JP2003156311A - Method and apparatus for detection and registration of alignment mark - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enhance the accuracy and reliability of the detection of an alignment mark by an image processing operation, irrespective of irregularities in the contamination of a substrate or in the production quality of the mark. SOLUTION: When the alignment mark 14 is detected by using a pattern matching operation, edge parts of the alignment mark 14 are covered with a plurality of rectangles S1 , S2 , S3 , S4 , the pattern matching operation is applied to the plurality of rectangles, and the alignment mark is detected.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an alignment mark detecting and registering method and apparatus, and is particularly suitable for use in aligning an object to be processed (referred to as a work) by a laser beam machine. An alignment mark detecting method for detecting an alignment mark using pattern matching, a position detecting method therefor, a registration method for registering a template for detecting the alignment mark, and an alignment mark detecting device using these,
The present invention relates to a position detection device and a registration device.

[0002]

2. Description of the Related Art Generally, in a laser processing machine such as a laser drilling machine for processing a work using a laser beam,
For example, it is necessary to accurately align the positions of the workpiece arranged on the XY stage and the processing laser beam. Therefore, conventionally, as illustrated in FIG. 1, the alignment mark 14 formed in the vicinity of the laser processing hole 12 of the printed circuit board 10 to be processed is imaged by the CCD camera 20 or the like and normalized with an image registered in advance. The pattern matching of the image processing using the correlation is used to detect the mark including only the step of determining the degree of coincidence with the alignment mark registered by the user. This pattern matching is
This is a very easy function for the user because he / she only has to register the image to be detected. In the figure, 22
Is a lighting device.

[0003]

However, in the pattern matching, the brightness pattern of the registered image is compared with the brightness of the image captured from the camera 20,
Since the algorithm is such that the evaluation is high if there are many places where the brightness matches in the image, if the image capturing environment such as the illumination light amount is different from that at the time of registering the mark or the mark 14 is dirty, it is possible to know where to detect. It had the problem of not having it.

For example, as shown in FIG. 2, when the user registers a circular image 16 as an alignment mark and the camera image is a circular image 18 smaller than the registered image 16, the operator determines the degree of coincidence. It is expected that the (correlation value) will be high in the part where the centers of the circles of both pieces coincide, but in reality, the pattern of the background of the registered image 16 is also affected, and thus the case where the centers coincide. Does not always have the highest degree of coincidence, and the photographed images 18, 18 ', and 18''all show similar correlation values, and the center may not be specified. Furthermore, in the extreme case, the registered image is shown in FIG.
Even if the photographed image has a circular shape as shown in (A) and the photographed image has a square shape as shown in FIG. 3 (B), since the background coincidence is also seen as shown in FIG. 3 (C), the coincidence between them is high. Therefore, it may be detected as a match depending on the determination level.

Therefore, the marks cannot be stably detected due to stains on the printed circuit board 10, variations in the manufacturing quality of the marks 14 on the substrate, and the like, which is becoming a problem in a laser processing machine that requires highly accurate positioning. there were.

The present invention has been made to solve the above-mentioned conventional problems, and an object of the present invention is to improve the accuracy and reliability of alignment mark detection by image processing.

[0007]

According to the present invention, when detecting an alignment mark using pattern matching, the edge portion of the alignment mark is covered with a plurality of rectangles, and pattern matching is applied to at least the plurality of rectangles. Then, the above problem is solved by detecting the alignment mark.

In addition, the pattern recognition results of the plurality of rectangles,
Further, when the relative positions of the plurality of rectangles coincide with those of the alignment mark registered in advance, it is determined as the alignment mark.

Further, the matching of the relative positions of the plurality of rectangles is determined by the matching of the relative distances of the plurality of rectangles and the relative positions of the centers of these centers.

The present invention also covers the edge portion of the alignment mark with a plurality of rectangles when detecting the position of the alignment mark using pattern matching.
At least, when pattern matching is applied to the plurality of rectangles and it is determined that they match the alignment mark registered in advance, the absolute position of the center of the centers of the plurality of rectangles is calculated, and the absolute center position is calculated. The problem is solved by detecting the position of the alignment mark.

According to the present invention, when registering a template for detecting an alignment mark using pattern matching, at least a plurality of rectangles that cover the edge portion of the alignment mark and an image cut out by the rectangle are selected. The above problem is solved by registering as a pattern matching template.

According to the present invention, when detecting an alignment mark using pattern matching, a plurality of rectangles which are registered in advance and cover at least the edge portion of the alignment mark, and an image cut out by the rectangle are used as a template. Pattern matching as
The above problem is solved by detecting the alignment mark.

