JPS63222245A - Defect inspector for mouth of bottle - Google Patents

Abstract

PURPOSE:To enable fast and highly accurate inspection, by expanding an inspection area of one screen with the development of a cylinder area of the inner surface of the mouth of a bottle to reduce the frequency of inputting images per bottle. CONSTITUTION:A bottle 1 is lighted with a lighting device 2 through a diffusion plate 3 to take the mouth part thereof with an image sensor 5 and an image signal is stored into a memory 6B through an A/D converter 6A. A development address generation circuit 6E performs a fast reading out of a memory 6B and writes an inner surface image developed into a memory 6F. An computing circuit 6G differentiates a development image circumferentially while being averaged axially to be written into memories 6H and 6I. A peak value detection circuit 6J reads out a differentiated value from the memory 6H to measure a value of a defect. An axial density value from the memory 6I is processed with an addition circuit 6K to detect circumferential periodicity. A decision circuit 6M judges a possible defect from the value of a defect in terms of the circumferentially differentiated value and the periodicity of the screw part at the bottle mouth.

Description

[Detailed Description of the Invention] [Industrial Application Field] This invention relates to an inspection device that automatically inspects defects such as chips and burrs (a type of crack) in the mouth of a bottle with a threaded portion. .

[Conventional technology]

Various automatic inspections of bottle openings have been carried out for bottles without screws, but almost none have been carried out for bottles with screws, and at present the inspection is left entirely to manual inspection.

[Problem that the invention seeks to solve]

For this reason, there is a problem common to manual inspections, namely that high-speed and high-precision inspections cannot be performed.

Accordingly, an object of the present invention is to provide an inspection device that can inspect bottle openings having threaded portions at high speed and with high precision.

[Means for solving problems]

The operation of diffusing and illuminating a bottle mouth with a threaded part from the back and taking an image from a direction at a predetermined angle to the plane containing the bottle mouth and extracting a binarized image of the bottle mouth is performed for one circumference of the bottle mouth. In order to sequentially process each binarized image and inspect it for defects, there is an image developing means for straightening the cylindrical portion of the bottle mouth, and an image developing means for straightening the cylindrical portion of the bottle mouth, and for differentiating this developed image in the circumferential direction and for inspecting the defective image in the axial direction. A first detection means for detecting a defect extending in the circumferential direction and a second detection means for detecting a defect extending in the circumferential direction by utilizing the periodicity of the threaded portion are provided, and the outputs from the rain detection means are integrated to detect the defect. Detect.

[Effect]

By expanding the cylindrical (elliptical) area on the inner surface of the bottle mouth onto a plane to expand the inspection area on one screen, and reducing the number of times images are captured for one bottle, high-speed and highly accurate inspection is possible.

〔Example〕

FIG. 1 is a block diagram showing an embodiment of the present invention.

In the same figure, 1 is a bottle, 2 is an illuminator, 3 is a diffuser plate, and 4
is a rotation mechanism, 5 is an image sensor such as a television camera, 6 is
is an inspection device. Note that the inspection device 6 is an analog/digital (A/D) converter 6A.

Image memory 6B, position detection circuit 6C1 expansion function processor 6D, expansion address generation circuit 6E, expansion image memory 6F, arithmetic circuit 6G, circumferential differential value memory 6H
It is composed of a uniaxial concentration value memory 6■, a peak value detection circuit 6J, a circumferential addition circuit 6K, a periodicity detection circuit 6L, a defect determination circuit 6M, a rotation synchronization circuit 6N, a comprehensive determination circuit 6P, and the like.

A transparent or translucent bottle 1 to be inspected is placed on a diffuser plate 3
The mouth of the bottle 1 is imaged by the image sensor 5. The video image of the bottle opening is A/
The image is converted into a digital image by the D converter 6A and stored in the image memory 6B, while the position of the bottle mouth on the screen is detected by the detection circuit 6C. When setting the inspection device 6 on the inspection line, the expansion function calculation processor 6D calculates the inner diameter DA of the bottle mouth and the depression angle θ of the image sensor 5.

