JPH0582618A - Image data defect detector - Google Patents

Abstract

PURPOSE:To provide an image data defect detector which does not require a large quantity of judging data and avoids a false detection due to the deviation in location of image data. CONSTITUTION:A comparison circuit 4 compares the image data from a standard variable-density image data storage circuit and the image data from an input device 3 for each picture element, outputting differential image data to a labeling circuit 6 through a binary circuit 5. The labeling circuit 6 extracts picture elements most suitable for defects and outputs the data about the location of those picture elements as labeling image data to a distributed value selection circuit 10. Meanwhile, a distributed value of the image data inputted into a storage circuit 7 is calculated by a distributed value calculating circuit 8. Based on the labeling image data, the distributed value selection circuit 10 selects the distributed value for the picture element which has been decided most suitable for defects and a judging circuit 12 determines whether the selected distributed value is within the range of defects stored in a register 11.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image data defect detecting device for detecting a defect in appearance of a printed circuit board on which a semiconductor chip is mounted, a lead frame or the like.

[0002]

2. Description of the Related Art FIG. 3 is a schematic block diagram of a conventional image data defect detecting apparatus. The conventional image data defect detection apparatus includes a reference grayscale image data storage circuit 31 in which image data of a printed circuit board having no defect is stored in advance, an input device 33 for inputting image data of a pattern 32 to be inspected, and a comparison circuit 34. A defect determination circuit 35, a determination data storage circuit 36, and an inspection result display section 37.

The comparison circuit 34 compares the image data from the reference grayscale image data storage circuit 31 and the image data from the input device 33 for each pixel, and changes the difference between the two image data into difference image data. The difference image data is output to the defect determination circuit 35. The defect determination circuit 35 reads the determination data (threshold value) for determining whether the product is a non-defective product or a defective product from the determination data storage circuit 36, compares the difference image data for each pixel with the determination data, and exceeds the value of the determination data. Regarding the difference image data, it is determined that the pixel has a defect. Then, the judgment result is displayed on the inspection result display unit 37.

[0004]

As described above, the prior art is as follows.
Judgment data (threshold value) to detect defects in image data
However, since the difference image data and the judgment data are compared for each pixel, the judgment data needs to have the same amount of information as the image data of the reference grayscale image data storage circuit 31. Therefore, a large capacity memory is required to store the determination data.

Further, the alignment of the image data from the reference grayscale image data storage circuit 31 and the image data from the input device 33 is extremely difficult because one pixel is usually as small as about 5 μm on each side. For this reason, even if the image data is a good product, when the positional deviation occurs when the image data from the reference grayscale image data storage circuit 31 is compared, the difference image data corresponding to the contour portion of the wiring pattern or the like becomes large. , May be certified as defective. As described above, the conventional apparatus has a problem in that the positional deviation causes an erroneous detection and the defect detection accuracy is not good.

The present invention has been made in view of the above circumstances, and provides an image data defect detecting apparatus which does not require a large amount of determination data and can prevent erroneous detection due to positional deviation of image data. The purpose is.

[0007]

In order to achieve the above object, an image data defect detecting apparatus according to the present invention has, for each pixel of reference image data, a degree of change in brightness within a certain area including the pixel. A first variance value calculating means for calculating a variance value, a storage means for storing an upper limit and a lower limit of the variance value for judging an image as a defect based on the variance value, and image data of an inspection pattern. Second dispersion value calculating means for calculating, for each pixel, a dispersion value indicating the degree of change in brightness within a certain area including the pixel; and a dispersion value for which the dispersion value of the pattern to be inspected is stored in the storage means. It is characterized in that it is provided with a judging means for judging whether it is within the range between the upper limit and the lower limit.

[0008]

According to the present invention, with the above-described structure, as a determination target when detecting a defect in image data, a variance value indicating the degree of change in brightness for each pixel of image data within a certain area including the pixel. Is used to determine whether the variance value for each pixel of the image data of the pattern to be inspected is within the defect range, and to determine the defect of the image data. This eliminates the need to store judgment data for each pixel,
It is sufficient to store only the upper limit and the lower limit of the variance value commonly used for all pixels as the determination data. Further, by setting the upper limit of the variance value of the determination data to a value smaller than the variance value of the image due to the positional deviation, it is possible to prevent erroneous detection due to the positional deviation.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic block diagram of an image data defect detecting apparatus which is an embodiment of the present invention, and FIG. 2 is a diagram for explaining a variance value calculating method.

