JP3038176B2 - Inspection equipment by pattern matching - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an inspection apparatus for inspecting an inspection object such as an IC by a pattern matching method.

[0002]

2. Description of the Related Art Conventionally, for example, in a visual inspection for inspecting a character drawn on the surface of an IC as shown in FIG. 7, a pattern matching method is often used. In the pattern matching method, a registered pattern is created based on an image taken from a non-defective IC, and the registered pattern and the IC of the inspection object are created.
If the degree of coincidence between the two is higher than a predetermined value, the IC of the inspected object is determined to be good, and if the degree of coincidence is lower than the predetermined value, the IC of the inspected object is defective. Image processing method.

In general, a plurality of inspection patterns are created based on an image taken from an IC of one inspection object, and the plurality of patterns whose matching degree between a registered pattern and the inspection pattern exceeds a predetermined value is used. If there is at least one, the IC of the inspection object is regarded as a non-defective product.

The degree of coincidence between the two can be calculated by various methods. One of them is a calculation method using normalization such as normalized correlation. The calculation method using this normalization is suitable for obtaining a stable result because it is not greatly affected by the fluctuation of the brightness of the illumination illuminating the inspection object.

However, when the degree of coincidence of pattern matching is obtained by a calculation method using normalization, for example, even when a character drawn on the surface of the IC of the inspection object is blurred, the character is not blurred. The degree of coincidence is calculated with little effect. In this case, when it is desired to make the blur of the character defective, the defect cannot be detected.
This is inconvenient for the appearance inspection of C.

[0006]

As described above, conventionally,
Although the degree of coincidence of pattern matching was obtained by a calculation method using normalization, this method is not greatly affected by fluctuations in the brightness of the illumination illuminating the inspection object, that is, the density of the image of the inspection object. For example, there is a drawback that a result in consideration of blurring of characters drawn on the surface of the IC of the inspection object cannot be obtained.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned drawbacks, and has as its object to propose a pattern matching method in consideration of density information capable of obtaining a result in consideration of blurring of characters of an inspection object. That is.

[0008]

In order to achieve the above object, an inspection apparatus of the present invention comprises: means for creating a registration pattern of N × M pixels; means for creating an inspection pattern of N × M pixels; The pattern matching between the registered pattern and the inspection pattern is expressed by the following equation.

[0009]

(Equation 7) And the result of the pattern matching PTN
And a means for judging the quality of the inspection pattern based on the result PTN of the pattern matching.

[0010] The inspection apparatus of the present invention comprises means for creating a registration pattern of N × M pixels, means for creating an inspection pattern of N × M pixels, and pattern matching between the registration pattern and the inspection pattern by the following equation.

[0011]

(Equation 8) And the result of the pattern matching PTN
And a means for judging the quality of the inspection pattern based on the result PTN of the pattern matching.

In the recording medium of the present invention, a step of creating a registration pattern of N × M pixels, a step of creating an inspection pattern of N × M pixels, and a pattern matching between the registration pattern and the inspection pattern are performed as follows. formula

[0013]

(Equation 9) And the result of the pattern matching PTN
And a result PT of the pattern matching
Determining the quality of the test pattern based on N.

In the recording medium of the present invention, a step of creating a registration pattern of N × M pixels, a step of creating an inspection pattern of N × M pixels, and a pattern matching between the registration pattern and the inspection pattern are performed as follows. formula

[0015]

(Equation 10) And the result of the pattern matching PTN
And a result PT of the pattern matching
Determining the quality of the test pattern based on N. According to the pattern matching method of the present invention, the pattern matching between the registration pattern of N × M pixels and the inspection pattern of N × M pixels is represented by the following equation.

[0016]

[Equation 11] Perform based on. The pattern matching method of the present invention comprises:
The pattern matching between the registration pattern of N × M pixels and the inspection pattern of N × M pixels is represented by the following equation.

[0017]

(Equation 12) Perform based on.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an inspection apparatus using pattern matching according to the present invention will be described in detail with reference to the drawings. FIG. 1 schematically shows the configuration of the inspection apparatus of the present invention.

The camera 10 collects an image of a subject (good IC or IC of the inspection object) 16 and supplies the image to the image processing device 11. The camera 10 includes a video camera, a digital camera, and the like, and includes an image capturing device such as a scanner.

The image processing device 11 includes an image memory unit 12,
It has a processing unit 13 and a control unit 14. Images of non-defective ICs or inspected ICs are stored in the image memory unit 12 respectively.
Is stored. The processing unit 12 creates a registration pattern based on the image of the non-defective IC, and creates an inspection pattern based on the image of the IC of the inspection object. Further, the processing unit 12 performs pattern matching between the registered pattern and the inspection pattern based on a pattern matching method described later.
The control unit 14 controls input and output of image data in the image memory unit 12 and calculation in the processing unit 13.

The illumination device 15 plays a role of irradiating scattered light to illuminate the subject 16. Next, the operation (pattern matching method) of the inspection apparatus of the present invention will be described with reference to FIG.

Although the present invention can be applied to various inspections by image processing, in the present embodiment, in order to make the explanation easy to understand, a case where characters attached to the surface of an IC (package) are inspected will be described. I do.

