JP2003075366A - Device and method for inspecting substrate - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve a problem that an erroneous detection rate for a tear-off condition of an ACF is difficult to be reduced while narrowing a pitch of a lead and saving a space thereof, in tear-off inspection in the ACF using a transmission type sensor or a reflection type sensor. SOLUTION: A substrate 1 bonded with the ACF 3 is image-picked up by a video camera 6, and an obtained image is binalized in an image processing part 7. The number of white picture elements or black picture elements of a binalized image therein is counted by a counting part 10, and a counted value therein is compared with a threshold value held preliminarily in a threshold value holding part 12 by a comparison circuit 11, so as to determined normality of an ACF-bonded condition or abnormality in which the ACF is torn off.

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 substrate and an inspecting method therefor, and more particularly to an apparatus and method suitable for inspecting the attachment state of ACF.

[0002]

2. Description of the Related Art TCP (Tape) is formed on a glass substrate, which is conventionally known as a process for manufacturing a liquid crystal display panel or the like.
In the OLB process of pressure-bonding electronic parts such as Carrier Package, etc., an anisotropic conductive film (Aniso
There is a step of attaching a tropic conductive film (hereinafter referred to as "ACF") to the lead portion of the glass substrate.

FIG. 7 is a schematic configuration diagram showing a conventional ACF attachment state inspection device (hereinafter referred to as an ACF inspection device). In FIG. 7, after the ACF 3 is attached to the lead portion of the glass substrate 1, the attachment state of the ACF 3 is inspected.

The glass substrate 1 to which the ACF 3 is attached in the previous step is conveyed by a cell stage (not shown),
It moves to the inspection position (Fig. 7) and the transmission type sensor 2
The turning state of CF3 is inspected.

The transmissive sensor 2 comprises a light emitting portion 21 located above the glass substrate 1 and a light receiving portion 22 located below the glass substrate 1, which are arranged opposite to each other with the glass substrate 1 interposed therebetween. Therefore, when the attached ACF 3 is turned over at the end of the glass substrate 1, the light 50 emitted from the light emitting unit 21 is emitted from the glass substrate 1 as shown in FIG. 7A.
Since the light passes through between the leads 4 and is received by the light receiving unit 22, the output of the light receiving unit 22 becomes a high level indicating that the ACF 3 is not detected.

On the other hand, when the ACF 3 is normally attached to the glass substrate 1, the light 50 emitted from the light emitting section 21 is emitted.
Is blocked by the ACF 3 and does not reach or is attenuated in the light receiving section 22 below the glass substrate 1, and the output of the light receiving section 22 becomes a low level indicating the detection of the ACF 3. afterwards,
When the transmissive sensor 2 detects the ACF 3, the cell stage moves to the next process, and the transmissive sensor 2 moves to the ACF 3
If is not detected, an error is output.

FIG. 8 is a schematic configuration diagram showing another conventional ACF inspection apparatus, and the same parts as those in FIG. 7 are designated by the same reference numerals. The ACF inspection apparatus shown in this example uses a reflective sensor 5 for detecting the ACF 3, and the reflective sensor 5 is located above the glass substrate 1. The reflection type sensor 5 has a light emitting unit and a light receiving unit (not shown) arranged in one place, and the light receiving unit receives the reflected light of the light 50 emitted from the light emitting unit and detects the presence or absence of the ACF 3.

Therefore, when the attached ACF 3 is turned over at the end of the glass substrate 1, as shown in FIG. 8A, the light 50 emitted from the light emitting portion is led by the lead of the glass substrate 1 (for example, gold plating). The reflected light is received by the light receiving unit, and the output of the light receiving unit is AC.
It becomes a high level indicating the non-detection of F3.

On the other hand, as shown in FIG. 8B, when the ACF 3 is normally attached to the glass substrate 1, the light 50 emitted from the light emitting portion is absorbed by the ACF 3 and is hardly reflected. Almost no reflected light is received by the portion, and the output of the light receiving portion becomes a low level indicating the detection of ACF3. After that, when the reflective sensor 5 detects the ACF 3, the cell stage shifts to the next process, and when the reflective sensor 5 does not detect the ACF 3, an error is output.

