JPS61198374A - Device for comparing and checking two-dimensional picture - Google Patents

Abstract

PURPOSE:To attain quick and accurate check without being affected by deformation or distortion by correcting the moving quantity when an object moves to the next image pickup position even if distortion or deformation in size being a picture detection resolution or over exists in the check object. CONSTITUTION:The position is corrected at each pickup pattern. That is, check objects 11, 12 having a 2-dimension pattern are fixed to vernier tables 31, 32 provided on an XY table 21 driven by XY table drive motors 22, 23 in the X and Y directions. The image pickup devices 41, 42 are fixed to a position picking up image of the same position of the check object. An output signal of the image pickup device is stored in storage elements 61, 62 arranged in a matrix as a binary-coding video signal via binary-coding and noise rejection circuits 51, 52. The signal is fed to a defect discriminating circuit 7, where a defect in the pattern is extracted. Then an output of a noise rejection circuit is sent to a position deviation detecting circuit 8 to measure the position deviation. The storage device 62 is controlled in response to the measured deviation to correct the deviation for accurate check.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to an apparatus for comparing a plurality of two-dimensional patterns and inspecting patterns for external defects, and is particularly applicable to semiconductors, printed circuit boards, and their manufacturing. The present invention relates to a two-dimensional image comparison inspection device suitable for inspecting used circuit patterns.

[Background of the invention]

In the conventional two-dimensional image comparison inspection apparatus, positioning is performed visually or automatically using a part of the pattern before the start of the inspection, and thereafter the inspection proceeds without any alignment.

However, with recent large printed circuit boards, dimensional distortion due to shrinkage of the printed circuit board itself (for example, ±0.31 m for 300 mm)
degree) and deformation. Furthermore, displacement and rotation due to errors in positioning performed before the start of inspection are added. Compared to the required resolution of 10 to 20 μm in the inspection of high-density, fine patterns, the positional deviation caused by the above factors has become too large to be ignored.

In addition, a two-dimensional image comparison inspection device that uses a positional deviation detection circuit that uses exclusive OR may malfunction if the amount of positional deviation is unconditionally detected based on the contents of a counter that counts the result of exclusive OR. there were. If a part of the inspection target has a small pattern area ratio or has no pattern at all, it may be caused by incorrect alignment due to defects in the inspection target or noise in the binary video signal. Conceivable.

[Purpose of the invention]

A first object of the present invention is to provide a two-dimensional image comparison inspection device that can quickly and accurately inspect an object to be inspected without being affected by dimensional distortion or deformation that exceeds the image detection resolution. It is.

A second object of the present invention is to quickly and accurately inspect a pattern to be inspected without erroneous alignment operations, even if there is a small pattern area ratio or no pattern at all. The object of the present invention is to provide a two-dimensional image comparison inspection device that can perform the following tasks.

[Summary of the invention]

In order to achieve the first object, the present invention minimizes the number of mismatched bits when the information of a storage element storing a plurality of two-dimensional images is shifted by 0 to several bits in each direction and superimposed. Focusing on the fact that the amount of movement corresponds to the amount of positional deviation, this structure corrects the amount of movement when moving to the next imaging position so that it is not affected by dimensional distortion or deformation of the printed circuit board, etc. This is what we propose.

In order to achieve the second object, the present invention also provides a counter that performs exclusive logical interest counting between binarized video signals of two two-dimensional patterns to be compared when the ratio of the pattern area to the imaging area is small. Focusing on the fact that the added value of the content becomes a small value, when the added value is less than a certain value, the alignment operation is not performed at all, and the part of the pattern to be inspected has a small or no pattern area ratio. The present invention proposes a configuration that prevents erroneous alignment operations even if there is a problem.

[Embodiments of the invention]

A first embodiment of the present invention will be described below with reference to FIG. The inspection object 11° 12 having a two-dimensional pattern is a fine adjustment table 31.3 provided on an XY table 21 movable in the XY direction by XY table drive motors 22, 23.
It is fixed at 2.

