KR20180022619A - PCB panel inspection method and apparatus - Google Patents

Abstract

Disclosed are a PCB panel inspection method, and an apparatus thereof. The PCB panel inspection method using the panel inspection apparatus comprises: a step of aligning locations of a reference image and an image of a panel to be inspected; a step of extracting an abnormal portion existing in an interest region by using a difference between the reference image and the panel image; a step of identifying a parameter with respect to the abnormal portion existing in the interest region; and a step of determining the abnormal portion to be normal or defective based on similarity between a group including one or more parameters with respect to normality or defect and the parameter with respect to the abnormal portion.

Description

Technical Field [0001] The present invention relates to a PCB panel inspection method and apparatus,

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to a PCB (Printed Circuit Board) panel inspection method and apparatus, and more particularly, to a method and apparatus for automatically checking whether an abnormal region recognized through an optical inspection is an actual defect .

The PCB panel includes a plurality of circuit elements connected in a circuit pattern on an insulating substrate. Since such a PCB substrate is produced in a large quantity by an automated production line, it may cause a disconnection of a circuit pattern or a pattern width defect. Therefore, an operator identifies the PCB pattern individually during the inspection process and displays a defective PCB using marking ink do.

Registration Utility Model No. 20-0375821

An object of the present invention is to provide a method and an apparatus for inspecting a PCB panel that can accurately determine whether an abnormal part existing in a PCB panel is an actual defect by an automated method.

According to another aspect of the present invention, there is provided a method of inspecting a PCB panel, the method comprising: matching a position of a reference image and a panel image to be inspected; Extracting an abnormal region existing in a region of interest through a difference between the reference image and the panel image; Extracting a parameter for an abnormal region existing in the region of interest; And determining a normal or defect of the abnormal region based on the cluster including the at least one parameter for normal or defective and the similarity of the parameter for the abnormal region.

According to an aspect of the present invention, there is provided an apparatus for inspecting a PCB panel, the apparatus comprising: a position matching unit for matching positions of a reference image and a panel image to be inspected; An interest area determination unit that extracts an abnormal region existing in a region of interest through a difference between the reference image and the panel image; A parameter extracting unit for extracting a parameter of an abnormal region existing in the ROI; And a defect judgment part for determining a normal or defect of the abnormal part based on a cluster including at least one parameter for normal or defect and a similarity of the parameter for the abnormal part.

According to the present invention, it is possible to accurately determine whether an abnormal part existing in a PCB panel is an actual defect by an automated method. In addition, it is possible to accurately match the positions of the two images in the pixel unit when the abnormal region is detected by comparing the panel image of the PCB panel to be inspected and the reference image which is the image of the PCB panel having no defect. In addition, when alignment of the panel image and the reference image is performed, alignment marks can be used to minimize the time required for alignment. In addition, since it is learned automatically when the abnormal part is normal and when it is defective, it is possible not only to increase the accuracy of the panel inspection according to the learning time, but also to obtain accurate test results for various abnormal parts.

1 is a view illustrating an example of a PCB panel used in an embodiment of the present invention;
2 is a view illustrating an example of an optical inspection system including a panel inspection apparatus according to an embodiment of the present invention.
3 is a flowchart illustrating an example of a PCB panel inspection method according to an embodiment of the present invention.
4 is a view illustrating an example of a position matching method using an alignment mark according to an embodiment of the present invention.
Figure 5 shows the values of each pixel of the alignment mark boundary,
6 is a diagram illustrating an example of image brightness correction according to an embodiment of the present invention.
7 is a diagram illustrating an example of a position matching method using pixel values according to an embodiment of the present invention.
8 is a view illustrating an example of a method for extracting an abnormal region in a region of interest according to an embodiment of the present invention.
9 is a view illustrating an example of a cluster according to an embodiment of the present invention,
10 is a diagram showing the configuration of an embodiment of a panel inspection apparatus according to the present invention.

Hereinafter, a method and an apparatus for inspecting a PCB panel according to an embodiment of the present invention will be described with reference to the accompanying drawings.

