JPH0731131B2 - Method for spotting periodic pattern - Google Patents

Description

The present invention relates to a shadow mask used for a cathode ray tube for a color television, a color separation filter for a color image pickup device, a color filter for a liquid crystal display panel, and an electron tube. Units (constant unit patterns) that have certain optical properties and shapes, such as used mesh electrodes, VDT filters, leak filter mesh filters, rotary encoders, linear encoders, IC photomasks, Fresnel lenses, and lenticular lenses, are primary. Industrial products that are regularly arranged in the original direction or two-dimensional direction, or unit patterns,
The optical properties, shapes, and the one-dimensional direction and the two-dimensional direction of the array pitch are gradually changed and repeated, and the optical properties due to the shape, size, or position of the unit pattern of the industrial product are arrayed. The present invention relates to a method for inspecting spots caused by nonuniformity.

(Prior Art) Conventionally, in the inspection of industrial products having the above-mentioned periodic pattern, generally, the elements such as the shape, size, array pitch, or color of the unit patterns constituting the periodic pattern are measured with a dedicated measuring instrument. However, even if the measurement result satisfies the specified quality standard, if the non-uniformity of the optical property is remarkable when the whole product in which a large number of the unit patterns are arranged is viewed, the product is selected. It becomes necessary to remove it as a defective product.

For example, in the case of inspecting a shadow mask used in a cathode ray tube for a color television or a color separation filter for a color image pickup device, the above-mentioned elements are inspected for each unit pattern by a dedicated measuring device, and further, an artificial means in the final product. There is a step of inspecting the spot visually recognized by.

(Problems to be Solved by the Invention) According to the conventional method, in the step of inspecting the spots of the industrial product having the unit pattern in the periodic pattern composed of the repeating arrangement of the unit pattern, the naked eye or a microscope is used. Although it is usually performed, there are problems that a large amount of manpower is required to inspect a large number of products and that the sensory inspection is lacking in reliability and inspection accuracy. .

Therefore, it is an object of the present invention to provide a method that enables efficient and highly accurate inspection of spots on the industrial products.

[Structure of the Invention] (Means for Solving Problems) The present invention relates to the shape, size, arrangement pitch of the unit pattern of a periodic pattern composed of a repeating arrangement of unit patterns
Alternatively, in a method of automatically inspecting spots caused by non-uniformity of color or the like, the periodic pattern is divided into a plurality of pixels by an image input device and stored in an image memory as digital data for each pixel in frame units. However, a means for obtaining frame-by-frame image data, a random noise component generated in the step of obtaining the image data, a shading component due to the luminance distribution of the light source, and pulsed noise due to stains attached to the optical system or the product. Means for facilitating recognition by suppressing (hereinafter referred to as fixed noise) or the like and extracting a spot component of the periodic pattern, and a product quality based on the image data obtained as described above.
Means for determining a defect are provided to achieve the above-mentioned object.

(Operation) According to the present invention described above, by applying the digital image processing technique to extract the spot component of the periodic pattern, the spot component can be emphasized and displayed. In the process, if used as an auxiliary means for visual recognition, there is an effect that it is possible to easily recognize spots that are difficult to catch by the conventional method.

Furthermore, since the spot inspection can be automated by comparing the extracted spot component and the determination standard with an electric means, a large amount of manpower is not required, and the accuracy can be quickly adjusted according to the flow of the manufacturing process. There is also the effect that a good inspection can be performed.

(Examples) Hereinafter, the present invention will be described in more detail based on the illustrated examples.

