JP2000258353A - Defect inspection method and device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To stably detect a defect by obtaining the average concentration of concentration data for each small region, judges a pixel to be a defective pixel when the concentration data for each pixel does not exist in a set concentration region, and finally judges the defect according to the judgment result. SOLUTION: An inspection target O is placed on a traveling stage 1 and is illuminated by lighting 2 for picking up an image with a CCD camera 3, and an image detection processing part 4 generates concentration data for each pixel from the pickup image. Before starting, a reference concentration data is generated by the image detection processing part 4 from a conforming image and is stored in an image memory 5. A CPU 8 controls an entire device. An inspection arithmetic processing part 6 divides an inspection region into, for example, a local region with 10×10 pixels, and calculates the average value of the concentration data for each local region. When the average value is not located between upper and lower thresholds being set from the reference concentration data, all pixels in the local region are judged to be defective and binarization processing is made. In this manner, when the image concentration is far away from the reference concentration, the part is judged to be defective.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a defect inspection method and apparatus for inspecting, for example, defects such as wiring patterns and bonding pads on a printed circuit board. The present invention relates to a defect inspection method and an apparatus for inspecting a defect of the inspection object based on obtained density data.

[0002]

2. Description of the Related Art As a method for detecting defects such as scratches, foreign matter, dents, and dirt on the surface of a printed circuit board or the like, conventionally, a captured image of an inspection object and a reference image obtained from a non-defective product have been used. Is widely used to determine the density difference (absolute density difference) between the two over the entire image and determine that a portion where the density difference exceeds a predetermined threshold value is a defect.

[0003]

By the way, as for the inspection object, for example, when the production lot changes, the average density (absolute density) may slightly change even if the density change state (relative density) is the same. In such a case, if the above-described conventional defect inspection method is used, a defect is determined based on the absolute density difference between the target image and the reference image. There was a possibility of doing it. Further, if the threshold is set wide to prevent such erroneous recognition, the accuracy of defect determination is reduced, and conversely, the possibility of erroneously recognizing a defective product as a good product is increased. Note that even in the same production lot, the absolute density value may vary depending on the location of the captured image, and in this case, the same can be said. The present invention has been made in view of the above circumstances, and its purpose is to stably detect defects even when the inspection object has a concentration variation depending on a location or a concentration variation between production lots. An object of the present invention is to provide a defect inspection method and device capable of performing detection.

[0004]

In order to achieve the above object, the present invention provides a defect inspection for inspecting a defect of the inspection object based on density data for each pixel obtained from a picked-up image of the inspection object. In the method, an area dividing step of dividing the imaging range of the inspection object into predetermined small areas, and an average density of density data of the inspection object is calculated for each of the small areas obtained in the area dividing step. Average concentration calculation process,
It is determined whether or not the density data of each pixel of the inspection object is included in a first density range set in advance based on the average density in each small area obtained in the average density calculation step. Then, based on the first determination step of determining that the pixel is a defective pixel when the pixel is not included in the range, and the determination result for each pixel obtained in the first determination step, the inspection is performed. And a final determination step of performing a final defect determination of the object. It is desirable to change the first concentration range to an appropriate value for each production lot, for example, in order to maintain high inspection accuracy. Further, it is determined whether or not the average density for each of the small areas calculated in the average density calculation step is included in a second density range preset for each of the small areas based on predetermined reference density data. A second determining step of determining that all the pixels in the small area are defective pixels when the average density is not included in the range, and in the final determining step, If the final defect judgment of the inspection object is performed based on the judgment results obtained in the judgment step and the second judgment step, the density of the picked-up image is at least partially and extremely reduced to the reference density. If the data is far from the data, it can be determined that the part is defective. Even if the second density range is set to be slightly wider, the defect is reliably recognized by the relative determination in the first determination step, and there is no fear that the accuracy of the defect determination is reduced. It is desirable to change the second concentration range to an appropriate value for each production lot, for example, in order to maintain high inspection accuracy. The order of the first and second determination steps is the same as that of the second determination step.
Is performed first, and the processing in the first determination step is performed only on the pixels in the small area that has not been determined to be defective. This can reduce the calculation time of the entire defect detection processing. It is possible. Furthermore, in the final determination step, if the defective pixels are labeled and a final defect determination is performed based on the labeling result, the defective pixels can be regarded as a unit, so that the defective pixels usually appear to some extent. It is possible to easily and accurately determine a defect that often occurs. In this final determination, for example, only when the number of defective pixels of the same label obtained by the above labeling exceeds a predetermined threshold, it is conceivable to determine a defective pixel relating to the label as defective. Furthermore, by performing, for example, a masking process on the density data of the inspection object and performing the defect determination based on only the density data in a predetermined inspection target area, high-speed processing without waste can be performed.

An apparatus capable of implementing the defect inspection method is a defect inspection apparatus for inspecting a defect of the inspection object based on density data for each pixel obtained from a captured image of the inspection object. Area dividing means for dividing the imaging range of the object into predetermined small areas; average density calculating means for calculating an average density of the density data of the inspection object for each of the small areas obtained by the area dividing means; Whether the density data for each pixel of the inspection object is included in a first density range set in advance based on the average density of each small area obtained by the average density calculation means. A first determining unit that determines that the pixel is a defective pixel when the pixel is not included in the range, and a determination result for each pixel obtained by the first determining unit. Final defect judgment of inspection object It is configured as a defect inspection apparatus characterized by comprising; and a final determination means for performing.