First, as shown in FIG. 3, a non-defective IC image is fetched and registered in the image memory as a reference image 21A. In addition, an area including substantially the entire surface of the IC (package) in the reference image 21A is registered as a registration pattern 22A. Registration pattern 2
2A has a size of N × M pixels as shown in FIG. 4 (steps ST1 and ST2).

Next, as shown in FIG. 5, an image of an IC to be inspected is captured and registered in an image memory as an inspection image 21B. In addition, IC of the inspection image 21B
A region including substantially the entire surface portion of (package) is registered as the inspection pattern 22B. Inspection pattern 22B
Has the same size of N × M pixels as the registered pattern, as shown in FIG. 6 (steps ST3 to ST4).

Next, pattern matching is performed on the registered image and the inspection image (step ST5). In the present invention,
The pattern matching is performed by the following two methods while considering the density information of each pixel of the registered image and the inspection image.

In the first method, the density of each pixel of the registration pattern (N × M pixels) is set to R (I, J), and the inspection pattern (N
× M pixels), the density of each pixel is defined as P (I, J), and the pattern matching result P
TN.

[0027]

(Equation 13)

I = 0, 1,..., N-1 J = 0, 1,..., M-1 where the parameter s is a positive real number. Since the pattern matching result PTN takes a value from −1 to 1, it is normalized in the same manner as the result of pattern matching by normal normalization (however, density information of each pixel is considered. ).

In the second method, the density of each pixel of the registered pattern (N × M pixels) is R (I, J), and the average density is Rav.
g, the density of each pixel of the inspection pattern (N × M pixels) is P
When (I, J) and the average density are defined as Pavg, PTN is obtained as a result of pattern matching by the following equation (2).

[0030]

[Equation 14]

I = 0, 1,..., N-1 J = 0, 1,..., M-1 where the parameter s is a positive real number. Since the pattern matching result PTN takes a value from −1 to 1, it is normalized in the same manner as the result of pattern matching by normal normalization (however, density information of each pixel is considered. ).

Therefore, the quality of the inspection object can be determined based on the PTN calculated by one of the above two methods (step ST6). If the range of the PTN as a result of the pattern matching is to be set to a range from 0 to 1, the following (3)
The calculation shown in the equation may be performed. PTN ′ = (PTN + 1) / 2 (3)

[0033]

As described above, according to the inspection apparatus using pattern matching of the present invention, pattern matching is performed in consideration of density information. Therefore, IC
It is very suitable for inspecting unevenness in density (uniformity of density in a predetermined area) of an inspection object, such as inspection of characters on the surface of a (package). That is, in the pattern matching in consideration of the density information, the result of the pattern matching becomes small when there is a faint character or the like, and the possibility of being determined to be defective is increased.

[Brief description of the drawings]

FIG. 1 is a view schematically showing a configuration of an inspection apparatus according to the present invention.

FIG. 2 is a view showing the operation of the inspection device of FIG. 1;

FIG. 3 is a view showing a reference image.

FIG. 4 is a diagram showing a registration pattern.

FIG. 5 is a diagram showing an inspection image.

FIG. 6 is a diagram showing an inspection pattern.

FIG. 7 is a diagram showing an example of characters on the surface of an IC.

[Explanation of symbols]

 10: Camera, 11: Image processing device, 12: Image memory unit, 13: Processing unit, 14: Control unit, 15: Illumination device, 16: Subject, 21A: Reference image, 22A: Registration pattern, 21B: Inspection image, 22B: Inspection pattern.

Claims (6)

(57) [Claims] 1. A means for creating a registration pattern of N × M pixels, a means for creating an inspection pattern of N × M pixels, and a pattern matching between the registration pattern and the inspection pattern are represented by the following equation: And the result of the pattern matching PTN
And a means for judging the quality of the test pattern based on the PTN result of the pattern matching. 2. A means for creating a registration pattern of N × M pixels, a means for creating an inspection pattern of N × M pixels,
The pattern matching between the registered pattern and the inspection pattern is expressed by the following equation: And the result of the pattern matching PTN
And a means for judging the quality of the test pattern based on the PTN result of the pattern matching. 3. A step of creating a registration pattern of N × M pixels, a step of creating an inspection pattern of N × M pixels, and a pattern matching between the registration pattern and the inspection pattern are represented by the following equations. And the result of the pattern matching PTN
And a result PT of the pattern matching
Determining a quality of the inspection pattern based on N. 4. A step of creating a registration pattern of N × M pixels, a step of creating an inspection pattern of N × M pixels, and a pattern matching between the registration pattern and the inspection pattern are represented by the following equations. And the result of the pattern matching PTN
And a result PT of the pattern matching
Determining a quality of the inspection pattern based on N. 5. A pattern matching between a registration pattern of N × M pixels and a test pattern of N × M pixels is represented by the following equation. A pattern matching method characterized in that the pattern matching method is performed based on 6. A pattern matching between a registration pattern of N × M pixels and a test pattern of N × M pixels is represented by the following equation. A pattern matching method characterized in that the pattern matching method is performed based on