JP3199984B2 - Optical member inspection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for inspecting a transparent optical member made of plastic, and more particularly to an apparatus for detecting a flaw generated in a mold by using an image processing technique.

[0002]

2. Description of the Related Art When a plastic optical member is removed from a mold, if the molded product is not sufficiently cured, the plastic may adhere to the mold and remain. When the plastic adheres to the mold, a convex portion is formed in the mold. If the molding is continued without removing the convex portion, an unnecessary concave portion is formed as a mold flaw in the molded product. In addition, if the plastic piece adheres to the mold and cures for a long time, a concave scratch may remain on the mold even after the plastic piece is released. Furthermore, the mold is configured so that the lens molding part can be removed alone,
When this part is assembled to the mold body, the part may be scratched. In this specification, a concave defect of a molded product caused by a plastic piece remaining in a mold in this way, or a convex defect generated in a molded product by shaving the mold itself is generally referred to as a mold flaw. And

[0003] When a mold flaw is found in a molded product that is unacceptable as a product, it is necessary to immediately feed it back to the production line to remove the plastic adhered to the mold.

[0004]

However, conventionally,
Since the inspection of optical members was performed visually by the inspector,
There is a problem that it is difficult to distinguish the mold flaw from other defects, the feedback to the production line is delayed, and the product yield is deteriorated.

[0005]

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems of the prior art, and provides an optical member inspection apparatus capable of detecting a mold flaw by separating it from other defects using an image processing technique. The purpose is to provide.

[0006]

SUMMARY OF THE INVENTION An optical member inspection apparatus and an inspection method according to the present invention focus on the fact that a mold flaw is a defect that continuously occurs in the same position on an optical member in the same shape, and uses image processing technology. Is used to determine a mold flaw from a binarized image by a statistical method.

[0007] An input image obtained by photographing a transparent optical member (test object) such as a plastic lens is binarized to extract defects serving as mold flaw candidates, and molded into the same mold. Defects appearing at the same location on the optical member are considered as mold defect candidates, and the number of appearances is counted by an appearance counter. Since there is a possibility that multiple type scratch candidates may appear,
A plurality of appearance counters are prepared so that the number of appearances of each type defect candidate can be counted independently. Then, when the count value of the appearance counter reaches a predetermined number of times of recognition of the type defect, the defect of the type defect candidate counted by the counter is recognized as a type defect.

Further, even when the appearance of a defect at the same location is temporarily interrupted, if the defect reoccurs within a predetermined allowable number of discontinuities, it is assumed that the interruption is due to omission of detection, and the number of appearances before the interruption is reduced. It can also be configured to restart counting by using this. In this case, a plurality of non-appearance counters are provided corresponding to the appearance counters, and if the defect of the type defect candidate no longer appears before the count of the corresponding occurrence counter reaches the number of times of type defect recognition, it does not appear. Count the number of times. The appearance counter is reset when the value of the non-appearance counter reaches a predetermined allowable number of discontinuities, and the non-appearance counter is used when a defect reappears at the same location before the value reaches the allowable number of discontinuities. Reset.

It is desirable to provide a warning means for notifying an inspector or the like when a mold flaw is recognized. By issuing a warning using sound or light, the inspector can know that a mold flaw has occurred.

[0010] An input image is obtained by diffusing a light beam from a light source by a diffusing means having a central region having a low diffuse transmittance and a peripheral region having a high diffuse transmittance to be incident on a test object and transmitting the light beam through the test object. Is obtained by photographing the test object by a photographing means provided at a position where.

By making the diffusion transmittance of the diffusion means have the above distribution, light from the central region and light oblique to the optical axis from the peripheral region enter the test object. The image of the test object is formed mainly by light from a low-luminance central region.

If there is a defect such as a flaw generated on the surface such as a mold flaw, low-brightness light from the central region is scattered and attenuated in a portion corresponding to the defect, but is high from the peripheral region. Since the light of the brightness is scattered and reaches the photographing means, the amount of light of the defective portion on the image plane increases as a result, and appears as a region of higher luminance than the surrounding component region in the photographed image.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the optical member inspection apparatus according to the present invention will be described below. In the apparatus of the embodiment, a plastic optical member is inspected. First, a schematic configuration of an optical member inspection apparatus according to the present invention will be described with reference to FIG.