FIG. 1 is a block diagram showing an example of an apparatus for carrying out the method according to the present invention. In the figure, 1 is a TV camera, 2 is an image processing device, 3 is a control unit, 4 is an image display, and 5 is a drive mechanism. , 6 is an object to be inspected, 7 is a diffusion plate, 8 is an incandescent lamp, 9 is a DC power supply, 10 is XY
Show the stage. The operation of the present invention will now be described. First, the inspected body 6 is illuminated by the transparent illuminating section composed of the incandescent lamp 8 and the diffusion plate 7 which are turned on by the DC power supply 9,
A video signal obtained by photographing the inspection area in the inspection object 6 with the TV camera 1 is input to the image processing device 2, and the video signal is A / D converted in the image processing device to generate a frame. The inspection area is stored in the memory, the processing described later is performed, and the inspection area is moved by the inspection object moving mechanism including the drive mechanism 5 and the XY stage 10 under the control of the control device 3, and the spot inspection of the inspection object 6 is performed.

FIG. 2 shows an example of the inspection object 6, which is a periodic pattern having periodic openings as the unit pattern 11.

FIG. 3 schematically shows the data corresponding to each processing step as an analog waveform in order to explain the processing contents of the image processing apparatus 2 and the effect thereof.

FIG. 3A is an example of the aperture ratio distribution of the object 6 to be inspected, which is an imaging condition in which the shape of the unit pattern 11 is ignored in the video signal waveform of the TV camera 1, for example, the frame memory of the image processing apparatus 2. The figure shows the case where the imaging conditions are such that there are about 10 unit patterns 11 on the inspected object 6 corresponding to one pixel. In the figure, A indicates a portion where the aperture ratio changes slowly, and B indicates the aperture ratio. Periodically changing part, C, shows a large change in aperture ratio and also an isolated part.

FIG. 3B shows the TV camera 1 input to the image processing apparatus 2.
The video signal waveform for one scanning line corresponds to the aperture ratio distribution shown in FIG. 3 (a), and the illuminance distribution of the transillumination unit is non-uniform and the sensitivity distribution of the imaging surface is non-uniform. Signal component that shows a gradual change due to sex (shading component)
And a random noise component generated in the video signal processing circuit and a local change component (fixed noise portion) D of the video signal due to dirt of the optical system.

FIG. 3 (C) shows image data obtained as a result of suppressing the random noise component by the image processing apparatus 2. The video signal is stored in a frame memory and N frames are added to each other to thereby obtain an inter-frame interval. While the signal component that does not change in increases, the random noise component has no correlation between frames, so the average amplitude is The random noise component is suppressed by utilizing the attenuation to. Although N frame memories may be used, the same effect can be obtained by using one frame memory in a cyclic structure.

Further, in FIG. 3 (d), after the inspection object 6 is slightly moved,
The result of suppressing the above-mentioned random noise component is shown.

FIG. 3 (e) shows the image data as a result of subtracting the image data shown in FIG. 3 (d) from the image data shown in FIG. 3 (C), and FIG. 3 (c). , (D), which are components included in the image data, such as shading components and fixed noise components, which do not change due to minute movements of the inspection object 6 are erased, and components due to changes in the aperture ratio of the inspection object 6 are extracted. It shows that. In addition, the inspection object 6
The amount of movement of the minute movement depends on the state of the spot to be detected, but the amount of movement of the inspection object 6 also increases with an increase in the period in which the spot changes, and in this embodiment, the number of pixels in the frame memory increases. The converted value is in the range of 2 to 20 pixels.

The above processing is performed based on the image data obtained by slightly moving the inspected body 6 once in the one-dimensional direction and capturing images before and after that.
It shows a method of extracting a spot component.

FIG. 4 schematically shows the moving positions for explaining the moving method when the inspected body 6 is slightly moved. In FIG. 4, P ₀ to P ₈ correspond to the imaging positions of the inspected body 6. is doing. For example, the positions imaged by the above movement are P ₀ and P _{1 in} the figure.
Equivalent to. However, when processing is performed only with image data obtained from these two locations, spots caused by a change in the aperture ratio in the H direction in the figure are detected, but changes in other directions, such as the V direction, occur. Although there is a defect that anisotropy of detection sensitivity is caused in that a large spot with a small change in the H direction is not detected, for example, it was obtained by imaging at P ₀ , P ₁ , P ₂ , P ₃ , P ₄ in the same figure. Based on the image data, P ₀ and P ₁ , P ₀ and P ₂ , P ₀ , and P ₃ , P ₀ and P ₄
The anisotropy is eliminated by performing the processing described in FIG. 3E for each of the combinations. Also the fourth
Even if the detection process is performed based on the image data obtained by subtracting the total of the image data at each position arranged on the circumference from the image data at the center position P ₀ as shown in P ₁ to P _{8 in} the figure,
The above-mentioned anisotropy can be avoided, and a value that approximates the curvature of the brightness distribution that is visually responsive can be obtained, resulting in an inspection result close to a visual inspection.