[0010]

By the way, in the conventional ACF inspection apparatus as described above, the determination area is wide because of detection by the spot light regardless of whether the transmission type sensor 2 or the reflection type sensor 5 is used. Can't take, ACF3
There is a high rate of false detection of curling, and there is a problem that the ACF 3 curls up and the glass substrate 1 that is not properly attached moves to the next process, which deteriorates the yield of products.

Therefore, there is also considered a method in which a video camera is used to pick up the attached state of the ACF 3, the image is captured, and the turned-up state of the ACF 3 is detected by using a multi-value matching method or an edging method. However, in this method, even if the determination area can be wide, when the ACF attachment state (position) varies, a difference may occur between the captured image and the reference image. In such a case, the image determination process using the conventional multi-value matching method or the edging method still cannot reduce the false detection rate of the ACF 3 in the turned-up state. Therefore, the ACF inspection device using the transmissive sensor 2 and the reflective sensor 5 is still used at present.

On the other hand, recently, due to the narrowing of the pitch of the leads 4, the range through which the sensor light emitted from the transmissive sensor 2 is transmitted is narrowed, and the space saving of the device saves the transmissive sensor 2 and the reflective sensor 5 itself. The space for installing the is disappearing. Therefore, the development of an ACF inspection device that can reduce the false detection rate of the ACF3 turn-up state while saving space by using an existing video camera used for positioning etc. for the ACF inspection. Has been requested.

An object of the present invention is to detect a curl of an inspection object by using an image picked up by an image pickup means,
It is an object of the present invention to provide a substrate inspection device and a substrate inspection method that can increase the detection rate and can cope with a narrow lead pitch.

[0014]

In order to achieve the above object, the invention of claim 1 provides an image pickup means for picking up an image of a substrate on which an inspection object is attached, and an image picked up by the image pickup means. Image binarizing means for binarizing each pixel in black and white; and counting means for counting at least one of the number of white pixels or the number of black pixels of the monochrome binarized image binarized by the image binarizing means, It is characterized by further comprising: a determining unit that compares the count value counted by the counting unit with a preset threshold value to determine whether the attachment state of the inspection target object is normal or abnormal. It is a substrate inspection device that does.

According to a second aspect of the present invention, in the first aspect, the range of the black-and-white binary image in which the number of white pixels or the number of black pixels is counted by the counting means is abnormal in the pasted state of the inspection object. It is characterized in that it is a range that is likely to occur.

According to a third aspect of the present invention, an image pickup step for picking up an image of the substrate on which the inspection object is attached, and an image binarization step for binarizing the image picked up by the image pickup step for each pixel in black and white. And a counting step of counting at least one of the number of white pixels or the number of black pixels of the monochrome binary image binarized by the image binarizing step, and the count value counted by the counting step is preset. And a threshold value for determining whether the attachment state of the inspection object is normal or abnormal by comparing the threshold value with the threshold value.

The invention of claim 4 is characterized in that, in the image binarization step of claim 3, the black-and-white binarization level is changed according to the detection condition.

According to a fifth aspect of the present invention, in the counting step of the third or fourth aspect, whether the pixel to be counted is a white pixel or a black pixel is selected according to a detection condition.

[0019]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 shows the AC according to this embodiment.
It is the block diagram which showed the structure of the F inspection apparatus.

The ACF inspection apparatus 100 includes a video camera 6 for picking up an image of the attachment portion of the ACF 3 which is an inspection object,
An image processing unit 7 that divides captured image data obtained from the video camera 6 into a predetermined number of pixels and binarizes each pixel into black and white.
An image memory 8 for storing the binarized image data, a monochrome binarization level setting unit 9 for holding a preset monochrome binarization level, and the binarized image data stored in the image memory 8. A counting unit 10 that counts the number of white pixels or black pixels, and a comparison circuit 11 that determines whether the attached state of the ACF 3 is normal or abnormal from the count value of the pixels of the binarized image data. Of these, the comparison circuit 11 has A
A threshold value holding unit 12 is provided which holds a threshold value that serves as a reference when determining whether the attached state of CF3 is normal or abnormal. In addition, in this example, the image captured by the video camera 6 is input to the monitor 13 such as a CRT and displayed.

Next, A by the above ACF inspection device 100
Specific examples of the CF inspection will be described in Examples 1 to 3. In FIG. 1, the ACF is attached to the lead portion of the glass substrate 1 in the previous step.
It is assumed that the glass substrate 1 to which 3 is attached has moved to the flip-up detection position by a cell stage (not shown).