The fixing is done by a mechanical method using a guide hole or an end as a reference, and positioning is possible with an accuracy comparable to that of an O12 dragon. The imaging devices 41 and 42 are fixed at positions for imaging the same position ii of the inspection object. The output signals of the imaging device pass through binarization and noise removal circuits 51 and 52, respectively, and are stored as binarized video signals in storage elements 61 and 62 arranged in a matrix. From there, the signal is sent to a defect determination circuit 7, where defects such as disconnections, short circuits, irregularities, and pinholes in the pattern are extracted.

However, as described above, if the imaging position differs due to misalignment, two-dimensional distortion, deformation, or other causes, erroneous inspections will occur frequently and accurate inspections will not be possible. Therefore, the output of the noise removal circuit is sent to the positional deviation detection circuit 8 to measure the amount of positional deviation. controlling the storage device 62 according to the measured amount of deviation;
To enable accurate inspection by correcting the amount of positional deviation.

The positional deviation detection circuit 8 converts the two-dimensional image into m bits in the vertical direction and n bits in the horizontal direction.
When the bits are moved, the number of mismatched pits within the imaging plane is counted, the result is stored in the counter at the (m, n)th address, and the minimum number of counters (mo.

Detecting the no) number, the positional deviation is 1 vertical m,) bit horizontal n □
It detects bits.

The number of counters is determined according to the amount of positional deviation to be corrected, so as mentioned above, if the maximum positional deviation is 0.3 m,
If the resolution is 10 μm, then the vertical and horizontal ±30 bits will be corrected, and (30X2+1)” = 3721 counters will be required.This is equivalent to the vertical and horizontal ±4 bit correction. The amount is approximately 50 times larger, and it takes several thousand μs to determine the minimum value, which is not realistic.

One screen captured by an imaging device ranges from a few parrots to a dozen or so, and even if a positional deviation of 0.31500 occurs, for example, in a range of 5 frames, the positional deviation will be only 3 μm, and the resolution will be 10 μm. It is small enough in comparison and does not pose a problem.

In other words, it is not a problem within the range that fits on one screen, but
It is necessary to correct the amount of positional deviation that is accumulated when the inspection object is moved.

Therefore, the output of the positional deviation detection circuit 8 is sent to the table control circuit 9, and the fine adjustment table 31.3 is sent to the table control circuit 9 each time the positional deviation is measured.
By controlling the motors 33 and 34 that drive 2, the position is corrected, and the influence of dimensional distortion and deformation 2 alignment error is eliminated, thereby reducing erroneous inspections.

The problem with this method is that the moving XY table 21
The backlash of the fine adjustment table provided above is a cause of positional deviation. This can be adequately addressed by appropriately selecting the preload of the ball screw and rail.

Furthermore, since it is sufficient for the correction operation to correct one screen movement, that is, a few bits, the determination to detect the minimum counter from within 100 counters is completed within 1 μs.

Since the fine adjustment table also only moves a few tens of μm, the movement of the fine adjustment table is completed within 0.1 seconds, which is the time required for the XY table to advance one screen, and does not affect the inspection time at all.

Figure 1 shows an example in which the left channel is moved vertically and the right channel is moved horizontally to perform the correction operation, but the reverse may also be used, or the same effect can be achieved by moving one channel vertically and horizontally. produces the effect of

In addition, as the imaging device, a television camera or a linear sensor may be used1.Furthermore, in the above embodiment, the imaging device is fixed and the inspection object is moved, but the inspection object is moved. A method in which the imaging device is fixed and moved can also be adopted.

According to this embodiment, position correction is performed for each imaged screen, so even when dimensional distortion and deformation are considerably large compared to the resolution, such as when inspecting a large printed circuit board, the This enables quick and accurate testing.

