JPH04236315A - Inspecting apparatus of pattern - Google Patents

Abstract

PURPOSE:To enable execution of inspection without huge-scale hardware by conducting computation of a pattern image and a mask image and by detecting a cut or a projection by extracting characteristic points by the computation. CONSTITUTION:An inner peripheral mask is subtracted from a pattern to be inspected and an inner peripheral mask subtraction signal (i) is outputted, while an outer peripheral mask is subtracted from an inverted image of the pattern and an outer peripheral mask subtraction signal l is outputted. Then, first and second characteristic points (j) and (m) having thin-lined pattern information of the signals (i) and l and also information on a distance between end points in the case when the former information has the end points are outputted respectively. When an input image signal and a thin-lined pattern of the signal (j) are located inside the outer peripheral mask and the distance between the end points is lower than a threshold based on the diameter of the pattern to be inspected, a cut detection signal (k) is outputted, and when the same situations are brought forth in respect to the signal (m), a projection detection signal (n) is outputted. According to this constitution, inspection can be executed without huge-scale hardware.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pattern inspection apparatus, and more particularly to a pattern inspection apparatus for detecting protrusions and chips in patterns composed of circles of the same diameter, such as ground and power patterns of printed circuit boards.

0002

2. Description of the Related Art As a conventional technique, for example, a comparison device using two sets of optical systems is used to compare a non-defective pattern and an inspection pattern mainly using hardware.

0003

[Problems to be Solved by the Invention] The conventional pattern inspection apparatus described above requires precise alignment in order to perform comparative inspection. The disadvantage is that it requires specialized hardware.

0004

[Means for Solving the Problems] The pattern inspection apparatus of the present invention includes: (A) an inner circumferential mask subtraction circuit that subtracts an inner circumferential mask from a pattern to be inspected in an input image signal and outputs an inner circumferential mask subtraction signal; (B) an outer periphery mask subtraction circuit that creates an inverted image of the pattern to be inspected, subtracts an outer periphery mask from this inverted image, and outputs an outer periphery mask subtraction signal; (C) a thin line obtained by thinning the inner periphery mask subtraction signal; a first feature point signal having thinning pattern information, information as to whether or not end points exist in the thinning pattern, and distance information between the end points if the end points exist, and the outer periphery mask subtraction signal. A feature point that outputs a second feature point signal having thinning pattern information, information as to whether or not an endpoint exists in the thinning pattern, and distance information between the endpoints if the endpoint exists. an extraction circuit, (D) based on the input image signal and the first feature point signal, the thinning pattern is located inside the outer peripheral mask, and the distance between the end points is determined from the diameter of the pattern to be inspected; (E) an outer circumferential mask confirmation circuit that recognizes a chipping when it is less than a threshold value and outputs a chipping detection signal; an inner circumferential mask confirmation circuit that recognizes a protrusion and outputs a protrusion detection signal if the distance between the end points is less than or equal to a threshold determined from the diameter of the pattern to be inspected. .

[0005]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the present invention will be explained in detail with reference to the drawings.

FIG. 1 is a block diagram showing one embodiment of the present invention. The pattern inspection apparatus shown in FIG. 1 (A) subtracts an inner circumferential mask from a pattern to be inspected in an input image signal;
Inner circumference mask subtraction circuit 1 that outputs inner circumference mask subtraction signal i
, (B) an outer periphery mask subtraction circuit 2 that creates an inverted image of the pattern to be inspected, subtracts an outer periphery mask from this inverted image, and outputs an outer periphery mask subtraction signal l; (C) an inner periphery mask subtraction signal i as a thin line; a first feature point signal j having thinning pattern information, information as to whether or not the thinning pattern has endpoints, and, if the endpoint exists, distance information between the endpoints; and a peripheral mask. a second feature point signal m having thinning pattern information obtained by thinning the subtraction signal l, information as to whether or not an end point exists in the thinning pattern, and information on the distance between the end points if the end point exists. (D) Based on the input image signal and the feature point signal j, the thinning pattern is inside the outer peripheral mask, and the distance of the end point is from the diameter of the pattern to be inspected. If it is less than a predetermined threshold value, it is recognized as a chip and outputs a chip detection signal k. an inner circumferential mask confirmation circuit 5 that recognizes a protrusion if it is located inside the circumferential mask and the distance between the end points is less than a threshold value determined from the diameter of the pattern to be inspected, and outputs a protrusion detection signal n; Consists of.