The present invention is also an alignment mark detecting apparatus for detecting an alignment mark using pattern matching, and an image pickup means for picking up a detection target mark or an image, and a detection target input by the image pickup means. Using a mark, at least a plurality of rectangles that cover the edge portion, a registration unit that saves an image cut by the rectangles as a pattern matching template, and a detection target image input by the imaging unit The problem is solved by including a detection unit that performs pattern matching using the template and detects an alignment mark.

The present invention is also an alignment mark position detecting device for detecting the position of an alignment mark by utilizing pattern matching, and an image pickup means for picking up a mark to be detected or an image, and the image pickup means. Using the detection target mark, at least a plurality of rectangles that cover the edge portion, a registration unit that saves an image cut by the rectangle as a template for pattern matching, and a detection target image input by the imaging unit On the other hand, if pattern matching is performed using the template and it is determined that the alignment marks match in advance, the absolute position of the center of the centers of the plurality of rectangles is calculated, and the absolute center position is calculated. By providing a detecting unit for setting the position of the alignment mark, It is obtained by solving.

The present invention is also an alignment mark registration apparatus for registering a template for detecting an alignment mark by utilizing pattern matching, and a plurality of rectangles that cover at least the edge portion of the alignment mark and the rectangles. The problem is solved by providing a registration unit that saves the image cut out by the method as a pattern matching template.

The present invention is also an alignment mark detecting apparatus for detecting an alignment mark by utilizing pattern matching, wherein a plurality of rectangles which are registered in advance and which cover at least an edge portion of the alignment mark are provided. The above problem is solved by providing a detection unit that detects an alignment mark by performing pattern matching using an image cut out by a rectangle as a template.

[0017]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described in detail below with reference to the drawings.

In the present embodiment, as shown in FIG. 4, a CCD camera 20 similar to the conventional one for photographing a detection target mark or an image and a personal computer for processing the output of the CCD camera 20, for example, are used. Image processing device 3 configured as
Contains 0 and.

As shown in detail in FIG. 5, the image processing device 30 includes an image capturing section 32 for capturing an image from the CCD camera 20, which is composed of an image capture card mounted on a personal computer, for example. A mouse 34 and a keyboard 36 which are user IFs for giving various inputs, a calculation unit 38 for performing processing such as pattern matching on a captured image, and a captured image and a calculation result are displayed to an operator. Monitor 40, which is a display unit for storing data, and storage unit 42, which stores the calculation results and the like.
And a bus 44 connecting them.

The operator performs various operations while looking at the image displayed on the monitor 40.

The operation of registering a template for pattern matching will be described below with reference to FIG.

First, in step 100, for example, by moving an XY stage (not shown) on which a printed circuit board is placed, the alignment mark 14 to be detected is placed in the image field 21 of the camera as shown in FIG. Next, in step 102, the capture button is pressed by the keyboard 36, for example. Then, the image processing device 30 performs step 20.
At 0, the image data is captured from the camera 20 via the image capturing section 32, and at step 202, the captured image is displayed on the monitor 40.

The operator then, in step 104,
Press the template registration button. Then, in step 204, the image processing apparatus 30 displays the cursor 50 in the display area of the captured image, as shown in FIG.

The operator then, in step 106,
The cursor 50 is moved by the mouse 34, and one of the positions for specifying the area of the target mark 18, for example, the upper left point is clicked. Then, the image processing apparatus 30 acquires the clicked image coordinate position in step 206 and displays the clicked position as indicated by the broken line in FIG.

The operator then, in step 108,
With the mouse 34, as shown by the solid line in FIG.
0 is moved, and the remaining position for specifying the area of the target mark 14, for example, the lower right point is clicked. Then, in step 208, the image processing apparatus 30 defines a quadrangle in which the upper left vertex is specified by the point clicked first and the lower right vertex is specified by the point clicked this time.
As shown in 0, the image is displayed in the image display area, the clicked image coordinate position is acquired, and the fixed rectangle ABCD is acquired.
Is displayed as enlarged in FIG.

Next, the image processing apparatus 30 proceeds to step 21.
0, the length of one side is 1 / √2 × a × L (L is the length of one side of the quadrangle ABCD determined by the operator, a is a coefficient smaller than 1 such as 0.95), and the operator is determined. Similar to the quadrangle ABCD, the position coordinates of the four corners of the inner quadrangle EFGH whose centers coincide with each other are calculated.