Next, a second embodiment of the present invention will be described with reference to FIGS. 2 and 3. Two inspection objects 11 and 12, such as printed circuit boards, having two-dimensional patterns are fixed to fine adjustment tables 31 and 32 provided on an XY table 21 that is movable in XY71i directions. Fixing is a mechanical method based on guide holes or edges, and in the case of printed circuit boards, 0
．． Positioning is possible with an accuracy of about 2 rat. The two imaging devices 41.4'2 are fixed at positions for imaging the same position of the two inspection objects 11.12. The output signals of the imaging device are respectively processed by binarization and noise removal circuits 51.
52, storage devices 61 arranged in a matrix
．． 62 as a binary video signal. These signals are subjected to feature extraction circuits 71 and 72 in which features such as boundaries and minute portions corresponding to each pattern are extracted.

The features of the two images are compared by a comparison/determination circuit 73, and if the features of both images completely match, it is determined that there is no defect.
If they do not match, it is determined that there is a defect. for example,
There is a disconnection in one of the patterns of inspection target 11.12,
If there are defects such as short circuits, irregularities, or pinholes, features corresponding to those defects are extracted, so it is determined that the two features do not match and there is a defect. In this comparative determination method, in order to determine whether or not there are identical features at the same position or extremely close positions, it is necessary to align the two patterns to be compared with an accuracy of about ±10 to 20 ttm. Therefore, the outputs of the noise removal circuits 51 and 52 are sent to the positional deviation detection circuit 8 to measure the amount of positional deviation. The storage device 62 is controlled according to the measured amount of deviation to correct the amount of positional deviation, and the output of the detection circuit 8 is sent to the table control circuit 9 to drive the fine adjustment tables 31 and 32 each time the positional deviation is measured. The motors 33 and 34 are controlled to correct cumulative positional deviations due to dimensional distortions and deformations of the inspection objects 11 and 12.

Here, the positional deviation detection circuit 8 uses a circuit to which exclusive OR is applied. FIG. 3 shows an example of the configuration of the positional deviation detection circuit. The positional deviation detection circuit is (m columns horizontally x n rows vertically)
+ ((m-n) horizontally x I rows) shift register circuit 81A, and (m columns horizontally x (2 vertically)
A second shift register circuit 81B composed of shift registers (rows 11+1) and a first shift register circuit 8
The contents of the register in (n+1) row and (m-n) column of 1A and (m -2n-1~ of second shift register circuit 81B)
m) Calculate the exclusive OR with each register configuring the cutout circuit in the (1 to 2n+1) columns of rows (2n+
1) x (2n+1)@ exclusive OR circuit 82 and exclusive OR circuit 82 calculation results are individually counted (
2n+1 ) x (2n+1) counter circuits 83
and a deviation 1 calculation circuit 84 that detects the counter circuit with the minimum count value during the counter of the counter circuit 83 and calculates the amount of deviation.
It consists of

The number m of registers in the horizontal direction of the shift register circuits 81A and 81B is the number of binarized signals corresponding to one horizontal scan of the imaging device 41, 42. ``In addition, a noise removal circuit 51 is installed in the register in the first row and first column of the first shift register circuit 81A.
The output signal of is input, and the second shift register circuit 81B
The output signal of the noise removal circuit 52 is input to the register in the first row and first column. Shift register circuits 81A, 81
The output of row ith column (i<m) of B is connected to the input of row 2 and column (i+1), and the output of row k and column m is connected to the input of row 1 (k+1) and column 1. There is. When the binary video signal is serially input from the noise removal circuits 51 and 52 to the shift register circuits 81A and 81B having this configuration, the shift register circuit 81A.

In 81B, a binary video signal is stored in an array corresponding to the shape of the pattern on the inspection object 11.12 imaged by the imaging device 41.42, and the next 1-bit signal is stored in 1 row 1.
Each time the image signals are input to the column register, the image signals of all the rows are shifted one bit at a time.