A CCD as a photographing means for photographing a test object
An output signal from the camera 30 is input to an image processing device 40 as a defect extracting unit. The image processing device 40 binarizes the input image to extract a defect of the test object, displays the extracted information on the monitor display 50, and outputs the information to the control device 60 as a mold flaw inspection device.

The control device 60 controls a warning means 70 for giving a warning by light when a mold flaw occurs, and a component supply device 80 for supplying an optical component to the inspection position.
A plurality of appearance counters respectively provided in the memory 90;
The occurrence of a mold flaw is detected using the non-appearance counter.

The control unit 60 determines the position and size of a scattering defect serving as a mold flaw candidate from the defects extracted by the image processing unit 40, and determines the same location on the optical element molded with the same mold. And the number of appearances is stored in an appearance counter in the memory 90.

When the value of the appearance counter has reached the predetermined number of times of recognition of the type defect, the control device 60 recognizes the defect as a type defect and outputs a signal to the alarm device 70 to generate an alarm. The value of the occurrence counter is reset if the defect is completely discontinued, but if the defect recurs after a relatively short interruption, the value of the occurrence counter is not reset assuming that it is due to missing detection. Use the previous count value. For this reason, the control device 60 counts the number of unappearances by an unappearance counter provided in the memory 90 in correspondence with the appearance counter, and controls the number of occurrences of the unappearance counter to be equal to or less than the predetermined allowable discontinuity number. If the defect is found again, the counting is restarted without resetting the value of the appearance counter.

The appearance counter is reset when its value reaches a predetermined number of times of recognition of a type defect, and when the value of the non-appearance counter reaches a predetermined allowable number of discontinuities. The non-appearance counter is reset when the value reaches a predetermined allowable number of discontinuities and when a defect serving as a mold flaw candidate is found before reaching this value.

For example, assuming that the number of times of type defect recognition is 20, and the allowable number of discontinuities is 5, a defect appearing at the same location is determined to be a type defect, not only when a defect appears continuously more than 20 times, Unless the number of non-appearances does not reach 5 consecutively even when the appearance is interrupted, a case where a total of 20 occurrences occur sporadically is included.

The control device 60 includes a certifying means for certifying a defect of a mold flaw candidate as a mold flaw under the above conditions, a first reset means for resetting a value of an appearance counter, and a second resetting means for resetting a value of a non-appearance counter. 2 resetting means.

Subsequently, the CCD camera 30 will be described with reference to FIG.
The configuration of an optical system for forming an image photographed by the camera will be described.

The optical system of the apparatus comprises a light source 10 and this light source 1
First and second diffusion plates 2 for diffusing the light beam emitted from zero
And a diffusion means 20 composed of 1 and 22;
The CCD camera 30 captures and captures a light beam transmitted through the diffusing unit 20 and transmitted through the positive lens 1, which is the test object, and a light beam transmitted around the test lens 1.

The light source 10 and the test lens 1 are arranged on the optical axis of the CCD camera 30. CCD camera 30
Is composed of a photographing lens 31 and a CCD sensor 32, and is adjusted so that the vicinity of the center in the thickness direction of the test lens 1 is set as the focus plane P. That is, the focus plane P and CC
The image receiving surface of the D sensor 32 is optically conjugate via the taking lens 31, and the image of the test lens 1 on the focus plane P is
It is formed in the range indicated by the symbol O on the CCD sensor.

In order to secure the amount of light taken in the CCD camera 30, the diffusing means 20 and the CCD camera 30
It is desirable to provide a condenser lens for condensing the divergent light flux between them. In this example, the positive lens 1 arranged as a test object functions as a condenser lens.

The first and second diffusing plates 21 and 22 are both substantially similar to the planar shape of the lens 1 to be measured, and the area of the second diffusing plate 22 is smaller than that of the first diffusing plate 21. . These diffusion plates 21 and 22 are provided perpendicular to the optical axis so that the optical axes pass through their respective centers. Also,
These diffusers have the same or different diffuse transmittances. Therefore, when the diffuser 20 is considered as a whole, the central region where the first and second diffusers overlap has a low diffuse transmittance, and the diffuser has a low diffuse transmittance. In the peripheral region where no light is transmitted, the diffuse transmittance becomes relatively high.