Next, in step 212, as shown in FIG. 12, each side of the calculated inner quadrangle EFGH and the side of the quadrangle ABCD determined by the operator 4 are formed.
The positions of the four quadrangles S1, S2, S3, and S4 (for example, the coordinates of the upper left vertex and the lower right vertex) are calculated. Next, in step 214, the four quadrangles S1, S2, S3, S4 and the image cut out by the quadrangle ABCD determined by the operator as shown in FIG. 13 are stored in the storage unit 42 as a template for pattern matching. , Finish the registration work.

The mark detection in the first embodiment of the present invention using the template registered in the procedure shown in FIG. 6 is performed in the procedure shown in FIG.

That is, first, in both the user mode and the automatic operation mode, the operator performs step 300.
Then, the detection target mark 14 is placed in the camera field 21. Then, in the user mode, the detection button is pressed in step 302. On the other hand, in the automatic operation mode, fetching is instructed in step 304.

The image processing apparatus 30 then proceeds to step 40.
0 through the image capturing unit 32, the CCD camera 20
Import image data from.

In the user mode, when the operator pushes the detection button in step 302 or when the operator instructs the detection in step 306 in the automatic operation mode,
First, in step 402, the image processing apparatus 30 searches the captured image for the entire image by pattern matching using a circular template determined by the operator. Next, in step 404, a pattern matching is performed using four rectangular templates in the vicinity of the position where the entire image is detected.

Next, at step 406, the positional relationship of the detected four rectangles is checked. For example, if the relative distance between the facing quadrangles S1, S2, S3, and S4 deviates by a predetermined value X% or more from the setting, it is determined that the detection has failed.
Returning to step 402, the next whole image is searched.

On the other hand, when it is determined in step 406 that the relative distance between the rectangles is within the allowable range, the process proceeds to step 410, where the coordinates of the upper left and lower right vertices of each rectangle are
As shown in FIG. 13, the centers C _{1 to} C ₄ of _four rectangles are calculated, and the center C ₀ (= C ₁ + C ₂ + C ₃ + C ₄ ) / 4) of the centers is calculated.
Is calculated as an alignment mark position.

Next, in step 412, the captured image is displayed on the monitor 40, the calculated center image is displayed as a numerical value, and the absolute center coordinates are displayed on the coordinates of the captured image with a cross cursor to confirm the operator. Ask for.

In this way, the position of the alignment mark can be accurately detected.

Incidentally, as in the second embodiment shown in FIG.
In step 406 ′, both of the relative distance between S1 to S4 and the relative position of the center C ₀ of the center with respect to S1 to S4 (referred to as relative center position) in the previously registered alignment mark. It is also possible that the alignment mark has been detected when the deviation from and is within the allowable range.

According to the second embodiment, it is possible to detect the alignment mark with higher accuracy than in the first embodiment. .

In each of the above embodiments, first, pattern matching using the whole image is applied as in the conventional case, and then pattern matching using rectangles is performed, so that highly accurate position detection can be performed quickly. It can be carried out. Depending on the shape of the alignment mark, it is possible to omit the pattern matching using the whole image and directly perform the pattern matching using a rectangle.

Although the alignment mark is circular in the above-mentioned embodiment, the shape of the alignment mark is not limited to a circle, and it may be a quadrangle (square or rectangle) as shown in FIG. It may have a triangular shape as shown, as well as other shapes. In any case, the edge portion is covered with a plurality of rectangles, and pattern matching is applied to the rectangles.

In the above embodiment, the present invention is applied to the alignment mark detection of the laser processing machine, but the application target of the present invention is not limited to this, and it is applicable to general alignment mark detection other than the laser processing machine. Obviously, can be applied similarly.

[0041]

According to the present invention, it is possible to improve the accuracy and reliability of alignment mark detection by image processing, because it has robustness against noise due to dirt on the object to be detected and variations in alignment mark creation. For example, the alignment accuracy and processing accuracy of the laser processing machine can be improved. Further, the operator can be freed from complicated settings and the working time can be shortened.

[Brief description of drawings]

FIG. 1 is a perspective view showing how alignment marks are detected.

FIG. 2 is a plan view for explaining an example of conventional problems.

FIG. 3 is a plan view for explaining another example of conventional problems.

FIG. 4 is a block diagram showing the overall configuration of an embodiment of an alignment mark detection device according to the present invention.

FIG. 5 is a block diagram showing a detailed configuration of the image processing apparatus.

FIG. 6 is a flowchart showing a procedure of registration work in the embodiment of the present invention.

FIG. 7 is a diagram showing a state in which a detection target mark is put in a camera image field in registration work.