Therefore, the shift register circuit 81B ((m-2n
-1) to m) Columns (1 to 2n+1) rows of the registers of the extraction circuits also pass through the entire video signal while being shifted one bit at a time in an arrangement corresponding to the pattern shape.

This cutting circuit and the first shift register circuit 81B (n
+1) Exclusive OR of the row (m-n) column register (the coordinates of this register are equal to the coordinates of the register at the center of the cutout path of the second shift register circuit) and the entire video signal passing through the cutout path. When the counter with the minimum integrated value is detected, the video signal that has passed through the first shift register circuit 81A, that is, the pattern imaged by the imaging device 41 and the counter that has passed through the second shift register circuit 81B. Positional differences between the video signal, that is, the pattern imaged by the imaging device 42, can be detected. For example, the register at the center of the cutout circuit ((n+1) rows (m-n
If the value of the counter that is the sum of the exclusive OR with the register (J column) is the minimum, the first and second patterns will not be misaligned, and the register (n row (m-n+21st column register) If the value of the counter that has accumulated the exclusive OR with The position is shifted by +2 bits in the -1 bit column direction.

However, when the area of the imaged pattern is small compared to the total area of the imaged pattern (2
(If the ratio of H and L of the converted video signal is unbalanced), there is a possibility that the amount of positional deviation will be detected incorrectly. The cause of this false detection will be explained below. The operating principle of the positional deviation detection circuit is to move the second pattern on the comparison side in all directions around the first pattern on the reference side, and detect the direction of movement that minimizes the mismatch with the fourth pattern at that time. Since they are equivalent, if the ratio of the area of the imaged pattern is small to the total area imaged, most of the non-pattern areas will match, so the count of the amount of mismatch will be a small value for all counters. At this time, if a defect pattern exists in one of the inspection targets or noise enters the binary video signal, a discrepancy unrelated to the actual positional deviation will be detected and the count value of the counter will be increased. Therefore, the count value of the counter corresponding to the amount of positional deviation will no longer be the minimum value, causing erroneous detection.

When the ratio of the area of the pattern is large, the number of mismatch counts for all counters becomes large, and the difference between the minimum count and other counts becomes large, so the possibility of false detection becomes extremely low. .

This embodiment focuses on which phenomenon, and as shown in FIG. 3, all the count values of the counter circuit 83 that counts the output of the exclusive OR circuit 82 are added by the addition comparison circuit 85, and the added value is obtained. If it is less than the set value, the result of the deviation 1 calculation circuit 84 is forced to be output as 0, that is, there is no positional deviation for both patterns. Here, if the addition value of the addition comparison circuit 85 is greater than a certain value, the error 1 calculation circuit 8
The result of step 4 is output as is.

In this way, in the detection circuit of this embodiment, the addition and comparison circuit 8
By adding 5, it is determined that there is no positional shift when the ratio of the imaged pattern area is small, so that erroneous positioning operations can be prevented.

Further, the addition/comparison circuit 83 does not need to add all the digits of the counter, but may only add, for example, the most significant one or two digits.

In addition, in general, misalignment of the inspection target, for example in the case of printed circuit boards, is caused by deformation such as elongation or shrinkage during molding, and is often a gradual accumulation of misalignment from a positional perspective. There is almost no problem in practical use, even if positioning is not performed at a part of the way where there are few patterns, as long as positioning is performed at the next point where a pattern appears.

Note that the imaging device in this description has the same effect whether it is a television camera using an imaging tube, a two-dimensional semiconductor image sensor, or the like, or a linear sensor.

Further, in the embodiment, the explanation has been made so that everything is configured using a hardware circuit, but the same effect can be obtained even if a part is processed by software using a microprocessor.

According to this embodiment, when the pattern area is small relative to the imaged area, an incorrect positioning operation is not performed, so there are some parts of the pattern where the ratio of the pattern area is small or where there is no pattern at all. It has the effect of being able to accurately test even the objects being tested.

