JPH05256791A - Inspecting apparatus of defect of board - Google Patents

Abstract

PURPOSE:To inspect the defects of the surface of a board accurately. CONSTITUTION:A prescribed lightness for separating all picture elements of a pickup image of the surface 1a of a board into picture elements showing a pattern 2 and ones showing a background 3 is determined beforehand. Then, the image of the surface 1a of the board being an object of inspection is picked up and the picture elements of the prescribed lightness or above are removed from this image picked up. It is judged that the lightness of the picture elements left after the removal is outside the range of the aforesaid determined lightness. Then, it is determined that the picture elements judged to be outside the range of the lightness represent a defective part of the pattern 2. The defective part of the background 3 is determined in the same way.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention inspects defects of a pattern having a predetermined shape on a substrate such as a printed wiring pattern on a printed wiring board of an electronic circuit and an electrode pattern on a liquid crystal substrate, and defects on the surface of the substrate other than the pattern. The present invention relates to a device for automatically performing.

[0002]

2. Description of the Related Art Wiring patterns in electronic circuits and ITO (transparent electrode) patterns in liquid crystal substrates are "thin" or "thick" due to defects in the process of attaching a conductive material to the substrate, the etching process, etc. "Break",
May cause defects such as "shorts". In addition, the surface of the substrate other than the pattern (hereinafter referred to as “background”) may also have the same defect due to the above-mentioned problem. Therefore, it is indispensable for product quality assurance to inspect the pattern and background defects, determine whether the substrate is a good product or a defective product, and ship only the good product. At present, the inspection of such defects is performed visually by a person looking through a microscope, and automation is in the development stage.

On the other hand, with the recent development of pattern forming technology, there is a further tendency toward finer patterns.

[0004]

However, a human inspection is apt to make an error due to the limit of concentration, etc., and it is a human judgment, so that there is a difficulty in accuracy. In addition, the judgment also varies from worker to worker. Further, the work speed is not sufficient, and there is a problem that the cost for the work is high. Moreover, these problems have become more serious in response to the recent trend of pattern miniaturization.

The present invention has been made in view of the above circumstances, and is capable of coping with the tendency of pattern miniaturization and inspecting defects of the substrate accurately and promptly and at low cost. Its purpose is to provide.

[0006]

Therefore, according to the present invention,
By capturing an image of the substrate surface on which a pattern of a predetermined shape is formed on the substrate, the pattern of the substrate surface is taken as a dark image, the background excluding the pattern is captured as a bright image, and the pattern is obtained based on these bright and dark images. In a defect inspection apparatus for a substrate configured to inspect defects, in advance, while determining a predetermined lightness for separating all pixels of a captured image of the substrate surface into pixels showing the pattern and pixels showing the background, Reference setting means for determining a lightness range of pixels showing a pattern, and pixels having lightness higher than or equal to the predetermined lightness are removed from an image obtained by imaging the surface of the substrate to be inspected, and the removed residue Determination means for determining that the lightness of the pixel is outside the lightness range, and the pixel judged by the judgment means to be outside the lightness range is the pattern. It is determined to be a defective portion of the.

With the same configuration, the defective portion of the background is also determined.

[0008]

The predetermined brightness for separating all the pixels of the imaged image on the surface of the substrate into the pixels indicating the pattern and the pixels indicating the background is determined in advance, and the brightness range of the pixels indicating the pattern is also determined. After that, the surface of the substrate to be inspected is imaged, and pixels having a predetermined brightness or higher are removed from the captured image. Then, it is determined that the brightness of the remaining removed pixels is outside the brightness range determined above. Then, the pixel determined to be outside the brightness range is determined to be the defective portion of the pattern. The defective portion of the background is similarly determined.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a substrate defect inspection apparatus according to the present invention will be described below with reference to the drawings.

FIG. 1 shows the configuration of the apparatus of the embodiment.