A coordinate expansion conversion table is calculated and stored in the memory within the expansion address generation circuit 6E. The developed address generation circuit 6E reads the contents of the image memory 6B at high speed and writes the developed internal image into the developed image memory 6F. When the arithmetic circuit 6G differentiates the developed image in the circumferential direction, -
Also, the axial direction averaging process is performed, and the circumferential direction differential value is stored in the memory 6H, and the axial direction local average density value is stored in the memory 6H.
Write each in I. The peak value detection circuit 6J reads the differential value from the circumferential differential value memory 6H and measures the amount of defects. Further, the axial density values from the memory 6■ are added in the circumferential direction by an adding circuit 6K, and the periodicity in the axial direction is detected by a detection circuit 6L. The determination circuit 6M determines the presence or absence of a defect using the amount of defects based on the circumferential differential value and the periodicity of the bottle mouth thread.

With the above steps, defect determination for one screen is completed, so
While synchronizing with the rotating mechanism 4, a synchronization circuit 6N captures an image at a different position, repeating the same process as above to inspect the entire inner peripheral surface for defects, and outputting a comprehensive judgment result via the judgment circuit 6P. .

A detailed explanation will be given below with reference to FIGS. 2 to 16. Furthermore, Fig. 2 is an explanatory diagram showing an example of a digital image of the bottle opening, Fig. 3 is a block diagram showing a specific example of a bottle opening position detection circuit, and Fig. 4 is a principle for explaining the principle of detecting the bottle opening position. Figure 5 is a block diagram showing a specific example of the expansion address generation circuit, Figure 6 is an explanatory diagram for explaining the relationship between the expansion area of the bottle opening and the group of elliptical lines for expansion, and Figure 7 is the bottle opening. Explanatory diagram for explaining the coordinate system developed for the inner surface of the part, No. 8
The figure is a conceptual diagram showing an expanded function memory, Figure 9 is a block diagram showing a specific example of an expanded image memory and an expanded image differential circuit, and Figure 10 is for explaining a two-dimensional local memory with a size of 3 x 3. , Figure 11 is a conceptual diagram showing the differential operator, Figure 12 is a conceptual diagram showing the average operator, and Figure 13 is a conceptual diagram showing the average operator.
The figure is a block diagram showing a specific example of a peak value detection circuit.
Fig. 4 is a graph for explaining the relationship between differential value histograms and defects, Fig. 15 is a block diagram showing a specific example of a circumferential addition circuit, and Fig. 16 is a histogram showing the result of circumferential addition of density values. .

The digital image input via the image sensor 5 is a grayscale image of the bottle mouth as shown by reference numeral 11 in FIG. 2, and the position detection circuit 6C determines the center position of the inner surface of the bottle mouth. Specifically, the position detection circuit 6C includes a binarization circuit 6C1°X-direction projection circuit 6C2, a bottle mouth center YA calculation circuit 6C3, a bottle mouth side edge detection circuit 6C4, a center XA calculation circuit 6C5, etc. as shown in FIG. configured. Here,
Assuming that the image binarized by the binarization circuit 6C1 is shown as 13 in FIG.
The vicinity of the tip of the bottle is detected from the projected length histogram obtained in C2, and the starting coordinate Y2 and center coordinate YA are set at positions EA pixel and E8 pixel apart from the Y coordinate Y1 where the X direction projected length is, for example, WA pixel or more. and seek. Next, at the detection position W, 'y c of the bottle opening, the side edge X of the screw opening,
and X2 are determined by the detection circuit 6C4, and their center coordinates are determined as the average value ((X++Xt)/2) to determine the center position 1.
Determine the coordinates (XA, YA) of 4.

A specific example of a circuit for developing the cylindrical surface of the inner surface of the bottle mouth is shown in FIG. That is, the expansion function processor 6D uses the inner diameter DA (mm) of the bottle and the depression angle θ of the sensor 5 at the time of setting.

From this, the major axis W and minor axis H of the inner elliptical area are determined. That is, from the converted value k [pixels/mm], W=h DA , H=W sin θ.

W and H are obtained from the following relational expression. The expansion function calculation processor 6D obtains an elliptic line group from the following equation.

(i=o, 1, 2..., N,) here, PA
indicates the pitch of the elliptic line group. An example of the elliptic line group 16 is shown in FIG. In the figure, ellipse line E116 (i=o, 1.2.
......, N,) has a pitch of PA pixels,
It is defined within the development area 15. In addition, the number of elliptical lines N
, is determined by the following formula.