1 is a view illustrating an example of a PCB panel used in an embodiment of the present invention.

Referring to FIG. 1, a PCB panel 100 includes at least one strip 120, and each strip 120 includes at least one unit 110. The unit 110 has circuit patterns and electronic elements. The PCB panel 100 includes at least one alignment mark 130, 132, 134, 136 for aligning the position or orientation of the panel. In this embodiment, the '+' shape is shown as an example of the alignment marks 130, 132, 134, and 136. However, this is an example only and various shapes such as triangles and circles, numbers, symbols, and characters can be implemented. The PCB panel 100 may be composed of one or more layers. PCB panel 9100) is composed of a plurality of layers, it is possible to check whether each layer is defective.

The size of the unit 110 may vary according to the embodiment, but it is several millimeters to several tens of millimeters, so it is difficult for a worker to visually check the defects of circuit patterns or the like existing therein. As an example of the optical inspection method for solving this problem, an image obtained by photographing the PCB panel 100 (hereinafter referred to as a 'panel image') and a PCB panel having no defect are taken, (Hereinafter, referred to as an "inspection target area") by comparing an image (hereinafter, referred to as a "reference image") created by the graphic tool with an inspection target area .

However, even if the operator verifies only the inspection target area detected through the optical inspection, not the entire PCB panel, the inspection time is delayed due to the manual labor of the operator. Accordingly, embodiments of the present invention provide a method and apparatus for determining whether an inspection target area is defective through an automated method.

2 is a diagram illustrating an example of an optical inspection system including a panel inspection apparatus according to an embodiment of the present invention.

Referring to FIG. 2, the optical inspection system includes at least one optical inspection apparatus 200, 202, 204, at least one panel inspection apparatus 210, 212, 214, and a marking apparatus 220.

The optical inspection apparatuses 200, 202 and 204 compares the panel image obtained by photographing the PCB panel 100 with the reference image and grasps the unmatched region as the inspection target region. Here, the panel image may vary according to the embodiment, such as an image of a unit, an image of a strip, or an image of an entire panel. The reference image may also be an image for the entire unit, strip, or panel. For example, the optical inspection apparatuses 200, 202, and 204 can compare the panel image obtained by sequentially photographing each unit 110 of the PCB panel 100 according to a predetermined direction and the reference image to grasp the inspection target area. The optical inspection apparatuses 200, 202 and 204 can generate log data including positional information of the inspection target region (coordinate values, identification information of each unit, etc.) and the like.

The panel inspecting apparatuses 210, 212 and 214 designate a certain region including an inspection target region identified by the optical inspection apparatuses 200, 202 and 204 as a region of interest (ROI) and accurately position the panel image and the reference image The anomalous region in the region of interest is extracted through the difference image of the posterior panel image and the reference image, and whether or not the abnormal region is defective is determined on the basis of the similarity of the cluster to be discussed later.

The PCB panel 100 has areas that have no relation to defects and areas that affect defects in the product. For example, an empty area without a circuit pattern in the unit is irrelevant to whether or not the panel is defective even if foreign matter exists therein. There may also be defective and non-defective areas in the periphery of the strip. Therefore, the panel inspecting apparatuses 210, 212 and 214 refer to the database including the positional information of the defective or non-defective area in the PCB panel 100 and determine whether the inspection target area is defective or not It is possible to perform a process of determining whether or not a defect exists through a cluster similarity or the like.

The marking apparatus 220 marks an abnormal region identified as a defect by the panel inspecting apparatuses 210, 212, 214 using ink, laser, or the like. The marking apparatus 220 can perform marking on an abnormal region as a defect occurrence position. However, if the abnormal region is close to the boundary surface of the unit, the marking region can be spread to other units. Therefore, the marking apparatus 220 can perform marking at a predetermined position in the unit in which the abnormal region exists, rather than the abnormal region. For example, the marking device 220 can perform marking in the center of the unit irrespective of whether the defective part is above or below the unit.

