JP2949089B2 - Image processing device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to an image processing apparatus for detecting an edge of an object such as a bonding pad on an IC chip and detecting the position of the object.

[0002]

2. Description of the Related Art Conventionally, a method often used for detecting an edge of a bonding pad (hereinafter, referred to as a pad) on an IC chip is a detection method based on a projection view. In this method, as shown in FIG. 9, a vertical projection diagram (projection) is applied to a detection area 10 designated in advance.
Then, a projection in the horizontal direction is created, and a point where each projection crosses a predetermined threshold value is defined as an edge of the pad 11.

[0003]

The method of detecting a pad based on a projection view is an effective method only when no object other than the pad exists in the detection area. But in fact,
In the vicinity of the pad, there are a lead pattern and another pad. Since these objects affect the projections in the vertical and horizontal directions in the detection area, the edges of the pad cannot be accurately detected.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described drawbacks, and has as its object to accurately detect an edge of a target object even if an object other than the target object (such as a pad) exists in a detection area. It is an object of the present invention to provide an image processing device capable of detecting the same.

[0005]

[Means for Solving the Problems]

A. In order to achieve the above object, an image processing apparatus according to the present invention includes: a first unit that determines a detection area and a threshold for a captured image; and a cross-sectional view of density for each line in the detection area. Second means for detecting an edge of an object in the detection area based on a threshold value and a sectional view of the density.

The second means sets a point in the detection area as an edge detection start point, and creates a cross-sectional view of the density for each line from the edge detection start point toward an end of the detection area.

The second means performs edge detection again after enlarging the detection area when an edge pair indicating both ends of the object cannot be detected from the cross section of the density created first.

[0008] The second means, when an edge pair indicating both ends of the object can be detected from the cross-sectional view of the density created from the second time on, when the next line deviates from the detection area, the detection area After the enlargement, edge detection is performed again.

In the image processing apparatus according to the present invention, when the size of the detection area is maximized, the size of the detection area is restored, the edge is detected again after the threshold value is changed. A third means for determining that the edge detection of the object is impossible when the threshold value exceeds an upper limit or a lower limit.

B. An image processing apparatus according to the present invention includes a first unit that determines a detection area and a threshold value for a captured image, and creates a cross-sectional view of density for each line in the detection area, and A second step of detecting an edge of the object in the detection area based on the sectional view;
Means, and third means for setting a re-detection area smaller than the detection area based on the detection result of the edge of the object,
A fourth step of creating a cross-sectional view of the density for each line in the re-detection area and re-detecting the edge of the object based on the threshold and the cross-sectional view of the density in the re-detection area
Means.

The third means sets the re-detection area to a range including at least all the edges of the object detected by the second means. The image processing apparatus of the present invention further obtains an average value of the position of the edge in the re-detection area when the detection rate of the edge of the object by the fourth means is equal to or more than a predetermined determination value. Fifth means for setting a value to a position of an edge of the object is provided.

The image processing apparatus of the present invention further enlarges the detection area determined by the first means when a detection rate of the edge of the object by the fourth means is less than a predetermined determination value. Thereafter, a fifth means for performing edge detection again is provided.

In the image processing apparatus according to the present invention, when the size of the detection area is maximized, the size of the detection area is restored, the threshold value is changed, and then the edge is again detected. A sixth means for performing detection is provided, and a seventh means for determining that edge detection of the object is impossible when the threshold value exceeds an upper limit or a lower limit.

[0014]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an image processing apparatus according to the present invention will be described in detail with reference to the drawings. FIG. 1 shows the overall configuration of the image processing apparatus of the present invention. The image capturing unit 1 captures a reference image and an image to be inspected, and includes a CCD camera, a video camera, a scanner, and the like. The image memory 2 stores a reference image and an image to be inspected, and stores image data of at least two frames.
Data storage capacity.