FIG. 2 is a plan view showing a mask used to extract pattern defects and protrusions. Inner circumference mask M
1 is a circular shape whose diameter is several pixels shorter than the pad diameter 6 of the pattern. The subtraction by the inner mask subtraction circuit 1 sets the area where the pattern and the inner mask M1 overlap to 0, the area containing only the inner mask M1 to 0, the area containing only the pattern to 1,
The background is defined as being converted to 0.

FIG. 3 shows an inner circumferential mask M1 for an input image.
, is a schematic diagram showing how defects are extracted while sequentially raster scanning the outer circumferential mask M2 from the upper left.

FIGS. 4(a) to 4(e) are schematic diagrams showing details of the operation of FIG. 3. 4(a) shows a pattern with a chip, FIG. 4(b) shows the inner peripheral mask M1 located inside the pattern shown in FIG. 4(a), and FIG. 4(c) shows the pattern shown in FIG. 4(b) shows how the inner circumferential mask M1 is subtracted from the pattern image, and FIG. 4(d) shows how the subtracted image shown in FIG. 4(c) is thinned and end points K1 and K2 are extracted. The outer periphery mask subtraction circuit 2 receives the binary image signal h. The outer circumferential mask subtraction circuit inverts 1 and 0 of the input image, then subtracts the image of the outer circumferential mask M2 shown in FIG. 2, and outputs an outer circumferential mask subtraction signal l. Further, subtraction in this case is defined as converting data 1 outside the outer peripheral mask M2 in the inverted image of the pattern to 0. Figure 7 (a
) shows an inverted image of a pattern with protrusions. FIG. 7(b) shows that the pattern shown in FIG. 7(a) is located inside the outer peripheral mask M2. Figure 7(c)
is a diagram obtained by subtracting the outer circumferential mask M2 from the pattern image shown in FIG. 7(b).

The feature point extraction circuit 3 receives the inner circumference mask subtraction signal i and the outer circumference mask subtraction signal l, performs line thinning processing on each of the input subtraction images, and then detects end points.
Output feature point signals j and m. Thinning refers to the process of converting a figure into skeleton lines, and is used as an effective means for detecting feature points of a figure. The number of times of line thinning is set to such an extent that the subtracted image when the centers of the inner circumferential mask M1 and outer circumferential mask M2 are located at the center of the pattern has the width of a pixel due to line thinning. The endpoint and the number of connections when considering the image in pixel units are 1
Refers to the pixel that is . The end point T in FIG. 10 is an example of an end point.

In FIG. 4(d), the subtracted image of the chipped pattern shown in FIG. 4(c) is thinned and detected as end points K1 and K2. In a normal pattern without chipping, no end points are detected when the inner circumferential mask M1 is inside the pattern.

FIG. 7(d) is a diagram in which the subtracted image of the pattern with protrusions shown in FIG. 7(c) is thinned and end points T1 and T2 are extracted. As shown in the figure, both ends of one protrusion are detected as end points T1 and T2. For a normal pattern without protrusions, no end points are detected when the pattern is inside the outer mask.

The outer periphery mask confirmation circuit 4 inputs the binary image signal h and the feature point signal j, and outputs a chipping detection signal k. The outer circumferential mask confirmation circuit 4 first uses the outer circumferential mask M2 to confirm whether patterns are scattered inside the outer circumferential mask M2. The outer periphery mask M2 takes into consideration the pad diameter of the pattern, and the outer circumference 1 of a circle whose diameter is several pixels longer than the pad diameter.
It is an image of pixels. If no pattern exists on the pixel of the outer circumferential mask M2, it is determined that a pattern exists inside the outer circumferential mask M2, or it is determined that no pattern exists at all. If this is not the case, it is determined that a pattern exists covering the outer periphery mask M2.