In this embodiment, the optical inspection apparatuses 200, 202, 204 and the panel inspection apparatuses 210, 212, 214 correspond to each other in a one-to-one correspondence. However, as another example, one panel inspection apparatus 210 may be connected to the plurality of optical inspection apparatuses 200, 202, 204. As another example, the panel inspecting apparatuses 210, 212, and 214 and the optical inspecting apparatuses 200, 202, and 204 may be implemented as a single apparatus.

 For example, when the PCT panel is composed of a plurality of layers and each layer is inspected through different optical inspection apparatuses 200, 202 and 204, one panel inspection apparatus 210 is connected to different optical inspection apparatuses 200, It is possible to automatically determine whether the PCB panel is defective after collecting the inspection results of each layer of the PCB panel inspected by the inspection unit. When the log data generation methods of the plurality of optical inspection apparatuses 200, 202, and 204 are different, it is necessary to integrate them, and details thereof are disclosed in Patent Application No. 10-2016-00114457, " Lt; / RTI >

3 is a flowchart illustrating an example of a PCB panel inspection method according to an embodiment of the present invention.

Referring to FIG. 3, the panel inspection apparatus matches the positions of the reference image and the panel image (S300). The size of the circuit pattern of the unit is very small. Therefore, in order to precisely grasp the defects such as the open or short of the circuit pattern by comparing the panel image with the reference image, . Therefore, the panel inspection apparatus of the present embodiment includes a process of accurately matching the positions of the reference image and the panel image on a pixel-by-pixel basis through a difference in alignment mark and / or pixel value, and this will be described again with reference to FIGS. 4 to 7 .

The panel inspection apparatus extracts an abnormal region in the region of interest through a difference image of the positionally matched reference image and the panel image (S310). 2, when the optical inspection apparatus compares the reference image with the panel image to find the inspection target area, the panel inspection apparatus sets a target area having a predetermined size based on the inspection target area, The abnormal region is extracted through the difference image between the reference image and the panel image.

The region of interest may be a certain region including the region to be inspected or a unit or strip where the region to be inspected exists. An example of extracting an abnormal region based on the difference image of the reference image and the panel image for the region of interest is shown in FIG.

When the abnormal region in the region of interest is extracted, the panel inspection apparatus grasps various parameters for the abnormal region (S320). The parameters of the abnormal region recognized by the panel inspection apparatus may vary according to the embodiment, and one example is summarized in the following table.

parameter Whether the lead wire is contacted on both sides Average brightness of the panel image in the area of interest Distribution of panel images in the region of interest Minimum brightness in the area of interest Radius size LBP histogram of contact pixels Maximum distortion position

If the parameter for the abnormal region is detected, the panel inspection apparatus compares the parameter with the cluster including the established normal or combination parameter to determine whether the defect is defective (S330). For example, the panel inspection apparatus compares the similarities of the parameters of the abnormal region and the cluster including the at least one parameter when the abnormal region is determined to be normal or defective, and if the parameter of the abnormal region is similar to the parameter of the defect cluster It is determined that the abnormal region is defective. For example, a panel inspection apparatus can grasp the similarity degree of a cluster using a learning algorithm such as a SVM (Support Vector Machine). Various other similarity comparison algorithms other than the above can be applied to the present embodiment.

The panel inspection apparatus additionally stores the result of the defect judgment in the existing cluster and uses it for the subsequent defect judgment. In this manner, the panel inspection apparatus can repeat the learning process of accumulating the determination result of the abnormal region in the cluster, thereby improving the accuracy of the defect determination using the cluster similarity. An example of a cluster used by the panel inspection apparatus for defects is shown in Fig.

FIG. 4 is a view illustrating an example of a position matching method using an alignment mark according to an embodiment of the present invention, and FIG. 5 is a diagram showing values of pixels of the alignment mark boundary.