By the way, the above description has been centered on the method of moving the inspection object 6, but the image data obtained by imaging at each position includes the random noise component, the shading component, and the fixed noise component. Next, a method for reducing these unnecessary components will be described. First, the random noise component can be suppressed by adding a plurality of frames at each moving position shown in FIG. 4 in the same manner as described above. Further, the shading component and the fixed noise component can be erased by calculating the image data so that the total number of frames of the image data becomes "0". For example 32 frames at a position P ₀ of FIG. 4, performs addition of 8 frames in P _1, is subtracted image data both images at P ₀ image data from the quadruple image data in P ₁ Data is 32
Since the image data corresponding to the frames is added, the total number of frames of both image data becomes “0”, and the shading component and the fixed noise component are erased. As another example, when screen addition for 4 frames is performed at each position of P ₁ to P _{8 in} FIG.
Since the addition result of the image data of P ₁ to P ₈ corresponds to the addition result of the image data of 32 frames, if subtracting from the addition result of the image data of 32 frames at the position of P ₀ , it is the same as the above example The shading component and the fixed noise component are eliminated, and the second derivative of the brightness distribution is obtained.

Furthermore, when smoothing processing is applied to image data with reduced shading components and fixed noise components by the above processing, random noise components are further reduced, making it possible to detect extremely small spot components. Further, it is also possible to suppress a change in image data due to a minute defect or a spot having a short period and select and detect a spot component having a long period.

The above-described processing is performed on the entire area of the inspection object 6, and the obtained image data of the entire inspection object 6 is displayed on the TV monitor to cause spots on the inspection object 6. Since the change in the brightness level appears only in the portion where the spots exist, the spots can be easily recognized even by visual inspection.

Next, a method of automatically determining whether the product is good or bad based on the image data subjected to the above image processing will be described.

FIGS. 3 (f) and 3 (g) show image data obtained by removing the DC component from the image data of FIG. 3 (e), and FIG. 3 (f) shows the image shown in FIG. 3 (e). Image data obtained by subtracting the average value of image data in a sufficiently wide area of data,
FIG. 3 (g) shows image data obtained by differentiating the image data of FIG. 3 (e). Then, the spots are automatically detected by setting a predetermined threshold value (S ₁ , S ₂ ) according to the nature of the spots to be detected.

In FIG. 3 (h), thresholds S ₁ and S ₂ are set for the image data of FIG. 3 (g), and those exceeding the threshold are shown as “1”, and those not exceeding the threshold are shown as “0”. It is binary data.
To further explain the relationship between the threshold value and spots, first, in the case of detecting an isolated spot as shown by C in FIG. A threshold value such as S ₁ may be set for the image data. As another example, when a synchronously varying spot as shown by B in FIG. 3 (a) is detected and the product is defective, the images of FIGS. 3 (f) and (g) are displayed. After setting a threshold value such as S _{2 for} the data, it is converted into the binarized data as shown in FIG. 3 (h), and the predetermined value is converted from the binarized data as shown in FIG. 3 (i). Convert the number of spots in the area, that is, density data,
To said seal of data, by performing a comparison with a predetermined threshold value S _3, detect only plaques that changes periodically, it is possible to good or bad determination of product from the result, close to the visual inspection The test result is obtained.