The threshold value determining circuit 3 is a circuit for determining a threshold value which is used as a reference when detecting a pad edge. The rough edge detection circuit 4 is a circuit that detects a rough position of a pad edge by an edge detection algorithm described below. The determination circuit 5 is a part that performs various types of determination performed when the approximate edge detection circuit 4 determines the approximate position of the edge.

The detailed edge detecting circuit 6 is a circuit which determines a range for obtaining an accurate edge position based on the approximate edge position, and further detects a pad edge position in the range. The determination circuit 7 is a part that performs various determinations performed when the edge position is obtained by the detailed edge detection circuit 6.

Next, the operation of the image processing apparatus of FIG. 1 will be described. In the following, a pad detection method for detecting the position of a pad will be described as an example.

First, the reference image and the image to be inspected are fetched into the image memory 2 by the image fetching unit 1, and a pad detection area 10 is set as shown in FIG. In this example,
For example, the upper left of the detection area 10 is set as the coordinate origin O. Further, the threshold value determining circuit 3 determines a reference threshold value when detecting the edge of the pad.

Next, the approximate edge detection circuit 4 and the determination circuit 5 detect the approximate edge position of the pad. This edge detection is performed so as to prevent erroneous detection due to foreign matter on the pad or other than the designated pad.

First, a predetermined point (pixel) in the detection area 10 is selected by an operator, and the predetermined point is set as an edge detection start point P. Further, as shown in FIG. 3, a cross-sectional view of the density in the horizontal direction (X direction) is created for each line (row) from the edge detection start point P toward the upward direction a of the detection area 10, and Regarding the cross-sectional view of each density, a point (pixel position) at which the cross-sectional view crosses the threshold value is roughly defined as an edge position.

Here, since it is considered that no edge is detected after the line k, if the edge is not detected continuously for the preset number of designated lines,
Assuming that the detection line has deviated from the pad, the edge detection in the upward direction a ends.

Further, even when the number of lines for which edge detection has not been performed has not reached the specified number of lines, but the detected line has deviated from the detection area 10, the edge detection in the upward direction a is terminated.

Similarly, from the edge detection start point P to the downward direction b of the detection area 10, one line (line) is formed for each line (row) in a cross-sectional view of the density in the horizontal direction (X direction). Regarding the cross-sectional view, a point (pixel position) at which the cross-sectional view crosses the threshold value is roughly defined as an edge position.

As for the downward direction b, similarly to the upward direction a,
If edge detection has not been performed continuously for the specified number of lines, or if the detected line has deviated from the detection area 10, the edge detection in the downward direction b ends.

When the edge detection in the upward direction a and the downward direction b is completed, for example, as shown in FIG. 4, an edge pair group (indicated by a cross) indicating the approximate position of the edge in the horizontal direction (X direction) of the pad Is obtained.

Further, the left direction c of the detection area 10 in FIG.
Also, for the right direction d, a cross-sectional view of the density in the vertical direction (Y direction) is created for each line (row) from the edge detection start point P toward the left direction c or the right direction d of the detection area 10. In addition, with respect to the cross-sectional views of each concentration, a point (pixel position) where the cross-sectional view crosses the threshold value is set as an approximate edge position.

When the edge detection in the left direction c and the right direction d is completed, for example, as shown in FIG. 5, an edge pair group (indicated by a cross) indicating the approximate position of the edge in the vertical direction (Y direction) of the pad Is obtained.

When detecting an edge in a predetermined direction from the edge detection start point P, if no edge pair is found, the detection area 10 is narrow with respect to the pad, that is, the pad protrudes from the detection area 10. it is conceivable that.

Therefore, in such a case, after the size of the detection area 10 is enlarged by a predetermined amount, the above-described general edge detection algorithm is applied again to detect the edge pair.

However, if the size of the detection area 10 exceeds the preset maximum detection area, the detection area 1
No expansion of 0 is performed. In this case, the detection area 10 is returned to the initial size, and the threshold value is increased (or decreased), and the above-described edge detection operation is performed again.