As shown in the figure, on the XY table 4, a substrate 1 (printed wiring board or the like) to be inspected for defects is provided.
Are mounted so that the substrate 1 can be arbitrarily moved in the XY two-dimensional plane. A pattern 2 having a predetermined shape is formed on the surface 1a of the substrate 1, and the substrate surface 1a is composed of the pattern 2 and the background 3 excluding the pattern 2. The CCD camera 5 is for enlarging and imaging the substrate surface 1a through the microscope 6, and the pattern 2 on the substrate surface 1a is used as a dark image, that is, each pixel of the pattern 2 has a low gradation level of brightness. To be caught. Further, the camera 5 uses the portion of the substrate surface 1a corresponding to the background 3 as a bright image, that is, the background 3
Each of the pixels is captured as a pixel having a high brightness gradation level. The image signal is output to the image processing device 7. Although not shown, the image processing device 7 includes an A / D converter that digitizes an image signal, a frame memory that stores the digitized image, a CPU that performs analysis processing of the stored image, a peripheral logic circuit, and a memory. A D / A converter for outputting images to the monitor TV8 and a signal superimposing unit for superimposing I / O, charts, menus, etc. on the monitor TV8 for communicating with external devices (such as the XY table 4). Etc. A monitor TV and a trackball 9 are connected to the image processing device 7. An interactive process is performed on the screen of the monitor TV8 by operating the trackball 9 by an operator, and an interactive process is performed on the screen of the monitor TV8. The result of is displayed.

In such a structure, when the substrate 1 is small and the captured image of the entire surface 1a can be obtained at one time, the XY moving mechanism in the XY table 4 may be omitted. It is possible. Further, the microscope 6 may be omitted in the substrate 1 in which the substrate surface 1a does not need to be enlarged.

FIG. 2 exemplifies a captured image of the substrate surface 1a displayed on the screen of the monitor TV8. As shown in the figure, the pattern 2 is displayed as a dark image of black color, and the background 3 is displayed as a bright image of white color. It is sufficient that the pattern 2 and the background 3 can be imaged with good contrast, and depending on how the illumination is applied and the material of the substrate 1, the pattern 2 may be imaged and displayed as a white bright image and the background 3 as a black dark image. Of course it is possible.

Next, the processing executed by the CPU in the image processing apparatus 7 will be described with reference to the flowcharts shown in FIGS.

Pretreatment (FIG. 5) The operator first selects a defect-free sample as the substrate 1 and images it. As a result, the captured image is displayed on the screen of the TV 8 (step 101). Then, the "search area setting" menu superimposed on the TV 8 screen is selected by operating the trackball 9 (see FIG. 2). Then, a rectangular box cursor (not shown) appears on the screen. By adjusting the size and position of the cursor by operating the trackball 9, a desired inspection target area in the captured image on the screen is set. Then, the frequency distribution of the brightness of each pixel in the area surrounded by the cursor is displayed on the screen as a histogram as shown in FIG.

FIG. 3 shows the brightness M of each pixel in the search area.
Is a horizontal axis and the corresponding pixel number N is a vertical axis. As shown in the figure, the lightness distribution has a mountain 10 located in a portion having a relatively low lightness corresponding to the pattern 2,
It is separated into a mountain 11 located in a portion having a relatively high brightness corresponding to the background 3. Then, the "background separation" menu on the screen is selected (see FIG. 2), and the threshold value of lightness for separating the mountain 10 and the mountain 11 is determined. That is,
By visually observing the lightness distribution on the screen and operating the trackball 9, an intermediate position M ′ between the two ridges 10 and 11 is determined as a threshold value.

Here, the threshold value M'can be selected from both a fixed threshold value and a relative threshold value that follows variations in ambient brightness. The fixed threshold is to store the threshold as an absolute value of brightness. On the other hand, the relative threshold is not the absolute value of the brightness but the peaks 10 and 11
It stores the percentage of the area of the threshold value set in the position. Therefore, according to the relative threshold value, even if the captured image is different and the distribution is different, the threshold value is set according to the change of the distribution.