Next, an example of automatically inspecting spots caused by the periodical change of the aperture ratio of the object 6 to be inspected by the above-mentioned embodiment will be shown. The unit pattern 11 of the inspection object 6 has a diameter of 100 μm and an arrangement pitch of 300 μm, and a periodic pattern having an arrangement as shown in FIG. 2 is illuminated with transmitted light as shown in FIG. If the TV camera 1 was set to 600 mm × 450 mm as the entire inspection area, the total of the imaging time and the image processing time was about 5 seconds, the change in brightness was 0.5%, and the change in aperture ratio was close to the detection limit of visual inspection. We were able to detect and accurately judge whether the product was good or bad.

As the image pickup method and the illumination method shown in the present invention, any method can be used as long as a video signal including spot information can be obtained. For example, an imaging tube or a solid-state imaging device (area sensor or line sensor) was used as the imaging method.
TV camera, imaging device using image detector and flying spot tube, X-ray TV imaging device, electron microscope imaging device, etc., and illumination methods include transmitted light illumination, transmitted dark field illumination, reflected dark field illumination, specular reflection A method using lighting or the like can be mentioned. Furthermore, the property of the illumination light may be selected so that the signal-to-noise ratio of the signal at the time of imaging the spot for which coherence, spectral characteristics, polarization, etc., are all to be detected is high.

〔The invention's effect〕

As described above, according to the present invention, delicate spots of various industrial products having a periodic pattern are photographed in a photographing field of view in which a unit pattern forming the periodic pattern is not resolved, and the image data is subjected to image processing. By extracting the spot component by doing so, the spot component that is difficult to visually recognize can be easily captured by displaying the spot component on the TV monitor, and the nature of the spot is determined, and the product quality is determined based on the result. -It became possible to automatically judge defects, and significant effects such as improvement of inspection accuracy, reliability and production efficiency were observed.

[Brief description of drawings]

FIG. 1 is a block diagram showing an example of an apparatus for carrying out the method according to the present invention, FIG. 2 is an explanatory diagram showing an example of a periodic pattern, FIGS. 3 (a) to (i) and FIG. FIG. 3 is a waveform diagram and an explanatory diagram showing an image processing means of the method according to the present invention and its effect. 1 …… TV camera 2 …… Image processing device 3 …… Control unit 4 …… TV monitor 5 …… Driving mechanism 6 …… Inspection object 7 …… Diffusion plate 8 …… Incandescent lamp 9 …… DC power supply 10 …… XY stage 11 …… Unit pattern

Claims (6)

[Claims] 1. A method for inspecting spots caused by irregularities in the uniformity of optical properties in a periodic pattern composed of a repetitive array of unit patterns, the periodic pattern is imaged, decomposed into pixels, and imaged in frame units. A means for obtaining data and a means for moving the periodic pattern to a predetermined position are provided, and the uniformity of the periodic pattern is inspected based on image data obtained by moving the periodic pattern to a plurality of positions. A method for inspecting spots of a periodic pattern, which is characterized in that 2. The periodic pattern spot inspection method according to claim 1, wherein the image data is image data obtained by adding image data of a plurality of frames. 3. The image data is image data obtained by processing so that the total number of frames of each image data obtained at each position of the periodic pattern becomes "0". A method for inspecting spots of a periodic pattern according to claim 1 or 2. 4. Image data obtained by moving the periodic pattern to a plurality of positions and processing the image data so that the image data has a value according to the curvature of the distribution of optical properties in the periodic pattern. Claims characterized in that the first
A method for inspecting spots of a periodic pattern according to item 2, item 2 or item 3. 5. The image data obtained by moving and processing the above-mentioned image data so as to have a distribution of differences in optical properties between two points having a predetermined positional relationship of the periodic pattern. The periodic pattern unevenness inspection method according to claim 1, 2, or 3. 6. The process described above is a process capable of recognizing a portion where a density of pixels in which a change in optical property of the periodic pattern is a predetermined level or more as a predetermined value or more, as a defect. Claims 1 to 3,
The periodic pattern spot inspection method according to the third, fourth, or fifth aspect.