〔Effect of the invention〕

According to the invention, position correction is performed for each imaged screen, so even when dimensional distortion and deformation are considerably large compared to the resolution, such as when inspecting a large printed circuit board, it is possible to This enables quick and accurate testing.

First, if the pattern area is small compared to the imaged area, position correction for each screen is canceled to prevent incorrect alignment, so if there is a part of the pattern that has a small ratio of pattern area or Even if there are many places to be inspected where there is no oxidation, it can be inspected quickly and accurately.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing a first embodiment of the invention, FIG. 2 is a block diagram showing a second embodiment of the invention, and FIG. 3 is a block diagram showing a second embodiment of the invention.
FIG. 2 is a circuit diagram showing an example of a positional deviation detection circuit according to an embodiment. 11.12...Inspection object, 21...XY table,
22.23...XY table drive motor, 31°32
...Fine adjustment table, 33, 34...Fine adjustment table drive motor, 41.42...Imaging device, 51.52...
・Binarization noise removal circuit, 61.62...Storage device,
7... Defect determination circuit, 71. 72... Feature extraction circuit, 73... Comparison/determination circuit, 81A, 81B... Shift register circuit, 82... Exclusive OR circuit, 83.
...Counter circuit, 84...Difference 1 calculation circuit, 85.
... Addition comparison circuit.

Claims (1)

[Scope of Claims] 1. An imaging device that captures two-dimensional images of a plurality of inspection objects roughly positioned in advance, and a storage circuit that samples each of the captured two-dimensional images and stores them as a binary video signal. , a two-dimensional image comparison inspection device equipped with a defect determination circuit that compares the stored contents and determines that there is a defect when the two do not match;
~ A positional deviation detection circuit that integrates the number of mismatched bits when moving several bits and calculates the amount of positional deviation between two-dimensional images from the integrated value, and a A two-dimensional image comparison/inspection apparatus comprising: a control circuit for correcting the amount of positional deviation by movement, and correcting the positional deviation for each imaging. 2. A two-dimensional image comparison inspection device according to claim 1, wherein the imaging device is a television camera. 3. A two-dimensional image comparison inspection device according to claim 1, wherein the imaging device is a linear sensor. 4. A two-dimensional image comparison inspection apparatus according to any one of the above claims, wherein the control circuit outputs a signal for moving an inspection object. 5. A two-dimensional image comparison inspection device according to any one of claims 1 to 3, wherein the control circuit outputs a signal for moving an imaging device. 6. Comparison of the storage contents of an imaging device that captures two-dimensional images of a plurality of inspection targets that have been roughly positioned in advance, and a storage circuit that samples each of the captured two-dimensional images and stores them as a binary video signal. In a two-dimensional image comparison inspection apparatus equipped with a defect determination circuit that determines that there is a defect when the two do not match, the stored contents are zero-coded in each direction.
- Adds up the number of mismatched bits when moving several bits, calculates the amount of positional deviation between two-dimensional images from the integrated value, adds the integrated values, and calculates the amount of positional deviation when the added value is outside the preset range. A positional deviation detection circuit that sets the positional deviation to zero, and a control circuit that corrects the positional deviation amount by moving the inspection object and the imaging device relative to each other prior to the next imaging, and detects the positional deviation when the positional deviation occurs. A two-dimensional image comparison inspection device characterized by correcting. 7. A two-dimensional image comparison inspection device according to claim 6, wherein the imaging device is a television camera. 8. A two-dimensional image comparison inspection device according to claim 6, wherein the imaging device is a linear sensor. 9. A two-dimensional image comparison inspection apparatus according to any one of claims 6 to 8, wherein the control circuit outputs a signal for moving an inspection object. 10. A two-dimensional image comparison inspection device according to any one of claims 6 to 8, wherein the control circuit outputs a signal for moving an imaging device.