The size of the second diffusion plate 22 is determined so that the range of light vertically emitted from the second diffusion plate 22 substantially matches the lens 1 to be measured. As a result, all the vertical emission components from the central region enter the lens 1 to be inspected, and the first
A component which vertically transmits through the diffusion plate 21 and does not pass through the second diffusion plate 22, that is, a vertical emission component from the peripheral region, does not enter the lens 1 to be measured. The reason why the planar shape of the first diffusion plate 21 is formed to be similar to the shape of the test object is to make oblique emission components from the peripheral region uniformly enter the test object from all directions.

FIG. 3 shows an example of the planar shape of the test lens and the diffusion plates 21 and 22. As shown in FIG. 3 (A-1), when the test lens is the finder lens 1a having a rectangular planar shape, the first and second diffusion plates 21a and 22 are used.
It is desirable that the planar shape of a be a rectangle as shown in FIG. When the test lens is a general circular lens 1b as shown in FIG. 3 (B-1), the planar shape of each of the diffusion plates 21b and 22b is shown in FIG. 3 (B-2). It is desirable to make the street circular.

Since the inspection apparatus of the embodiment is configured to inspect a plastic lens formed by a multi-cavity mold without cutting it off from a runner, a gate is provided on the lens to be inspected as shown in FIG. Runner R is connected via G.

As shown in FIG. 4, the captured image includes a high-luminance background area B mainly formed by high-luminance components in a peripheral area that does not pass through the second diffusion plate 22, and An image (component area) S of the test lens mainly formed by a low-luminance component in the central area transmitted through the two diffusion plates is included.

By adjusting the range of the vertical emission component from the central region to the shape of the lens 1 to be measured,
By making the shape of the second diffusion plate 22 similar to the planar shape of the test lens 1, the component region where the test lens is arranged and the background region in the image as described above are clearly separated. This makes it extremely easy to separate the target area in the image processing described later.

When the second diffusion plate 22 is not provided, the component region has a somewhat lower luminance than the background region due to absorption and diffusion by the component, but a clear luminance difference as in the embodiment is obtained. Does not occur.

Here, if a defect that absorbs light, such as black dust contained in the optical member, is present on the surface or inside of the lens 1 to be inspected, a part of the transmitted light from the central region where the lens image is formed is absorbed. As a result, since light does not reach the CCD sensor 32, a defect image DL having a lower luminance than the component region is generated in the component region S having the intermediate luminance as shown in FIG.

If a defect that scatters light on the surface of the lens 1 to be inspected, for example, white dust or a scratch (including a mold scratch) on the surface of the optical member, the light scatters due to the defect.
If there is no defect, the range O of the lens image on the CCD sensor 32
, A part of the high-intensity oblique emission component Lop from the peripheral region that does not reach the range O of the lens image, and as shown in FIG. Occurs.

For example, if a low-brightness image DL based on an absorptive defect and a high-brightness image DH based on a scatterable defect exist on a scanning line in the X-axis direction, the output of a pixel row along this scanning line is output. Is as shown in FIG. 7 (A). The image processing device 40 performs binarization using the two thresholds SH1 and SH2, thereby obtaining two different values as shown in FIGS. 7B and 7C.
Each type of defect can be extracted independently.

As described above, by setting the light amount distribution of the illumination light in two stages using the two diffusion plates, it is possible to reduce the absorption defect in the component area S having a lower brightness than the background area B. Can detect the defect as an image DL of lower luminance than the component area S, and in the case of a scattered defect, as a high luminance image DH. Therefore, it is possible to simultaneously detect defects having different properties from one image data. it can.

Next, the procedure of measurement using the above-described apparatus will be described with reference to a flowchart. As a preparation before inspection, information relating to a component corresponding to an optical member to be inspected is loaded as a data table. In addition to selecting a diffuser suitable for the part shape according to the information,
Set the shooting magnification.

The flow of the whole inspection is shown in the flowchart of FIG. Input an image from the CCD camera (Step A-
1) A component region corresponding to the image of the test lens is separated based on the distribution of luminance (step A-2).