When the threshold M'is determined in this way, a pixel having a lightness higher than the threshold M'is regarded as a "pixel corresponding to the background 3" and a lightness higher than the threshold M '. The low pixels are respectively stored as “pixels corresponding to pattern 2”. As will be described later, when the defect of the pattern 2 is inspected, the "memory pixel corresponding to the background 3" is masked, and when the background 3 is inspected, the "memory pixel corresponding to the pattern 2" is masked ( Step 102).

Next, the "non-defective pattern range" menu is selected (see FIG. 2), and the upper and lower limits of the brightness of the pixels showing the pattern 2 having no defect are determined. That is, as shown in FIG. 3, if there is no defect in the pattern 2, each pixel is distributed in the mountain 10 within a predetermined lightness range having a lower limit of M1 and an upper limit of M2. M2 is designated by the trackball 9. Therefore, as will be described later, if there is a pixel that is out of the lightness range for a pixel that shows pattern 2, the pixel portion is determined to be defective. For example, the upper limit value M of the brightness range
If bright pixels exceeding 2 are mixed, the corresponding pixel portion is judged to be “cut” in pattern 2. Conversely, if there is a dark pixel lower than the lower limit M1 of the lightness range, the corresponding pixel portion is detected. Is determined to be "foreign matter mixed" (step 103).

However, if this determination is uniquely made for all the pixels in the image pickup area including the background 3, all the pixels corresponding to the background 3 are determined to be defective. Therefore, a pixel portion that is known to be the background 3 in advance is subjected to a process of masking the background when inspecting the pattern 2 in order to exclude it from the determination target in advance. In this masking process, the pixels to be masked are identified by the above step 1.
Pixels corresponding to lightness higher than the threshold value M ′ set in 02 may be used. However, this may lead to too stringent inspection standards. That is, the pattern width usually has some variations depending on the sample, and if only the pixels corresponding to the background 3 are used as a mask, if the pattern becomes thinner than the non-defective sample, the background portion outside the mask is detected as a defect. Will be done. Therefore, it is necessary to slightly inflate the background mask by a predetermined number of pixels, and appropriate settings are made. In the expansion process, an expansion / contraction filter using a so-called morphology method is used.
The number of pixels in the background mask to be expanded is determined by how much variation in the pattern width is allowed (step 104).

Then, similarly to step 103, the lightness range M3 to M4 indicating the background 3 having no defect is determined as shown in FIG. As a result, it is determined that a dark pixel portion lower than the lower limit value M3 of the brightness range is a "pattern protrusion, short circuit" and a bright pixel portion higher than the upper limit value M4 is a defect of "foreign matter inclusion, pinhole". (Step 105). Then, in the same manner as in step 104, the setting process of the expansion of the mask of the pattern when inspecting the defect of the background 3 is performed, and the accurate inspection is performed even if there is a slight pattern thickening (step 106). ).

If the threshold M'as a reference for inspection, the lightness ranges M1 to M2, M3 to M4 as the reference for inspection, and the expansion setting are made on the basis of the non-defective non-defective sample as described above, it is required. Tests that apply the criteria are performed. That is, the “execution test” menu is selected on the screen (see FIG. 2), and some samples are actually imaged. Then, a lightness distribution similar to that shown in FIG. 3 is obtained, and the pixels in the search area are
The pixels corresponding to pattern 2 and the pixels corresponding to background 3 are separated. The inspection of the pattern 2 is executed by masking not only the pixels having the lightness of the threshold value M ′ or more but also the pixel expanded by the pixel set in the above step 104. Lightness range M1 for the remaining pixels after being masked
It is determined whether or not it is within M2. As a result, the pixels determined to be outside the brightness range are displayed blinking on the screen.
Here, the limitation may be given depending on the number of pixels. For example, if the number of pixels determined to be outside the brightness range is 1,
If it is a pixel, it can be determined that there is no defect, and if there are two or more pixels adjacent to each other that are outside the brightness range, it can be determined that there is a defect for the first time.

Similarly, the defect of the background 3 is determined using some samples (step 107).