Referring to FIGS. 4 and 5, the PCB panel includes at least one alignment mark 402. Since the boundaries of the alignment marks 402 in the high resolution panel image are composed of several pixels rather than being clearly separated by one pixel, the center 406 of the alignment mark 402 depends on how the boundaries of the alignment marks are determined It can be different.

For example, in the mark image 400, if the alignment mark 402 is a first color (e.g., white) and the remaining area 404 is a second color (e.g., black) The values (e.g., brightness values, hue values, saturation values, etc.) of each pixel constituting the boundary of the mark 402 may be different as shown in the first figure 500 of FIG. In this embodiment, the pixel values of the boundaries of the alignment mark 402 are shown to be constantly changed as if they are steps for convenience of explanation, but the pixel values of the actual boundaries may vary irregularly.

The panel inspection apparatus enlarges the mark image 400 of a predetermined size including the alignment mark 402 to obtain the center of the pixel boundary more accurately. For example, the panel inspection apparatus enlarges a mark image 400 of N * N (N is a natural number of 1 or more) pixel size to a mark image 410 of M * M (N> M) pixel size. Each pixel value of the enlarged mark image 410 may be determined by various conventional methods such as an interpolation method based on each pixel value of the original mark image 400. In addition, various conventional methods of enlarging an image can be applied to this embodiment.

Since the boundary of the enlarged mark image 410 is composed of more pixels than the original size mark image (510 in FIG. 5), the magnitude of change between each pixel value is smaller than the magnitude of change between each pixel value of the mark image of the original size Can be small. The panel inspection apparatus includes a curve equation representing a pixel value for each of the boundaries 430, 432, 434, 436, 438, 440, 442, 444 where each line of the rectangle 420 including a part of the alignment mark 412 in the enlarged mark image 410 meets the alignment mark Of 520). The curve equation connecting each pixel value can be generated by various interpolation methods such as spline, bezier curve and the like. Various conventional methods other than obtaining the curve equation can be applied to this embodiment.

The panel inspection apparatus grasps the positions of the centers of the respective boundaries by using the curve equations for the boundaries 430, 432, 434, 436, 438, 440, 442, 444 of the alignment marks in the enlarged mark image 410. For example, when the curvilinear equation is a three-dimensional equation, the panel inspection apparatus can grasp the coordinate value at which the value of the second derivative is 0 as the center value of the boundary.

When the center of each of the boundaries 430, 432, 434, 436, 438, 440, 442, 444 of the alignment mark is grasped, the panel inspection apparatus grasps the center 406 of the alignment mark based on the center. For example, the panel inspecting apparatus can determine the center of a line connecting the center values of two boundaries (430 vs 432, 434 vs 436, etc.) located on each line of the rectangle 420, As shown in FIG.

The panel inspection device aligns the center of the alignment mark of the reference image with the alignment mark of the panel image to match the positions of the two images.

6 is a diagram illustrating an example of image brightness correction according to an embodiment of the present invention.

Referring to FIG. 6, histograms of the brightness values of the reference image and the panel image are respectively shown. In this embodiment, it is assumed that the reference image and the panel image are binary images composed of white and black. When the brightness values of the reference image and the panel image are different, it is preferable that the brightness value of the image is corrected before extracting the abnormal region because it may affect the abnormal region extraction to be discussed later.

For example, when the position (A ', B') of the maximum value of the white and black histograms 610 of the panel image is the position (A, B) of the maximum value of the white and black histogram 600 of the reference image , The panel inspection apparatus corrects the brightness values of the white and black colors of the panel image 620 to match the reference image.

7 is a diagram illustrating an example of a position matching method using pixel values according to an embodiment of the present invention.

Referring to FIG. 7, the panel inspecting apparatus calculates the sum of the differences between the respective pixels of the corresponding regions (710 vs 730, 710, 732, 710, and 734, etc.) of the reference image and the pixel image to be minimum The position of the reference image or pixel image is shifted in pixel units. Where the pixel value may be a brightness value, a hue value, or a saturation value. Also, according to an embodiment, the reference image and the pixel image may be binarized images.