Example 1 In Example 1, the glass substrate 1
An example in which the reflectance of the lead 4 and the reflectance of the ACF 3 are lower than that of the lead 4 will be described.

The video camera 6 picks up an image of a predetermined area of the glass substrate 1 and outputs the obtained image data to the image processing section 7. At this time, the image captured by the video camera 6 is displayed on the screen of the monitor 13 as shown in FIG. 2, for example.

The image processing unit 7 divides the image data picked up by the video camera 6 into a preset number of pixels, and also divides the pixel data according to the monochrome binarization level held by the monochrome binarization level setting unit 9. Black and white binarization is performed for each image, and the obtained black and white binarized image is stored in the image memory 8. In this example, the black and white binarization level is set such that the glass substrate 1 is a white pixel and the leads 4 and the ACF 3 are black pixels. What is this black and white binarization level?
It is a gray level that is a reference when judging the brightness of an image, pixels having a brightness higher than a preset gray level are classified as white pixels, and pixels having a brightness lower than the gray level are black pixels. Classify. Therefore, in the first embodiment, the gray level is set such that the glass substrate 1 is a white pixel and the leads 4 and the ACF 3 are black pixels.

Next, the counting section 10 counts the total number of white pixels in the recognition area 30 shown in FIG. 3A of the black-and-white binary image stored in the image memory 8, and this count value is calculated. Output to the comparison circuit 11. Comparison circuit 11
Compares the input count value with a threshold value, which will be described later, held in the threshold value holding unit 12, and if the input count value is less than or equal to the threshold value, the ACF 3 is applied to the glass substrate 1.
Is determined to have been pasted normally, and the determination result is output.

The image in FIG. 3A shows the ACF on the glass substrate 1.
3 is an example of being pasted normally, and the recognition area 3
0 is almost covered by ACF3. Therefore, the number of white pixels is small, the input count value (the number of white pixels) is equal to or less than the threshold value, and it is determined that the ACF 3 is normally attached.

On the other hand, the video camera 6 is used to
When an image of CF3 not normally attached and the end of which is turned up is picked up, the black-and-white binary image stored in the image memory 8 is as shown in FIG. 3B, for example. The recognition area 30 has more white pixels than the example of FIG. This is because the ACF 3 is turned over at this portion and the surface of the white glass substrate 1 is exposed. Therefore, in this case, the count value (the number of white pixels) of the counting unit 10 becomes large and exceeds the threshold value held in the threshold value holding unit 12. Therefore, the comparison circuit 11
Determines that the ACF 3 is not normally attached to the glass substrate 1 and the ACF 3 is turned over and is a defective glass substrate.

A video camera 6 is used to take an image of a sample in which the ACF 3 is normally attached to the glass substrate 1.
By counting the number of white pixels in the obtained black and white binary image,
The count value at that time becomes a normal value. Therefore, a more accurate determination can be made by comparing the normal value and the count value, but in reality, even if the ACF 3 is normally attached, the position may vary. , A value considering the variation is set as the threshold value.

(Embodiment 2) In this embodiment 2, the lead 4 is used.
A column having a lower reflectance than the glass substrate 1 and the ACF 3 will be described. The reflectance of the leads 4 is the glass substrate 1
As compared with the case of ACF3 or ACF3, the case where the lead 4 has a darker color than the glass substrate 1 or ACF3 or the case of a material that easily absorbs light can be considered. In this example, the black and white binarization level set in the black and white binarization level setting unit 9 is set such that the glass substrate 1 and the ACF 3 are white pixels and the leads 4 are black pixels. Has been done. Further, the threshold value held in the threshold value holding unit 12 also corresponds to the set black and white binarization level.

When the black and white binarization level is set as described above, the black and white binarized image stored in the image memory 8 when the ACF 3 is normally attached is, for example, as shown in FIG.
As shown in (a). In this case, since the ACF 3 is attached to the recognition area 30, the number of white pixels increases.
In this example, in order to facilitate the determination, a dummy lead 41 having a wide width that becomes a black pixel when binarized is provided, and the ACF 3 is attached on the dummy lead 41. Recognition area 30 above 41
Is set.