When the threshold value exceeds a preset maximum threshold value (or falls below a minimum threshold value),
The pad detection is not possible (NG), and the edge detection operation ends.

Next, the detailed edge position of the pad is detected by the detailed edge detection circuit 6 and the determination circuit 7. The purpose of this edge detection is to accurately detect the edge of the pad by checking the linearity of the edge, and to improve the accuracy of the edge detection position by obtaining the average value of the edge positions.

First, re-detection areas A, B, C and D are set near the four edges of the pad. This redetection area is determined as follows. As shown in FIGS. 4 and 5, the edge whose horizontal position is farthest from the horizontal position of the edge detection start point P is detected from the edge pair group in the horizontal direction (X direction), and the edge is detected. Are assumed to be the approximate horizontal edge coordinates Emax and Emax '. Further, an edge whose horizontal position is closest to the horizontal position of the edge detection start point P in the horizontal edge pair group is detected, and the coordinates of the edge are converted to the horizontal approximate edge coordinates Emin, Emin.
'.

Similarly, of the edge pair group in the vertical direction (Y direction), the edge whose vertical position is farthest from the vertical position of the edge detection start point P is detected, and the coordinates of the edge are set in the vertical direction. The directions are assumed to be roughly edge coordinates Fmax and Fmax '. Further, an edge whose vertical position is closest to the vertical position of the edge detection start point P in the vertical edge pair group is detected, and the coordinates of the edge are determined by the vertical approximate edge coordinates F.
min and Fmin '.

The horizontal direction of the re-detection area A is at least E so that all the edge pairs in the horizontal direction are included.
It is set wider than the range from max to Emin, and the vertical direction is set from Fmax to Fmax '.

Similarly, the horizontal direction of the re-detection area B is at least E so that all the edge pairs in the horizontal direction are included.
It is set wider than the range from max 'to Emin', and the vertical direction is set from Fmax to Fmax '.

The vertical direction of the re-detection area C is at least Fma so that all edge pairs in the vertical direction are included.
It is set wider than the range from x to Fmin, and the horizontal direction is set from Emax to Emax '.

Similarly, the vertical direction of the re-detection area D is at least F so that all vertical edge pairs are included.
The width is set wider than the range from max 'to Fmin', and the horizontal direction is set from Emax to Emax '.

Next, as shown in FIG. 6, pad edges are detected again in the re-detection areas A, B, C and D. If the edge detection rates in the re-detection areas A, B, C, and D are equal to or greater than a predetermined determination value, the average value of the coordinate values of the edges detected in the re-detection areas A, B, C, and D Is the edge coordinate.

On the other hand, when the edge detection rates in the re-detection areas A, B, C, and D are less than a predetermined judgment value, the detection area 10 is enlarged by a predetermined amount, and then the above-described operation is started again from the beginning. An edge detection algorithm is applied to perform edge detection.

However, if the size of the detection area 10 exceeds the preset maximum detection area, the detection area 1
The threshold value is increased (or decreased) while 0 is returned to the initial value, and the above-described edge detection operation is performed again.

If the threshold value exceeds the preset maximum threshold value (or falls below the minimum threshold value), the pad detection is impossible (NG) and the edge detection operation is terminated.

After detecting the edge coordinates (average value) in the re-detection areas A, B, C, and D, the position of the pad is obtained based on the coordinates of the edge. FIGS. 7 and 8 show flowcharts of the above-described edge detection algorithm.

In the above embodiment, it is assumed that the object (pad) to be subjected to position detection is brighter than the background, but the reverse is also possible. Further, even if the object whose position is to be detected is not a square, the above-described edge detection algorithm can be applied.

[0045]

As described above, according to the image processing apparatus of the present invention, the following effects can be obtained. After performing rough edge detection in the detection area including the entire pad, a re-detection area is set near the edge, and detailed edge detection is performed again in the re-detection area.