NE = (2H/PA) ((Nigauss symbol)) When each ellipse line Ei in Fig. 6 is developed, the development surface 18 shown in Fig. 7 is obtained.
At the top, it becomes a straight line Li 21. Furthermore, when the point Q (X, Y) on the image area 17 shown in Fig. 6 is expanded, it becomes the straight line 21 point Q (T, Z) shown in Fig. 7, and when the elliptical area 15 is similarly expanded. This becomes area 20. in this way,
The developed coordinate system 19 consists of the inner circumferential direction T and the axial direction Z of the bottle mouth, and XI = -W cos (j-Pm)Y,
=i-Pa-H-sin (j-P++) (j=0.
1. ......, NF) The following relationship holds true. tl, , P6 is the pitch (r a d) in the circumferential direction. Also, the relationship with the developed coordinates (T, Z) is as follows.

T=j (j=o, 1, 2..., NF)Z
=i (i=o, 1, 2..., NE) ＼
Then, Nr'P-s = π. Also, in this case, PA=1 pixel, and no thinning of the elliptic line group is performed.

For example, as shown in FIG. 8, the expansion function memory 6E2 stores (T,
The (Xm, Ys) coordinates corresponding to 0≦T:5Nr, O≦Z≦N9 are stored.
-pA -H-5in (T-P+) (X,,'y,) is obtained from the equation and stored in the memory 6E2. If, then (Xs, Ys) is forcibly stored as (0,0). In this way, the processor 6D writes the expansion function into the memory 6E2, and reads out (T, Z) and (xs, y,) from the memory 6E2 during defect inspection, and also calculates the amount of positional deviation (XA, yA). ) is corrected by the arithmetic circuit 6E1.

XmXl +XA Y=Y, +YA Then, the contents of the image memory 6B are transferred to the address (x, y
) and write the density value F (X, Y) to the memory 6F at the address (T, Z).

Regarding the image G (T, Z) developed in this way,
FIG. 9 shows a circuit for calculating the circumferential (T) direction differential G and the axis (Z) direction average G2. That is, the density value F (X, Y) of the image memory 6B is stored as the expanded density value G (T, Z') in the memory 6F, and the two-dimensional local memory 6F2 stores it as a 3×3 density value as shown in FIG. A two-dimensional local region 22 is formed.

The output of the two-dimensional local memory 6F2 is processed by a circumferential (T) direction differential operator 2 in a differentiating circuit 6G1 as shown in FIG.
3, and averaged by an average value operator 24 as shown in FIG. 12 in an average value circuit 6G2, and the respective calculation results Gr and G2 are stored in memories 6H and 61, respectively.

The peak value detection circuit 6J shown in FIG. 13 calculates the area value HT of the defect from the contents of the circumferential differential value memory 6H. That is, for every absolute value 1Grl of the circumferential differential value GT, the number of pixels is added by an adder circuit 6J1, and a histogram 25 as shown in FIG. 14 is obtained by an arithmetic unit 6J2. Then, a peak value A26 is detected in a range where the differential value IGtl is smaller than GTA. The processor 6J3 uses this peak value At
Then, find the point on the histogram 25 corresponding to A7/2, and let the differential value of the point that intersects the 1Gtl axis of the straight line connected to the peak value be GTll, which has a differential value larger than the larger of GTA and GTH. The total sum Ht of pixels is output to the determination circuit 6M as the number of defective pixels.

Regarding the axis (Z) direction concentration value Gz (T, Z), the first
A density histogram is obtained by a circumferential addition circuit 6K shown in FIG. That is, the density value for each circumferential direction is read out to the axial address generation circuit 6 by 2, and this is added to the addition circuit 6.
1 is added to , and the result ΣG2 is stored in the memory 6 at 3. An example of the histogram stored in memory 6 at 3 is shown at 27 in FIG. In other words, since the bottle opening thread has periodic peaks as shown in the same figure, each peak position ZI, Zt, Z3...
..., and the periodicity detection circuit 6L detects the peak Z+
, Zt, Z3. The pitch of .

lZz　Z+　Pzl<δ +23 Zt Pzl<δ Here, δ indicates the upper limit value of the pitch deviation.

From the above, while defects having lengths mainly in the axial direction are detected by the circumferential differential value memory 6H, the peak value detection circuit 6J, etc., the axial density value memory 61. Since defects having a length in the circumferential direction are mainly detected by the circumferential addition circuit 6 and the periodicity detection circuit 6L, it is possible to detect defects even in bottles with screws at the mouth. become.