The counting section 10 counts the total number of white pixels in the recognition area 30 shown in FIG. 4 (a). Then, the comparison circuit 11 compares the count value of the number of white pixels with the threshold value held in the threshold value holding unit 12. Figure 4
In the example of (a), the number of white pixels exceeds the threshold value, so
The determination result indicating that the attachment of CF3 is normal is output.

On the other hand, in the case where the ACF 3 is turned over abnormally, the monochrome binarized image stored in the image memory 8 is as shown in FIG. 4 (b). In this case, since the ACF 3 is turned over in the recognition area 30 and is not in this portion, the black dummy lead 41 is exposed and the number of white pixels is reduced.

The counting section 10 counts the total number of white pixels in the recognition area 30 shown in FIG. 4 (b). Then, the comparison circuit 11 compares the count value of the number of white pixels with the threshold value held in the threshold value holding unit 12. Figure 4
In the example of (b), the number of white pixels is less than the threshold value.
The determination result indicating that the attachment of CF3 is abnormal is output.

(Embodiment 3) This embodiment 3 is a PCB (P
Printed Circuit Board) Substrate 1 '
An example in which the ACF 3 is attached to the above will be described, and the case where the lead 4 has a higher reflectance than the PCB substrate 1 ′ and the ACF 3. As the black and white binarization level held in the black and white binarization level setting unit 9, a black and white binarization level is set so that the PCB substrate 1 ′ and the ACF 3 become black pixels and the leads (gold plating) 4 become white pixels. ing. Further, the threshold value held in the threshold value holding unit 12 also corresponds to the set black and white binarization level.

When the glass substrate 1 to be inspected in this example is imaged by the video camera 1, the image displayed on the monitor 13 is as shown in FIG.
As shown in. When the black and white binarization level is set as described above, the black and white binarized image stored in the image memory 8 when the ACF 3 is normally attached is
As shown in FIG. 6A. In this case, the recognition area 3
Since ACF3 is attached to 0, there are many black pixels. In this example as well, in order to facilitate the determination, a dummy lead 42 (gold plating) that is wide and becomes white when binarized is provided, and AC is placed on the dummy lead 42.
F3 is attached and the recognition area 30 is set on the dummy lead 42.

The counting section 10 counts the total number of black pixels in the recognition area 30 shown in FIG. 6 (a). Then, the comparison circuit 11 compares the count value of the number of black pixels with the threshold value held in the threshold value holding unit 12. Figure 6
In the example of (a), since the number of black pixels exceeds the threshold value, A
The determination result indicating that the attachment of CF3 is normal is output.

On the other hand, in the case where the ACF 3 has an abnormality such that the sticking is turned over, the black and white binarized image stored in the image memory 8 is as shown in FIG. 6B. In this case, since the ACF 3 is turned over in the recognition area 30 and is not present in this portion, the white dummy lead 41 is exposed and the number of black pixels is reduced.

The counting section 10 counts the total number of black pixels in the recognition area 30 shown in FIG. 6 (b). Then, the comparison circuit 11 compares the count value of the black pixels with the threshold value held in the threshold value holding unit 12. Figure 6
In the example of (b), the number of black pixels is less than the threshold value.
The determination result indicating that the attachment of CF3 is abnormal is output.

According to the above-described embodiment, the image obtained by picking up the glass substrate 1 with the video camera 1 is binarized into a black-and-white binary image, and the number of white pixels in the recognition area of this black-and-white binary image. Alternatively, since the normal or abnormal state of the ACF 3 attachment state is determined by the number of black pixels, it is possible to set a wide determination area and the ACF 3 attachment state (position).
It is possible to improve the detection accuracy of the turn-over detection of the ACF 3 by the image picked up by the video camera 1 without being affected by the variation of No. 2 and to increase the turn-over detection rate of the ACF 3.

Further, since the image picked up by the video camera 1 is used, it is not necessary to pass light between the leads 4 and it is possible to cope with a narrow pitch of the leads 4. Furthermore, attach the ACF to the glass substrate or PCB substrate of the video camera 1 A
Space saving can be achieved by also using the existing camera for positioning which the CF attaching device has.

Further, as shown in each of the above embodiments, the black-and-white binarization level at the time of binarizing the picked-up image is changed in accordance with the detection condition, and it is judged by the number of white pixels or the black pixels. By selecting whether to judge by the number of
It is possible to perform the most effective ACF turn-up state detection with respect to various glass substrates, leads, ACF, and other detection conditions.