Also, in each edge detection, a sectional view of the density for each line is used without using a projection diagram (projection). Therefore, even if a bonding wire or a lead exists around the pad whose position is to be detected, the pad edge can be detected accurately.

Further, by setting the re-detection area, the linearity of edge detection can be checked, and thereby the edge of the pad can be detected accurately without erroneously detecting a foreign substance or the like on the pad. In addition, by taking the average value of the edge coordinates obtained in the re-detection area, the accuracy of edge detection can be improved.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of an image processing apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram showing an edge detection method according to the present invention.

FIG. 3 is a diagram showing an edge detection method according to the present invention.

FIG. 4 is a diagram showing an edge detection method according to the present invention.

FIG. 5 is a diagram showing an edge detection method according to the present invention.

FIG. 6 is a diagram showing an edge detection method according to the present invention.

FIG. 7 is a flowchart showing the operation of the image processing apparatus of the present invention.

FIG. 8 is a flowchart showing the operation of the image processing apparatus of the present invention.

FIG. 9 is a diagram showing a conventional edge detection method.

[Explanation of symbols]

 10: Detection area 11: Pad, A, B, C, D: Re-detection area.

Claims (11)

(57) [Claims] 1. A first means for determining a detection area and a threshold value for a captured image, and forming a cross-sectional view of density for each line in the detection area, and forming the cross-sectional view of the threshold value and the density. Second means for detecting an edge of the object in the detection area based on the detection of the edge of the object.
Re-detection area smaller than the detection area based on the output result
And third means for setting a value within the re-detection area.
Create a cross section of concentration for each line, and
Based on the cross-sectional view of the concentration in the re-detection area,
An image processing apparatus comprising: fourth means for re-detecting an edge . 2. A detection area for a captured image.
First means for determining a threshold value;
A cross-sectional view of the concentration for each line by
And the object in the detection area based on the cross-sectional view of the concentration
And a second means for detecting an edge of the detection area, the second means as one edge detection start point in the detection area,
A cross-sectional view of density is created for each line from the edge detection start point toward an end of the detection area.
The image processing apparatus according to. 3. The method according to claim 2, wherein when the edge pair indicating both ends of the object cannot be detected from the cross section of the density created first, the edge detection is performed again after enlarging the detection area. 3. The image processing apparatus according to claim 2, wherein: 4. The method according to claim 1, wherein when the edge pair indicating both ends of the object can be detected from the cross-sectional views of the density created after the second time, when the next line deviates from the detection area, the detection is performed. 3. The image processing apparatus according to claim 2 , wherein edge detection is performed again after the area is enlarged. 5. The method according to claim 5, wherein when the size of the detection area is maximized, the size of the detection area is returned to the original size, the threshold value is changed, and edge detection is performed again. the image processing apparatus according to claim 3 or 4, wherein that it comprises. 6. The image according to claim 5 , further comprising: fourth means for judging that the edge of the object cannot be detected when the threshold value exceeds an upper limit or a lower limit. Processing equipment. 7. The third means, wherein the re-detection area is
The image processing apparatus according to claim 1, wherein the set in the range that includes all of the edges of said detected object by at least the second means. 8. When an edge detection rate of the object by the fourth means is equal to or greater than a predetermined determination value, an average value of edge positions in the re-detection area is obtained, and the average value is determined for the object. the image processing apparatus according to claim 1, characterized by comprising a fifth means for the position of the edge. 9. When the detection rate of the edge of the object by the fourth means is less than a predetermined determination value, the detection area determined by the first means is enlarged and then edge detection is performed again. 2. The image processing apparatus according to claim 1, comprising five means. 10. A sixth means for resetting the size of the detection area when the size of the detection area is maximized, changing the threshold value, and performing edge detection again. The image processing apparatus according to claim 9, further comprising: 11. The apparatus according to claim 1, further comprising: seventh means for determining that edge detection of said object is impossible when said threshold value exceeds an upper limit or a lower limit.
0. The image processing apparatus according to item 0.