The panel inspection apparatus may sum the difference between the values of each pixel with respect to the whole of the reference image and the pixel image. However, since the reference image and the pixel image are high-resolution images, a time delay may occur when each pixel is calculated. Therefore, the difference in pixel values only for a certain region determined based on the inspection target region, Can be added. An example of summing differences between values of pixels for a certain area (M * M size) is expressed as follows.

Here, C (x, y) denotes a pixel value at the (x, y) position of the panel image, R (x, y) denotes a pixel value at the (x, y) Of the area. The panel inspection apparatus finds the position (i, j) at which the MAD (Mean Absolute Difference) value becomes minimum.

For example, the panel inspecting apparatus can obtain a difference between values of each pixel with respect to a strip region or a unit region in which an inspection target region identified through an optical inspection apparatus exists. Or the panel inspection apparatus can determine the region of interest as a region for obtaining a difference between each pixel.

The panel inspection apparatus repeatedly performs the process of summing the differences between the pixels of the predetermined regions 710 and 730 corresponding to each other while relatively moving the reference image 700 or the panel image 720 (for example, And 25 times for a 5 * 5 pixel size). The panel inspection apparatus matches the position of the panel image with the reference image when the sum of the differences between the values of the pixels becomes minimum.

8 is a diagram illustrating an example of a method for extracting an abnormal region in a region of interest according to an embodiment of the present invention.

Referring to FIG. 8, the panel inspection apparatus extracts an abnormal region 840 through a position-matched reference image and a difference image 830 for the region of interest 800 and 810 of the panel image.

For example, the panel inspection apparatus obtains a binarized image in which the difference image 830 is white and black based on a threshold value for white and black. The panel inspection apparatus finds a coordinate (maximum distortion coordinate) at which a value obtained by squaring each pixel value in a binarized image becomes a maximum. The panel inspection apparatus performs erosion and dilation operations with the maximum distortion coordinates as an initial starting point. For example, the panel inspecting apparatus removes pixel blocks having a certain threshold value or less through a 3 * 3 erosion operation, performs an expansion operation, and connects each pixel block to extract an abnormal region. Various preprocessing procedures other than erosion and expansion calculations can be performed as a method for removing noise and the like present in the difference image.

For example, if there are two lines of circuit patterns 802 and 804 in the region of interest 800 of the reference image and two lines of circuit patterns 812 and 814 and an anomalous region 820 are present in the region of interest 810 of the panel image , Only the abnormal region 840 is left when the difference image 830 of the two regions of interest 800 and 810 is obtained.

9 is a diagram illustrating an example of a cluster according to an embodiment of the present invention.

Referring to FIG. 9, the cluster is divided into a normal cluster 900 and a defect cluster 910. The normal cluster 900 and the defect cluster 910 are again divided into open 902 and 912 and short 904 and 914 and foreign substances 906 and 916 and scratches 908 and 918. Each cluster of the present embodiment is only one example, but a cluster according to various defect types may exist according to the embodiment. Each cluster 902, 904, 906, 908, 912, 914, 916, 198 comprises at least one or more parameters.

For example, when the parameter of the abnormal region is largest in similarity to the open defect cluster 912 among the plurality of clusters 902, 904, 906, 908, 912, 914, 916, and 198, the panel inspection apparatus determines that the abnormal region 840 is an open defect. Then, the panel inspection apparatus adds the parameter of the abnormal region 840 to the open coupling cluster 912.

10 is a diagram showing the configuration of an embodiment of a panel inspection apparatus according to the present invention.

Referring to FIG. 10, a panel inspection apparatus 1000 includes a position matching unit 1010, a region of interest detection unit 1020, a parameter extraction unit 1030, and a defect determination unit 1040.

The position matching unit 1010 aligns the positions of the reference image and the panel image in pixel units. For example, the position matching unit 1010 may adjust the positions of the reference image and the panel image using the alignment mark as shown in FIGS. 4 and 5, The position of the panel image can be matched. Alternatively, the position matching unit may perform the secondary position matching on a pixel-by-pixel basis using the pixel value difference of FIG. 7 after first matching the positions of the two images using the alignment mark.