As a result, if the defect inspection is satisfactory, it is judged that the inspection standard set in steps 102 to 106 is appropriate (YES in step 108), and the next automatic operation ( Although the process proceeds to step 109), if the defect inspection is far from the actual condition and is unsatisfactory, it is determined that the inspection standard set in steps 102 to 106 is not appropriate (step 10).
No. 8), the processing of steps 102 to 108 is repeated again to reset the inspection standard. When the appropriate inspection standard such as the threshold value M'is set in this way, the automatic operation shown in FIG. 6 is performed in order to inspect the defect of the substrate 1 which is not always a good product.

That is, the "automatic operation" menu on the screen is selected (see FIG. 2), and the surface 1a of the substrate 1 is imaged. In this case, the substrate 1 is positioned at a predetermined position by the XY table 4, and a predetermined region of the front surface 1a is captured as a captured image and displayed on the screen of the TV 8. In the automatic driving, the image processing in the image processing device 7 and the control of the moving mechanism of the XY table 4 are executed in synchronization with each other (step 201).

Thus, a lightness distribution similar to that shown in FIG. 3 is obtained as shown in FIG. 4 (a). here,
The mountain 12 is a mountain showing the pattern 2, and the mountain 13 is a mountain showing the background 3. Both are separated by a threshold M ',
Pixels having a brightness equal to or lower than the threshold M ′ and pixels having a brightness higher than the threshold M ′ are stored as a pixel corresponding to the pattern 2 and a pixel corresponding to the background 3 (step 202). Then, in FIG. 4A, the threshold value M
A pixel for inspecting the background 3 (hereinafter referred to as a “pattern mask”) is obtained by adding only the expanded pixels set in the above step 106 to pixels having lightness of ′ or less, that is, pixels corresponding to the pattern 2. ), And store.

Similarly, the pattern 2 is inspected by adding only the dilated pixels set in the above step 104 to the pixels having the lightness higher than the threshold M ', that is, the pixels corresponding to the background 3. Set and stored as a mask (hereinafter referred to as "background mask") (step 20).
3).

Next, the inspection of the defect of the background 3 is executed. In this case, of all the pixels of the captured image, the pattern mask portion set and stored in step 203 is removed. Then, it is determined whether or not the brightness of the remaining removed pixels is within the brightness range M3 to M4.

As a result, if it is judged that two or more pixels having a lightness smaller than the lower limit M3 of the lightness range are present adjacent to each other as shown by 13a in FIG. 4A, the corresponding pixel portion is detected. Is judged to be a defect such as "pattern protruding or short circuit" as shown at 3a in FIG. Further, as shown in 13b of FIG. 4A, if it is determined that two or more pixels having a lightness larger than the upper limit M4 of the lightness range are present adjacent to each other, the corresponding pixel portion is As shown in 3b of (b), it is determined that there is a defect such as "foreign matter inclusion or pinhole" (step 20).
4).

In the same manner, the inspection of the defect of pattern 2 is executed. In this case, of all the pixels of the captured image, the background mask portion set and stored in step 203 is removed. Then, it is judged whether or not the brightness of the remaining removed pixels is within the brightness range M1 to M2.

As a result, if it is judged that two or more pixels having a lightness smaller than the lower limit M1 of the lightness range are present adjacent to each other as shown by 12a in FIG. 4A, the corresponding pixel portion is detected. Indicates that “foreign matter is mixed” as shown in 2a of FIG.
It is determined to be a defect such as. In addition, 12 in FIG.
If it is determined that two or more pixels having a lightness larger than the upper limit value M2 of the lightness range are present adjacent to each other as shown in b, the corresponding pixel portion is as shown in 2b of FIG. Is judged to be a defect such as "pattern break" (step 205).

When it is determined that there is a defect as described above, the corresponding pixel is blinked and displayed on the screen. This allows the operator to recognize the inspection result (step 206).