FIG. 8 is a diagram showing a state in which the cursor is also positioned at the upper left of the area of the target mark.

FIG. 9 is a diagram showing a state in which the cursor is also positioned at the lower right of the area of the target mark.

FIG. 10 is a diagram showing a state in which the cursor is also placed over the area of the target mark.

11 is an enlarged view of a main part of FIG.

FIG. 12 is a diagram showing a state in which a rectangular template is being extracted during registration work.

FIG. 13 is a diagram showing an extracted rectangular template in the same manner.

FIG. 14 is a flowchart showing a procedure of alignment mark detection work according to the first embodiment of the present invention.

FIG. 15 is a flow chart showing a procedure of alignment mark detection work according to the second embodiment.

FIG. 16 is a plan view showing another example of the alignment mark and the template.

FIG. 17 is a plan view showing still another example of alignment marks and templates.

[Explanation of symbols]

10 ... Printed circuit board (work) 12 ... Laser processed hole 14 ... Alignment mark 18 ... Taken image 20 ... CCD camera 21 ... Camera image world 30 ... Image processing device 32 ... Image capturing section 34 ... Mouse 36 ... Keyboard 38 ... Calculation unit 40 ... Monitor (display section) 42 ... Storage section 50 ... Cursor S1, S2, S3, S4 ... Square (rectangle)

Claims (10)

[Claims] 1. When detecting an alignment mark using pattern matching, an edge portion of the alignment mark is covered with a plurality of rectangles, and pattern matching is applied to at least the plurality of rectangles to detect the alignment mark. A method for detecting an alignment mark, which is characterized in that 2. The alignment mark is determined when the pattern recognition results of the plurality of rectangles and the relative positions of the plurality of rectangles match those of an alignment mark registered in advance. The method for detecting the alignment mark described in. 3. The alignment according to claim 2, wherein the matching of the relative positions of the plurality of rectangles is determined by the matching of the relative distances of the plurality of rectangles and the relative positions of the centers of their centers. Mark detection method. 4. When detecting the position of an alignment mark using pattern matching, the edge portion of the alignment mark is covered with a plurality of rectangles, and pattern matching is applied to at least the plurality of rectangles and registered in advance. The position of the alignment mark, which is characterized by calculating the absolute position of the center of the centers of the plurality of rectangles and detecting the absolute center position as the position of the alignment mark when it is determined that they match the alignment mark. Detection method. 5. When registering a template for detecting an alignment mark using pattern matching, at least a plurality of rectangles covering the edge portion of the alignment mark and an image cut by the rectangle are
A registration method of an alignment mark, characterized by being registered as a pattern matching template. 6. When detecting an alignment mark using pattern matching, pattern matching is performed using a plurality of rectangles that are registered in advance and that cover at least the edge portion of the alignment mark and an image cut out by the rectangle as a template. A method for detecting an alignment mark, which comprises performing and detecting the alignment mark. 7. An alignment mark detecting apparatus for detecting an alignment mark using pattern matching, comprising: an image pickup means for picking up a mark to be detected or an image; and a mark to be detected inputted by the image pickup means. ,
At least a plurality of rectangles that cover the edge part,
A registration unit that saves the image cut out by the rectangle as a template for pattern matching, and a detection unit that performs pattern matching on the detection target image input by the imaging unit using the template and detects an alignment mark. An alignment mark detection device, comprising: 8. An alignment mark position detecting device for detecting the position of an alignment mark using pattern matching, comprising: an image pickup means for picking up a detection target mark or an image; and a detection target mark inputted by the image pickup means. Using,
At least a plurality of rectangles that cover the edge part,
A registration unit that saves the image cut out by the rectangle as a template for pattern matching, and pattern matching is performed on the detection target image input by the image capturing unit using the template, and matches the alignment mark registered in advance. When it is determined that the alignment mark is detected, an absolute position of the center of the centers of the plurality of rectangles is calculated, and a detection unit that uses the absolute center position as the position of the alignment mark is provided. Position detection device. 9. An alignment mark registration apparatus for registering a template for detecting an alignment mark using pattern matching, comprising: a plurality of rectangles that cover at least an edge portion of the alignment mark; and an image cut out by the rectangles. To
An alignment mark registration apparatus comprising registration means for storing as a pattern matching template. 10. An alignment mark detecting device for detecting an alignment mark by using pattern matching, wherein a plurality of rectangles which are registered in advance and cover at least an edge portion of the alignment mark, and the rectangle are cut out. An alignment mark detecting apparatus comprising a detection unit for detecting an alignment mark by performing pattern matching using an image as a template.