If a pattern exists over the outer circumferential mask M2, the end points detected by the feature point extracting circuit 3 are excluded from candidate points for chipping. In FIG. 4(e), the outer peripheral mask M2
Since no pattern exists above, the end points K1 and K2 are left as candidate points for chipping. As shown in FIG. 5A, if the inner circumferential mask M1 does not exist inside the pattern, non-missing portions are also detected as end points K11 and K22. In this case, since a pattern exists on the outer circumferential mask M2 as shown in FIG. 5(b), the end points K11 and K22 are excluded as candidate points.

Next, the outer circumferential mask confirmation circuit 4 calculates the distance between the end points K1 and K2, which are the remaining candidate points, and if the distance approximately corresponds to the diameter of the inner circumferential mask M1, they are removed from the candidate points of chipping. . FIG. 4(f) shows the distance d between the end points K1 and K2, and since the distance d is not equal to the diameter of the inner peripheral mask M1, it is determined that there is a chip. FIG. 6A shows a case where the inner circumferential mask M1 deviates slightly from the pattern, and the portions that are not chipped are detected as end points K111 and K222. As shown in FIG. 6(b), since the inner circumferential mask M1 and the outer circumferential mask M2 are only slightly shifted from the pattern and the center, there is no pattern on the outer circumferential mask M2, so the end point K1
11, K222 remains as a candidate point for chipping. but,
When the distance d between the end points K111 and K222 is determined, it is approximately equal to the diameter of the inner circumferential mask M1 since the inner circumferential mask M1 is only slightly shifted from the pattern. Therefore, the end point K11
1, K222 is excluded from the candidate points for chipping. The outer periphery mask confirmation circuit 4 has terminal points K1, K2, K11, K22,
Of K111 and K222, end points K1 and K2 caused by true chipping are output as chipping detection signals k.

The inner circumference mask confirmation circuit 5 inputs the binary image signal h and the feature point signal m, and outputs a protrusion detection signal n. The inner circumferential mask confirmation circuit first uses the inner circumferential mask M1 to confirm whether the inner circumferential mask M1 exists inside the pattern. The inner circumferential mask M1 is an image of one pixel around a circle whose diameter is several pixels shorter than the pad diameter, taking into account the pad diameter of the pattern. If the pattern does not exist on the pixels of the inner circumferential mask M1, it is determined that the inner circumferential mask M1 exists inside the pattern or completely outside the pattern. If not, the inner circumference mask M
It is determined that it is multiplied by 1 and exists.

If a pattern exists over the inner circumferential mask M1, the end points detected by the feature point extraction circuit 3 are excluded from the candidate points of the protrusion. In FIG. 7(e), the inner peripheral mask M1
Since there is no pattern in , the end points T1 and T2 are left as candidate points for protrusions. As shown in FIG. 8A, if the pattern does not exist inside the outer peripheral mask M2, portions that are not protrusions are also detected as end points T11 and T22. in this case,
As shown in FIG. 8(b), since a pattern exists on the inner circumferential mask M1, T11 and T22 are excluded as candidate points.