If it is determined that the part is not located at the predetermined inspection position and is missing during the part area separation processing, a loss flag is set during the part area separation processing, and this loss flag is set in the inspection loop. It is determined whether or not to continue the inspection depending on whether or not the inspection has been performed (step A-3).

The image of the separated component region is separated into a scattering defect having a higher luminance than the base luminance and an absorbing defect having a lower luminance by the dynamic binarization processing (step A-4). Are extracted (step A-5).

In steps A-6 and A-9, the occurrence and disappearance of the mold flaw are inspected (subroutines in FIGS. 9 and 10). If the mold flaw occurs or disappears, the inspector is immediately notified by the alarm device. (Steps A-7, 8, 10, 11).

The image processing device 40 judges the quality of the lens to be inspected based on the extracted result (step A-12), and displays the judgment result on the monitor display 50 (step A-13). In step A-14, it is determined whether there is a part to be inspected next, and if there is a part, the part is replaced (step A-15), and the processing from step A-1 is repeated.
If there are no parts, the inspection ends.

Next, a subroutine for detecting a mold flaw and managing a mold flaw included in the inspection flow is shown in FIGS.
It will be described based on.

In the pattern flaw detection processing shown in FIG. 9, pattern flaws are extracted from the binarized image by a statistical method, and in principle, a figure which continuously appears in the same place for a predetermined number of times or more is recognized as a pattern flaw. Is a process of detecting as When the test object is molded by a multi-cavity mold, the data of the mold flaw candidate is stored for each cavity, and the data is stored in the cavity where the test object to be inspected is molded. The corresponding data is sequentially read and used.

The input is the coordinates of the center of gravity of the figure in the binarized image,
The number M of figures, the barycenter coordinates of pattern flaw candidates, the number K of pattern flaw candidates, the position error margin R when judging the difference of a figure based on the barycenter coordinates, the number of pattern flaws recognized as a type flaw S, and the allowable number of discontinuities C (Step B-1).

In the following processing, in steps B-2 to B-17, it is determined whether or not there is a matching figure based on the pattern defect candidate, and an existing pattern defect candidate is registered as a pattern defect or deleted from the candidate. After this is completed, in steps B-18 to 26, a new mold flaw candidate is added based on the figure while avoiding duplication with the existing mold flaw candidate.
When both the number of pattern flaw candidates and the number of figures are “0”, after the processing and determination of steps B-1, 2, 3, 18, and 19, the inspection loop is executed without performing any substantial processing. Return.

If there is a pattern defect candidate, it is determined whether or not there is a figure matching the pattern defect candidate based on the distance between the barycenter coordinates of the pattern defect candidate and the barycenter coordinate of the figure.
(Steps B-2 to B-8). If there is a figure that matches the pattern defect candidate, the number of continuous appearances, which is an index of the frequency of appearance, is incremented, and the number of consecutive non-appearances is reset to 0 (steps B-9, 10).

If the number of continuous appearances is equal to or greater than the predetermined number of times of type scratch recognition S, it is determined that a type flaw has occurred, a type flaw detection flag is set, and this candidate is registered as a type flaw.
(Steps B-11 to 14). The registered data is the coordinates of the center of gravity of the mold flaw and the cavity number of the found part.

If there is no figure that matches the pattern flaw candidate even after collation with all figures, the number of consecutive non-appearances of the pattern flaw candidate is incremented. The candidate is deleted, and if it is smaller, the matching of the next type defect candidate is continued as it is (step B-5, 1
6,17,15).

When all the pattern defect candidates have been collated, it is determined whether or not there is a pattern defect candidate that matches each figure (steps B-18 to 25). Is added as a new type defect candidate (step B-26).

According to the above processing, the occurrence of die flaws can be detected by statistically evaluating continuously occurring defects.

The die flaw management process shown in FIG. 10 is a process for detecting disappearance of a generated die flaw. The plastic that has adhered to the mold that causes the mold flaws may adhere to the molded product and peel off from the mold while molding continues, and even if mold flaws occur during the mold flaw detection process, the If the flaws disappear, no feedback to the production line is required.

The mold flaw management processing is configured by simplifying the mold flaw detection processing, and it is determined whether or not there is a figure which matches all the mold flaws (steps C-2 to C-8). , The number of consecutive non-appearances is reset to 0, and the determination of the next pattern flaw is continued (steps C-9 and C10).