In addition, although the above explanation has mainly been given to a device having a threaded portion, it goes without saying that the present invention can be similarly applied to a device that does not have a threaded portion.

(Effects of the Invention) According to the present invention, by taking an image of the bottle opening from an oblique direction, it is possible to easily capture the threaded portion, etc. as an image, and by developing the inside of the bottle opening using an expansion function table. , the transformed image can be made into horizontal stripes,
Not only can defects be detected through relatively simple image processing, but since the inspection area can be widened, the entire inner periphery can be inspected with fewer side images. In addition, by performing differentiation in the circumferential direction and detecting defects when the number of pixels is greater than the density value determined from the differential histogram, detection accuracy is improved. , it becomes possible to detect defects even if there is a threaded part.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is an explanatory diagram showing an example of a digital image of a bottle opening, FIG. 3 is a block diagram showing a specific example of a bottle opening position detection circuit, and FIG. 4 5 is a block diagram showing a specific example of an expansion address generation circuit. FIG. 6 is a diagram explaining the relationship between the expansion area of the bottle opening and the group of elliptical lines for expansion. Figure 7 is an explanatory diagram to explain the coordinate system of the inner surface of the bottle opening, Figure 8 is a conceptual diagram showing the expansion function memory, and Figure 9 is the expansion image memory and expansion image. A block diagram showing a specific example of a differential circuit, Figure 10 is 3×
11 is a conceptual diagram showing a differential operator, Figure 12 is a conceptual diagram showing an average operator, and Figure 13 is a peak value detection circuit. 14 is a graph for explaining the relationship between the differential value histogram and defects, FIG. 15 is a block diagram showing a specific example of the circumferential addition circuit, and FIG. 16 is a graph for explaining the relationship between the differential value histogram and defects. It is a histogram showing the circumferential direction addition result. Code explanation ■... Bottle, 2... Illuminator, 3... Diffusion plate, 4...
...Rotation mechanism, 5... Image sensor, 6... Inspection device, 6A... A/D converter, 6B, 6E2.6F,
6F1゜6F2, 6H, 61, 6 to 3...Memory, 6
C...Position detection circuit, 6C1...Binarization circuit, 6C
2...X direction projection circuit, 6C3...YA calculation circuit,
6C4...Side edge detection circuit, 6C5...Island calculation circuit,
6D-... Processor for expansion function, 6E... Expansion address generation circuit, 6E1... Position shift correction calculation circuit,
6G... Arithmetic circuit, 6G1... Differential circuit, 6G2...
...Average value circuit, 6J1...Circumferential differential value addition circuit,
6J2... Histogram calculator, 6J3... Peak value pixel number calculation processor, 6K... Circumferential addition circuit,
6 to 1... Addition circuit, 6 to 2... Axial address generation circuit, 6L... Periodicity detection circuit, 6M... Defect determination circuit, 6N... Rotation synchronization circuit, 6P... Comprehensive judgment circuit, 11... Bottle opening, 12... Imaging screen, 13
... Bottle mouth binarization pattern, 14... Bottle mouth center position, 15... Ellipse development area, 16... Ellipse line group for development, 17... Image area, 18... Development Face, 19・
... Expanded marker system, 20... Expanded area, 21... Expanded line group, 22... Two-dimensional local memory area, 23... Differential operator, 24... Average operator, 25... ... Differential value histogram, 26 ... Peak value, 27 ... Density value histogram, 28 ... Periodic region. Agent Patent Attorney Akio Namiki Agent Patent Attorney Kiyoshi Matsuzaki Figure 3 C Figure 4 Figure 5 6, E Figure 6 Figure 7 7114

Claims (1)

[Scope of Claims] The mouth of a bottle having a threaded portion is diffusely illuminated from the back, and an image is taken from a direction at a predetermined angle with respect to a plane containing the bottle mouth, and at least a binarized image of the bottle mouth is obtained. An inspection device that sequentially performs a take-out operation around one circumference of a bottle mouth, processes each binarized image, and inspects the image for defects; , a first detection means for detecting a defect extending in the axial direction by differentiating the developed image in the circumferential direction; and a second detection means for detecting a defect extending in the circumferential direction by utilizing the periodicity of the threaded portion. , A defect inspection device for a bottle opening is characterized in that it detects a defect by integrating the outputs from the first and second detection means.