Next, the inner circumferential mask confirmation circuit 5 calculates the distance between the end points T1 and T2, which are the remaining candidate points, and if the distance approximately corresponds to the diameter of the outer circumferential mask M2, they are removed from the candidate points of the protrusion. FIG. 7F shows the distance d between the end points T1 and T2, and since the distance d is not equal to the diameter of the outer circumferential mask M2, it is determined that it is a protrusion. FIG. 9(a) shows the outer peripheral mask M
2 shows a case where there is a slight deviation from the pattern, and portions that are not protrusions are detected as end points T111 and T222. As shown in FIG. 9(b), the outer circumferential mask M2 and the inner circumferential mask M1 are only slightly shifted from the center of the pattern, so there is no pattern on the inner circumferential mask M1, so the end points T111 and T222 are candidate points for protrusions. remains as. However, when the distance d between the end points T111 and T222 is determined, the distance d between the end points T111 and T222 is approximately equal to the diameter of the outer circumferential mask M2, since the outer circumferential mask M2 is only slightly shifted from the pattern. Therefore, the end points T111 and T222 are excluded from the candidate points of the protrusion. The inner circumference mask confirmation circuit 5 checks the end points T1, T2, T11,
Among T22, T111, and T222, the end points T1 and T2 caused by the true protrusion are output as the protrusion detection signal n.

[0019]

[Effects of the Invention] The pattern inspection device of the present invention detects chips and protrusions by performing calculations on pattern images and mask images and extracting feature points, so a good pattern dictionary is not required. This has the advantage that inspection can be performed without requiring a large amount of hardware for precise alignment with a comparison pattern.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing one embodiment of the present invention.

FIG. 2 is a plan view showing a mask used to extract pattern defects and protrusions.

FIG. 3 is a schematic diagram showing how defects are extracted while sequentially raster scanning an input image with an inner circumferential mask M1 and an outer circumferential mask M2 from the upper left.

FIG. 4 is a schematic diagram showing details of the operation of FIG. 3 for a pattern having a chip.

FIG. 5 is a schematic diagram showing an example of excluding end points extracted by a feature point extraction circuit from candidate points for chipping.

FIG. 6 is a schematic diagram showing another example of excluding end points extracted by the feature point extraction circuit from candidate points for chipping.

FIG. 7 is a schematic diagram showing details of the operation of FIG. 3 for a pattern having protrusions.

FIG. 8 is a schematic diagram showing an example of excluding an end point extracted by a feature point extraction circuit from candidate points of a protrusion.

FIG. 9 is a schematic diagram showing another example of excluding end points extracted by a feature point extraction circuit from candidate points of protrusions.

FIG. 10 is a schematic diagram showing end points.

[Explanation of symbols]

1 Inner mask subtraction circuit 2. Outer mask subtraction circuit 3 Feature point extraction circuit 4 Outer mask confirmation circuit 5 Inner mask confirmation circuit h Binarized image signal i Inner mask subtraction signal j Feature point signal k Missing detection signal l Outer mask subtraction signal m Feature point signal n Protrusion detection signal M1 Inner circumference mask M2 Outer mask

Claims (1)

[Claims] 1. (A) An inner circumferential mask subtraction circuit that subtracts an inner circumferential mask from the inspected pattern in an input image signal and outputs an inner circumferential mask subtraction signal; (B) Creates an inverted image of the inspected pattern. an outer circumferential mask subtraction circuit that subtracts an outer circumferential mask from this inverted image and outputs an outer circumferential mask subtraction signal; (C) thinning pattern information obtained by thinning the inner circumferential mask subtraction signal; a first feature point signal having information as to whether or not the end points exist, and distance information between the end points if the end points exist; thinning pattern information obtained by thinning the outer periphery mask subtraction signal; a feature point extraction circuit that outputs a second feature point signal having information as to whether or not an end point exists in the pattern, and distance information between the end points if the end point exists; (D) the input image signal and the Based on the first feature point signal, if the thinned pattern is inside the outer peripheral mask and the distance between the end points is less than or equal to the threshold determined from the diameter of the pattern to be inspected, it is recognized as a chip. an outer circumferential mask confirmation circuit that outputs a detection signal; (E) based on the input image signal and the second feature point signal, the thinning pattern is inside the inner circumferential mask, and the distance between the end points is within the covered area; A pattern inspection device comprising: an inner circumferential mask confirmation circuit that recognizes a protrusion if it is less than a threshold value determined from the diameter of the inspection pattern and outputs a protrusion detection signal.