Next, the substrate surface 1 which has not been inspected
The XY table 4 is drive-controlled so as to image the other area of a. As a result, the above step 20 is performed for other areas.
The processes 1 to 207 are executed again, and the defects on the substrate surface 1a are similarly determined. Thereafter, similar processing is sequentially performed for each area, and the entire surface 1a of the substrate is inspected. As a result, it is determined whether the substrate 1 is a good product or a defective product according to the presence or absence of defects on the entire surface 1a. Thereafter, the substrate 1 to be inspected next is carried in on the table 4, and the same processing is executed.

Instead of outputting the inspection result of defects for each region of the surface 1a, it is also possible to output the entire inspection result when the inspection is completed for the entire substrate surface 1a. Further, in order to speed up the process, it is also possible to implement that the substrate 1 is determined to be a defective product at the stage when it is determined that there is a defect in a predetermined area, and the subsequent inspection is omitted. In addition, the output form of the inspection result is arbitrary, and for example, the image processing device 7 is connected by serial communication.
It is also possible to send the result from the host computer to the host computer, and it is also possible to send the data to the peripheral sequencer and actuator by digital I / O. Depending on the case, it is also possible to actually spray ink on the location of the defect on the substrate 1 to make a mark.

In the embodiment, the presence / absence of a defect is determined according to whether or not the brightness of a pixel is within a predetermined range. However, the difference value (differential value) of the brightness of adjacent pixels is a predetermined value. This may be determined depending on whether it is the following.

That is, in the pattern 2 having no defect, even if the brightness of each pixel showing the pattern 2 varies, the fluctuation is slight, and the difference value (differential value) of the brightness between adjacent pixels is small. ) Is in the range of a predetermined value or less. On the contrary, if the pattern 2 has a defect such as a scratch, the difference value of the brightness between the adjacent pixels on the boundary becomes larger than the above-mentioned predetermined value. Therefore, the predetermined value can be used as the inspection standard for determining the defect of the pattern 2. Similarly, the defect of the background 3 can be determined by determining whether or not the difference value of the brightness of the adjacent pixels is less than or equal to a predetermined value.

Specifically, instead of determining whether or not the pixels of pattern 2 are within the brightness range M1 to M2, it is determined whether or not the difference in brightness of the adjacent pixels of pattern 2 is less than or equal to a predetermined value. Good. Further, the pixels of the background 3 are in the lightness range M3 to M
Instead of determining whether or not the difference is within 4, it may be determined whether or not the difference in brightness between adjacent pixels of the background 3 is less than or equal to a predetermined value. Here, the Sobel calculation method can be applied to the image differential processing.

In the embodiment, when the substrate is inspected for defects, both the pattern and the background are inspected. However, when the background inspection is not required, only the pattern defect may be inspected. It is possible. Similarly, only background defects may be inspected.

[0039]

As described above, according to the present invention, a predetermined inspection standard is set and pixels other than the inspection target are removed so that the inspection of defects is automatically performed. Even if such a pattern is formed, it becomes possible to accurately and quickly inspect the substrate for defects and at low cost.

[Brief description of drawings]

FIG. 1 is a perspective view showing a configuration of an embodiment of a substrate defect inspection apparatus according to the present invention.

FIG. 2 is a diagram showing a display example of a screen of the monitor TV shown in FIG.

FIG. 3 is a luminosity frequency distribution diagram showing a relationship between luminosity and the number of pixels when an image of a substrate having no defect is picked up.

FIG. 4A is a luminosity distribution chart showing the relationship between luminosity and the number of pixels when a defective substrate is imaged, and FIG. 4B is a diagram showing defects on the surface of the substrate. ..

FIG. 5 is a diagram showing a procedure of preprocessing of the processing executed by the image processing apparatus shown in FIG.

6 is a diagram showing a procedure at the time of automatic operation among the processes executed by the image processing apparatus shown in FIG.

