1 to 3, there are the following two effects. (1) The captured grayscale image signal for one round of the inner peripheral surface is differentiated with respect to the circumferential direction of the mouth and the axial direction, and then binarized based on a predetermined threshold value, and the logical sum is calculated. Therefore, defects such as chips, cracks, and stains can be detected even when they are parallel to the threads of the threaded opening, and the reliability of the appearance inspection can be improved. (2) Since the defect determination is performed separately for each of the following first and second defects, the content of the defect can be analyzed and a clue for quality improvement can be obtained. That is, the first defect is one isolated and isolated from the screw part based on the number of labels obtained by the labeling process exceeding the reference number, and the second defect is a part except the end part of the screw. However, it is based on the existence of uneven portions that are not isolated but are above the allowable limit.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an appearance inspection apparatus as an application example of a method of the present invention.

FIG. 2 is a first-stage flowchart showing the operation of the appearance inspection device.

[FIG. 3] Similarly, a flowchart of the subsequent stage

4A to 4C are schematic diagrams of images in respective processing steps regarding processing by an appearance inspection device.

FIG. 5 is an explanatory diagram related to labeling of a screw portion in processing of an appearance inspection device.

FIG. 6 is an explanatory diagram related to the determination of uneven portions in the process of the appearance inspection device.

FIG. 7 is a block diagram of a conventional appearance inspection device.

FIG. 8: Regarding processing by a conventional appearance inspection device,
(c) is a schematic diagram of images related to each processing step

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Bottle 2 Rotating table 3 Illumination section 4 Imaging section 5 Original image capturing means 6 Differentiating / binarizing / logical summing means 7 Thinning means 8 Labeling means 9 Concavo-convex detecting means 10 Image processing section 11 Judging section

Claims (3)

[Claims] 1. A step of illuminating a threaded mouth portion of a translucent bottle which is rotated around an axis from a lateral outside, and imaging an inner peripheral surface portion of the illuminated mouth portion from an oblique outside. A step of capturing a captured grayscale image signal for one round of an inner peripheral surface of the mouth portion; a predetermined process after differentiating the captured grayscale image signal with respect to the circumferential direction of the mouth portion and the axial direction. A step of performing binarization processing based on a threshold value;
A process of taking a logical sum and then performing a thinning process on the binarized one, or taking a logical sum after the thinning process; and after the thinning process, the thinned part is separated from the others. A labeling process for labeling existing portions; a step of determining that there is a first defect based on the number of labels obtained by this labeling process exceeding a reference number; and; after the thinning process. A step of obtaining a concavo-convex portion exceeding an allowable limit in the thinned portion excluding the end portion of the screw and determining that there is a second defect based on the existence of the concavo-convex portion exceeding the allowable limit. Method of visual inspection of bottle mouth. 2. The method according to claim 1, wherein the labeling process of the thinned portion related to the screw is a thinned portion that has been developed for one round of the inner peripheral surface of the mouth, and each end in the direction of the development. A method for inspecting the appearance of a bottle mouth, which is performed based on the determination that parts having the same axial direction coordinate are continuous. 3. The method according to claim 1, wherein
The concavo-convex portion that exceeds the allowable limit is a line segment that connects the first pixel arbitrarily selected in the thinned portion and the second and third pixels that are separated from each other by a predetermined number of pixels on each side. A method for inspecting the appearance of a bottle mouth, which is obtained based on that an angle formed is equal to or less than an allowable value.