The region-of-interest detection unit 1020 extracts an abnormal region from the position-matched reference image and the panel image through the difference image of the region of interest. The ROI acquiring unit 1020 can acquire a certain size region as a ROI centered on a different region of the reference image and the PANI image captured through the optical inspection apparatus of FIG.

The parameter extracting unit 1030 extracts a parameter for an abnormal region in the ROI. For example, the parameter extracting unit 1030 may vary according to the embodiment, such as brightness, size, contact of the lead wire, and the like.

The defect determination unit 1040 determines whether the parameter of the abnormal region is high in similarity with any of the clusters constructed in advance as shown in FIG. 9, and determines whether or not the defect is defective. For example, if the parameter of the abnormal region is high in similarity to the normal shot cluster (904 of FIG. 9), the defect determination section 1040 determines that the abnormal region is normal, not a defect.

The present invention can also be embodied as computer-readable codes on a computer-readable recording medium. A computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored. Examples of the computer-readable recording medium include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage, and the like. The computer-readable recording medium may also be distributed over a networked computer system so that computer readable code can be stored and executed in a distributed manner.

The present invention has been described with reference to the preferred embodiments. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. Therefore, the disclosed embodiments should be considered in an illustrative rather than a restrictive sense. The scope of the present invention is defined by the appended claims rather than by the foregoing description, and all differences within the scope of equivalents thereof should be construed as being included in the present invention.

Claims (8)

Matching the position of the reference image and the panel image to be inspected;
Extracting an abnormal region existing in a region of interest through a difference between the reference image and the panel image;
Extracting a parameter for an abnormal region existing in the region of interest; And
Determining a normal or defect of the abnormal region based on a cluster including at least one parameter for normal or defect and a parameter similarity between the abnormal region and the abnormal region. 2. The method of claim 1,
Aligning a position based on an alignment mark existing in each of the reference image and the panel image; And
And moving the reference image or the panel image within a predetermined range such that a sum of the differences between the respective pixels at the corresponding positions of the reference image and the panel image is minimized. PCB panel inspection method. 3. The method of claim 2, wherein aligning based on the alignment mark comprises:
Extracting a mark image including an alignment mark in the panel image;
Enlarging the mark image to a predetermined size;
Determining a center of the boundary area based on a change in a pixel value with respect to a boundary area of the alignment mark in an enlarged image;
Determining a center of the alignment mark based on a center of the boundary area; And
And aligning a center of an alignment mark of the panel image with an alignment mark of the reference image. The method of claim 2, wherein moving the panel image comprises:
And determining a region for obtaining a difference between each pixel as a strip region or a unit region according to whether the region to be inspected is a stripe region or a unit region recognized through an optical inspection apparatus. 2. The method of claim 1, wherein determining the normal or defect comprises:
A database including positional information on a region to be judged as defective or an area on which a defect is not to be judged among the respective regions of the panel is referred to and it is judged whether or not the abnormal region corresponds to the region to be judged as defective And determining whether a defect is present based on the degree of similarity with the cluster. The method of claim 1, wherein extracting the abnormal region comprises:
Designating a region of a predetermined size as a region of interest based on an inspection target region identified through an optical inspection apparatus; And
And determining an abnormal region through a difference image between the reference image and the panel image corresponding to the region of interest. A position matching unit for matching positions of a reference image and a panel image to be inspected;
An interest area determination unit that extracts an abnormal region existing in a region of interest through a difference between the reference image and the panel image;
A parameter extracting unit for extracting a parameter of an abnormal region existing in the ROI; And
And a defect judgment unit for determining a normal or defect of the abnormal region based on a cluster including at least one parameter for normal or defect and a similarity of the parameter to the abnormal region, . A computer-readable recording medium storing a program for performing the method according to any one of claims 1 to 6.

Images