[Explanation of symbols]

 1 substrate 2 pattern 3 background 5 CCD camera 7 image processing device

Claims (6)

[Claims] 1. An image of a surface of a substrate on which a pattern of a predetermined shape is formed is captured to capture the pattern of the substrate surface as a dark image and the background excluding the pattern as a bright image. In a substrate defect inspection device configured to inspect a defect in the pattern based on an image, a predetermined lightness that separates all pixels of a captured image of the substrate surface into pixels indicating the pattern and pixels indicating the background is set in advance. Reference setting means for determining the brightness range of the pixels showing the pattern, and removing the pixels having the brightness of the predetermined brightness or more from the image obtained by imaging the surface of the substrate to be inspected, A determination unit that determines that the brightness of the remaining pixels that have been removed is outside the brightness range. Defect inspection apparatus of the substrate so as to determine that the defective portion of the pattern. 2. The reference setting means obtains a lightness frequency distribution for all pixels of an image obtained by imaging a surface of a substrate having no defect, and based on the lightness frequency distribution, all pixels of the captured image are subjected to the pattern. 2. The defect of the substrate according to claim 1, wherein a predetermined lightness for separating the pixel indicating the pixel and the pixel indicating the background is determined, and the lightness range of the pixel indicating the pattern is determined based on the lightness frequency distribution. Inspection equipment. 3. An image of a surface of a substrate on which a pattern of a predetermined shape is formed is imaged to capture the pattern of the substrate surface as a dark image and the background excluding the pattern as a bright image. In a substrate defect inspection device configured to inspect the background for defects based on an image, a predetermined brightness that separates all pixels of a captured image of the substrate surface into pixels indicating the pattern and pixels indicating the background is set in advance. Along with the determination, the reference setting means for determining the brightness range of the pixel indicating the background, and removing the pixels having the brightness less than or equal to the predetermined brightness from the image obtained by imaging the surface of the substrate to be inspected, A pixel which is judged to be outside the lightness range by the judgment means. Defect inspection apparatus of the substrate so as to determine that the defective portion of the scene. 4. The reference setting means obtains a frequency distribution of lightness for all pixels of an image obtained by imaging the surface of a substrate having no defect, and based on the lightness frequency distribution, all pixels of the captured image are subjected to the pattern. 4. The defect of the substrate according to claim 3, wherein a predetermined lightness for separating the pixel indicating the background and the pixel indicating the background is determined, and the lightness range of the pixel indicating the background is determined based on the lightness frequency distribution. Inspection equipment. 5. An image of a surface of a substrate on which a pattern of a predetermined shape is formed is captured to capture the pattern of the substrate surface as a dark image and the background excluding the pattern as a bright image. In a substrate defect inspection device configured to inspect a defect in the pattern based on an image, a predetermined lightness that separates all pixels of a captured image of the substrate surface into pixels indicating the pattern and pixels indicating the background is set in advance. Reference setting means for determining, and removing from the image obtained by imaging the surface of the substrate to be inspected, the pixels having the brightness of the predetermined brightness or higher, and the brightness of each adjacent pixel of the removed remaining pixels. Find each difference,
A determination unit that determines that the obtained brightness difference is equal to or greater than a predetermined value, and a pixel that is determined to be greater than or equal to the predetermined value by the determination unit is determined to be a defective portion of the pattern. Defect inspection system for printed circuit boards. 6. An image of a surface of a substrate on which a pattern of a predetermined shape is formed is imaged to capture the pattern of the substrate surface as a dark image and the background excluding the pattern as a bright image. In a substrate defect inspection device configured to inspect the background for defects based on an image, a predetermined brightness that separates all pixels of a captured image of the substrate surface into pixels indicating the pattern and pixels indicating the background is set in advance. Reference setting means for determining, and removing from the image obtained by imaging the surface of the substrate to be inspected, the pixels having a brightness less than or equal to the predetermined brightness, and the brightness of each adjacent pixel of the remaining remaining pixels. Find each difference,
A determination unit that determines whether the obtained brightness difference is equal to or greater than a predetermined value, and determines that a pixel determined by the determination unit to be equal to or greater than the predetermined value is a defective portion of the background. Defect inspection system for printed circuit boards.