The image data defect detection apparatus shown in FIG. 1 includes a reference grayscale image data storage circuit 1, an input device 3, a comparison circuit 4, a binarization circuit 5, a labeling circuit 6, a storage circuit 7, and a storage circuit 7. Distributed value calculation circuits 8 and 9 and distributed value selection circuit 10
The register 11, the determination circuit 12, and the inspection result display unit 13 are provided.

The reference grayscale image data storage circuit 1 is, for example, a memory for storing image data of a printed circuit board having no defect, and the input device 3 is composed of a CCD camera or the like. Image data is obtained from the inspection pattern 2.

The comparison circuit 4 compares the image data from the reference grayscale image data storage circuit 1 with the image data from the input device 3 for each pixel, and changes it into difference image data which is the difference between the two image data. The binarization circuit 5 sets a binarization level in advance and converts the difference image data into binary image data based on this binarization level. The labeling circuit 6 performs a labeling process on the binary image data. As a result, a grayscale difference having a certain magnitude or more is extracted as a defect candidate, and pixel position data for the defect candidate is output as labeling image data.

By the way, when the difference image data is obtained in the comparison circuit 4, generally, the image data from the reference grayscale image data storage circuit 1 and the image data from the input device 3 cannot be perfectly aligned. Therefore, even if the image data is a good product, the difference image data corresponding to the contour portion of the wiring pattern or the like may become large due to the positional deviation. Therefore, the above-mentioned labeling image data includes the position data of the pixel in which the gray level difference is caused due to the true defect and the position data of the pixel erroneously detected due to the position shift.

The variance value calculation circuits 8 and 9 calculate the variance value for each pixel of the image data from the input device 3 and the image data from the reference grayscale image data storage circuit 1, respectively. The variance value indicates, for each pixel, the degree of change in brightness in a certain area including the pixel.

For example, when one pixel is a square having a side of 5 μm as shown in FIG. 2, the pixel having the Nth in the vertical direction and the Mth in the horizontal direction is represented by (N, M). Now, a small area S of a pixel × b pixel size including one pixel P (n, m) is taken, and the brightness of any one pixel P _S (i, j) in this area S is _{En, ij} . At this time, the average value E _{nm , ij} of E _{nm, ij} for the pixel P (n, m)
Is given by the following formula.

[0016]

[Equation 1]

The variance value s for the pixel P (n, m)
_nm ² is given by the following formula.

[0018]

[Equation 2]

In the equations (1) and (2), i,
The sum for j is taken over the set of (i, j) that is in the region S.

Generally, if the dispersion value s _nm ² is large, it is judged that the degree of change in brightness in the minute area S is large, whereas if the dispersion value s _nm ² is small, the brightness in the minute area S is small. It is determined that the degree of change is small and the brightness is uniform. Specifically, the dispersion value s _nm ² becomes small when the region S is taken on a flat surface such as a substrate or a wiring pattern,
The area S becomes large when the area S is taken as a contour portion such as a wiring pattern or an uneven portion due to a defect.

When the minute area S is extremely large, the dispersion value does not change so much for each pixel, and when the minute area S is extremely small, the dispersion value greatly changes for each pixel. Therefore, unless the minute area S is set in an appropriate range, it is not possible to accurately determine whether or not the image is defective by the variance value. In this embodiment, the minute area S is set to the pixel P.
The size was set to 5 × 5 centered on.

In the variance value selection circuit 10, the variance value for the pixel which is a defect candidate is selected from the variance values for each pixel calculated by the variance value calculation circuit 8 based on the labeling image data from the labeling circuit 6. Selected.

On the other hand, in the register 11, the defect range of the common dispersion value is set for all pixels based on the dispersion value for each pixel calculated by the dispersion value calculation circuit 9. The defect range is a criterion for judging whether or not the image is defective from the variance value, and the upper limit of the defect range is set to a value slightly smaller than the variance value in the contour portion of the part. this is,
This is because, in general, the variance value in the contour portion of the component is larger than the variance value in the uneven portion due to the defect, so that the defect candidate due to the positional deviation in the labeling circuit 6 is removed. Further, the lower limit of the dispersion value is set, because when the dispersion value is smaller than this lower limit, it is considered that the dispersion value is within an allowable range rather than a defect that must be removed.

Next, the operation of detecting a defect in image data will be described. First, the image data from the input device 3 is output to the comparison circuit 4 and the storage circuit 7. The comparison circuit 4 compares this image data with the image data from the reference grayscale image data storage device 1 and outputs difference image data, which is the difference between the two image data, to the binarization circuit 5. The binarization circuit 5 converts the difference image data into binary image data, and the labeling circuit 6
Output to. Then, the labeling circuit 6 performs the labeling process on the binary image data, and obtains the position data for the defect candidate as the labeling image data. It should be noted that since only the grayscale difference between the two image data is equal to or larger than a certain size is extracted by the labeling process, the labeling circuit 6
Also plays the role of removing noise.