Since the recognition area 30 determines the range for counting the pixels of the binarized image, the count time is set to be short and the location where the ACF 3 is easily turned over is set to the above-mentioned count range. It is possible to improve the determination accuracy of the attached state.

The present invention is not limited to the above embodiment, but can be implemented in various other forms. Further, it can be applied to detect the attachment state when a tape or the like is attached to a PCB substrate or the like, and the same effect can be obtained.

In the above embodiment, the image processing unit 7 divides the imaged image data of the video camera 6 into a predetermined number of pixels and then binarizes each pixel into black and white. Is divided into a predetermined number of pixels, the minimum unit (1 pixel) of the imaging resolution of the video camera 6 is set to 1 as it is.
It may be processed as a pixel or a plurality of pixels may be processed as one pixel.

Further, the inspection position of the attached state of the ACF 3 has been described as an example in which the end portion is likely to be turned over, but the inspection position is not limited to only the end portion of the attached ACF 3, and it is necessary. It may be set at an arbitrary position according to. For example, both ends of the attached ACF 3 and a plurality of positions between them may be set as inspection positions, and the inspection may be performed for each inspection position. By doing so, it is possible to perform an inspection with a higher judgment system.

[0046]

As described above in detail, according to the present invention, the detection rate can be increased and the lead rate can be increased by detecting the curling of the inspection object using the image picked up by the video camera. It is possible to cope with the narrowing of the pitch.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an ACF detection device according to an embodiment.

FIG. 2 is an explanatory diagram showing an example of a captured image of Example 1 displayed on a monitor.

FIG. 3 is an explanatory diagram showing an example of a black-and-white binary image of the first embodiment held in an image memory.

FIG. 4 is an explanatory view showing an example of a black and white binarized image of Embodiment 2 held in an image memory.

FIG. 5 is an explanatory diagram showing an example of a captured image of Example 3 displayed on a monitor.

FIG. 6 is an explanatory diagram showing an example of a black and white binarized image of the third embodiment held in an image memory.

FIG. 7 is an explanatory diagram of an ACF attachment state detection method using a conventional transmissive sensor.

FIG. 8 is an explanatory view of a conventional ACF attachment state detection method using a reflective sensor.

[Explanation of symbols]

1 ... Glass substrate 3 ... ACF 4 ... Lead 6 ... Video camera 7 ... Image processing unit 8 ... Image memory 9 ... Monochrome binarization level setting section 10 ... Counter 11 ... Comparison circuit 12 ... Threshold holding unit 13 ... Monitor 41, 42 ... Dummy lead

   ─────────────────────────────────────────────────── ─── Continued front page    F term (reference) 2G051 AA65 AB14 CA04 CB01 EA11                       EA12 EA14 EB01 ED09                 2H088 FA12 FA16 FA17 FA24 FA28                       FA30 HA05 MA16                 5G435 AA17 EE33 EE42 KK05 KK10

Claims (5)

[Claims] 1. An image pickup unit for picking up an image of a substrate on which an inspection object is attached, an image binarization unit for binarizing the image picked up by the image pickup unit into black and white for each pixel, and the image 2 Counting means for counting at least one of the number of white pixels or the number of black pixels of the black-and-white binarized image binarized by the binarizing means, the count value counted by the counting means, and a preset threshold value. And a determining unit that determines whether the pasting state of the inspection target is normal or abnormal. 2. The range of the black-and-white binarized image in which the number of white pixels or the number of black pixels is counted by the counting means is a range in which an abnormality is likely to occur in the attachment state of the inspection object. The board inspection apparatus according to claim 1. 3. An image pickup step of picking up an image of a substrate to which an inspection object is attached, an image binarization step of binarizing the image picked up by the image pickup step for each pixel, and the image 2 A counting step of counting at least one of the number of white pixels or the number of black pixels of the black and white binarized image binarized by the binarizing step, and the count value counted by the counting step and a preset threshold value. And a determination step of determining whether the attached state of the inspection target is normal or abnormal. 4. The board inspection method according to claim 3, wherein in the image binarization step, the black-and-white binarization level is changed according to a detection condition. 5. The substrate inspection method according to claim 3, wherein in the counting step, whether to count a pixel as a white pixel or a black pixel is selected according to a detection condition. .