If there is no matching figure, the number of consecutive discontinuous occurrences is incremented (step C-11). If this is equal to or greater than the predetermined allowable number of discontinuities C, the registration of the pattern flaw is deleted and the pattern flaw is deleted. Set the erase flag (Step C-13,
14). When the collation for all the type flaws is completed, the process returns to the inspection loop and continues.

By the above-described mold flaw management processing, it is possible to monitor whether the mold flaw detected by the mold flaw detection processing continuously appears or disappears. Statistical management of scratches becomes possible.

[0055]

As described above, according to the present invention, it is possible to statistically detect a mold flaw generated in a test object by a plastic piece adhered to a lower valley based on an image obtained by photographing the test object. As a result, it is possible to speed up the feedback to the production line and improve the product yield.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a processing system of an optical member inspection apparatus according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing an optical system of the optical member inspection device according to the embodiment of the present invention.

FIG. 3 is an explanatory diagram showing the shape of a test object and the shape of a diffusion unit in the apparatus of the embodiment in comparison.

FIG. 4 is an explanatory diagram showing an image when the test lens captured by the apparatus of the embodiment has no defect.

FIG. 5 is an explanatory diagram showing an image in a case where the test lens has an absorptive defect photographed by the apparatus of the embodiment.

FIG. 6 is an explanatory diagram illustrating an image when a test lens captured by the apparatus according to the example has a scattering defect.

7A and 7B show examples of luminance distribution on one scanning line of an image captured by the apparatus of the embodiment, where FIG. 7A shows a signal of an original image, FIG. 7B shows a signal obtained by binarizing a low luminance component, and FIG. ) Is a signal obtained by binarizing the high luminance component.

FIG. 8 is a flowchart illustrating an entire inspection process of the apparatus according to the embodiment.

FIG. 9 is a flowchart showing a subroutine for die flaw detection in the inspection flow.

FIG. 10 is a flowchart showing a subroutine of die flaw management in the inspection flow.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Test lens 10 Light source 20 Diffusing means 30 CCD camera 40 Image processing device 50 Monitor display 60 Control device 70 Warning device 80 Parts supply device 90 Memory

──────────────────────────────────────────────────続 き Continuation of the front page (72) Inventor Atsushi Kida 2-36-9 Maeno-cho, Itabashi-ku, Tokyo Asahi Optical Industry Co., Ltd. (56) References JP-A-5-72143 (JP, A) JP Hei 6-148083 (JP, A) JP-A-60-70334 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G01N 21/84-21/958 G01M 11/00-11 / 08 B29D 11/00 B29C 33/74 G01B 11/00-11/30 G06T 7/00

Claims (4)

(57) [Claims] 1. A defect extracting means for binarizing an input image obtained by photographing a transparent optical member and extracting a defect which is a mold flaw candidate, and the same on an optical member molded with the same mold A plurality of appearance counters that count the number of appearances of the defect appearing at the position of the above as a type defect candidate; and when the occurrence counter reaches a predetermined number of type defect recognition times, the defect of the type defect candidate is identified as a type defect. A plurality of certification means are provided in correspondence with the appearance counter, and when the defect of the mold flaw candidate no longer appears before the count of the corresponding appearance counter reaches the number of times of mold flaw recognition, A non-appearance counter for counting; first reset means for resetting the value of the appearance counter when the value of the non-appearance counter reaches a predetermined allowable number of discontinuities; An optical member inspection apparatus characterized by comprising a second reset means for resetting the value of the non occurrence counter when defects reappeared in the same place before the value of the pointer reaches the discontinuous allowable number. 2. The optical member inspection apparatus according to claim 1, further comprising a warning unit for notifying a mold flaw when the mold flaw is recognized. 3. The input image, wherein a light beam from a light source is diffused by a diffusion unit having a central region having a low diffuse transmittance and a peripheral region having a high diffuse transmittance, and is incident on an optical member.
The optical member inspection apparatus according to claim 1, wherein the optical member is obtained by photographing the optical member by a photographing unit provided at a position where a light beam transmitted through the optical member reaches. 4. The optical member inspection apparatus according to claim 3, wherein a shape of the diffusing unit is substantially similar to a plane shape of the optical member.