The variance value of the image data stored in the storage circuit 7 from the input device 3 is calculated in the variance value calculation circuit 8. When the labeling image data obtained by the labeling circuit 6 is output to the variance value selection circuit 10, the variance value selection circuit 10 selects and fetches the variance value of the pixel which is the defect candidate. On the other hand, the image data from the reference grayscale image data storage circuit 1 is sent to the variance value calculation circuit 9 to calculate the variance value. The defect range (upper and lower limits of the variance value) common to all pixels is stored in the register 11 based on the variance value calculated here.

Then, the judgment circuit 12 judges whether or not the dispersion value selected by the dispersion value selection circuit 10 is within the defect range stored in the register 11. If the variance value is within this range, it is regarded as a true defect, and the pixel at which position is defective is displayed on the inspection result display unit 13.

In the present embodiment, in order to detect a defect in the image data, a variance value indicating the degree of change in brightness in a certain area including the pixel is used for each pixel, and the image data from the input device is used. Determine if the variance value for is within the defect range. Since the variance value is used as a target for determining a defect in this manner, it is not necessary to set determination data for each pixel, and the upper and lower limits of the defect range can be commonly used in the entire image. Therefore, unlike the conventional case, a large amount of information is not required as the determination data, and a register or the like can be used as a memory for storing the determination data,
The cost can be reduced.

Further, the defect is not judged based on the difference image data obtained from the comparison circuit, but it is judged whether or not the variance value of the image data of the pattern to be inspected is within a certain range. Therefore, by setting the upper limit value of the reference variance value to an appropriate value, it is possible to prevent the image data due to this positional deviation from being determined as defective data even if the positional deviation occurs. .. Therefore, it is possible to prevent erroneous detection due to positional deviation and improve the accuracy of defect detection.

Further, in this embodiment, instead of detecting whether or not the variance value is within the defect range for all pixels of the image data, a labeling circuit is provided to extract some of the pixels as defect candidates. Therefore, it is only necessary to determine whether or not the variance value is within the defect range for only some pixels, and thus it is possible to improve the processing speed.

[0030]

As described above, according to the present invention, as a target for determining whether or not image data is defective, for each pixel, the degree of change in brightness in a certain area including the pixel is determined. By using the variance value shown, the data common to all pixels can be used as the judgment data, so that a small capacity memory such as a register can be used as the memory for storing the judgment data, and the position shift can be made. Since it is possible to prevent erroneous detection due to, it is possible to provide an image data defect detecting apparatus capable of improving the accuracy of defect detection.

[Brief description of drawings]

FIG. 1 is a schematic block diagram of an image data defect detection apparatus that is an embodiment of the present invention.

FIG. 2 is a diagram for explaining a calculation method of a variance value.

FIG. 3 is a schematic block diagram of a conventional image data defect detection device.

[Explanation of symbols]

 1 reference grayscale image data storage circuit 2 inspected pattern 3 input device 4 comparison circuit 5 binarization circuit 6 labeling circuit 7 storage circuit 8, 9 variance value calculation circuit 10 variance value selection circuit 11 register 12 judgment circuit 13 inspection result display section

Claims (2)

[Claims] 1. A first variance value calculating means for calculating, for each pixel of reference image data, a variance value indicating a degree of change in brightness within a certain area including the pixel, and based on the variance value. Storage means for storing an upper limit and a lower limit of a variance value for determining an image as a defect, and a variance value for each pixel of image data of a pattern to be inspected, which indicates a degree of change in brightness within a certain area including the pixel A second variance value calculating means for calculating the variance value, and a determining means for determining whether or not the variance value of the pattern to be inspected is within an upper limit and a lower limit of the variance value stored in the storage means. An image data defect detection device characterized by: 2. Comparing means for comparing the lightness and darkness of each pixel of the reference image data and the image data of the pattern to be inspected to output difference image data, and binary image data obtained by binarizing the difference image data. Binarizing means, a labeling means for labeling the binary image data and outputting labeling image data, and a defect candidate from the second variance value calculating means based on the labeling image data. A variance value selecting means for selecting a variance value for a pixel, and the judging means determines whether the variance value selected by the variance value selecting means is within the upper and lower limits of the variance value stored in the storage means. The image data defect detecting device according to claim 1, wherein the image data defect detecting device determines whether or not.