[0006]

According to the present invention, in the first determination step, when the density data of each pixel is far from the average density in the small area, the pixel is determined to be a defective pixel. Since the determination is performed, even if there is a variation in the density depending on the location of the inspection object or a variation in the density for each production lot, the defect can be stably detected without being affected by the variation. . Further, in the second determination step, the density of the captured image is at least partially and extremely separated from the reference density data by performing the density determination using the second density range based on the absolute density in small area units. If so, the part can be determined to be defective. Even if the second density range is set to be slightly wider, the defect is reliably recognized by the relative determination in the first determination step, and there is no fear that the accuracy of the defect determination is reduced. Further, if the second determination step is performed before the first determination step, it is possible to reduce the calculation time of the entire defect detection processing. Further, if the defect determination is performed after removing the density data other than the inspection area by the masking processing or the like, high-speed processing without waste can be performed. Further, if a labeling process is performed on the defective pixel, the defective pixel can be regarded as a unit. Therefore, it is possible to easily and accurately determine a defect that usually appears in a certain unit.

[0007]

Embodiments and examples of the present invention will be described below with reference to the accompanying drawings to provide an understanding of the present invention. The following embodiments and examples are mere examples embodying the present invention, and do not limit the technical scope of the present invention. FIG. 1 is a flowchart showing an example of a processing procedure of a defect inspection method according to an embodiment of the present invention, and FIG. 2 is a block diagram showing a schematic configuration of a defect inspection apparatus A1 according to an embodiment of the present invention. 3 is an explanatory view showing an example of a set of local areas and test areas on the inspection object, Fig 4 is the average concentration G (X, Y) a threshold g _1, g _2,
5 is a density histogram showing an example of the relationship between a and b. As shown in FIG. 2, the defect inspection apparatus A1 according to the present embodiment includes a moving stage 1, illumination 2, a CCD camera 3, an image detection processing unit 4, an image memory 5, an inspection calculation processing unit 6, a master CPU 8, The drive control unit 7 is provided. An inspection object 0 such as a printed board is placed on the moving stage 1. The CCD camera 3 captures a grayscale image of the inspection object 0 laid on the moving stage 1 and illuminated by the illumination 2. The moving stage 1 can be moved in the X and Y directions by the drive of the drive control unit 7, whereby the inspection object 0 is positioned relative to the CCD camera 3. The image detection processing unit 4 generates density data for each pixel based on the image captured by the CCD camera 3.
Before starting the defect inspection, the image memory 5 stores the CCD.
Non-defective density data (hereinafter referred to as reference density data) generated by the image detection processing unit 4 based on the non-defective image captured by the camera 3 is stored in advance. In the inspection calculation processing unit 6, the density data of the inspection object 0 generated by the image detection processing unit 4 (hereinafter referred to as inspection object density data) and the reference density stored in the image memory 5 in advance. Based on the data, a defect inspection process (FIG. 1), which will be described later, is executed, and the defect determination of each inspection object 0 is performed. This inspection calculation processing section 6 is an average density calculating means,
Are examples of the determination means, the second determination means, and the final determination means. The master CPU 8 controls the entire device A1 including the inspection calculation processing unit 6 and the drive control unit 7.

Next, the procedure of the defect inspection processing mainly performed by the inspection arithmetic processing unit 6 will be described with reference to the flowchart shown in FIG. Here, for simplicity, the picked-up image of the inspection object 0 is 100 as shown in FIG.
Each pixel is represented by p (x, y) and density data of each pixel is represented by g (x, y). Further, a local area (corresponding to a predetermined small area) in units of 10 × 10 pixels is set on the captured image, and each local area is represented by P (X, Y). Further, in the captured image of the inspection object 0, an inspection area (shown by oblique lines in an enlarged view of the local area P (2, 9)) is set as shown in FIG. It is assumed that masking data for masking data is prepared in advance.

When the defect inspection process is started, first, the inspection target density data of the inspection object 0 is taken into the inspection calculation processing unit 6 via the CCD camera 3 and the image detection processing unit 4 (step S1). The inspection target density data taken into the inspection calculation processing unit 6 is masked using the masking data (specifically, an AND operation between the position data indicating the inspection area, which is an example of the masking data, and each pixel position is performed. After the density data outside the inspection area is eliminated (step S2), it is divided into the local area P (X, Y) of 10 × 10 pixels (step S2).
3: Equivalent to the area dividing step). Note that steps S2 and S
The order of processing 3 may be reversed. continue,
In the inspection operation processing unit 6, one local region P (X,
Y) is taken out, and the average value G (X, Y) of all the density data g (x, y) belonging to the local area is calculated (step S4: corresponding to an average density calculation step). And
It is determined whether the average value G (X, Y) satisfies the following equation (step S5: corresponds to a second determination step). g ₁ (X, Y) ≦ G (X, Y) ≦ g ₂ (X, Y) Here, the above g ₁ (X, Y) and g ₂ (X, Y) are obtained from a good-quality captured image. The threshold is a preset threshold for each local area based on the obtained reference density data, and both are set as absolute density values. The above g ₁ (X, Y) and g ₂ (X, Y)
Is an example of the second density range. If the average density G (X, Y) does not satisfy the above expression, all the pixels in the local area are determined to be defective pixels, and the binarization processing is performed on all the pixels in the local area. (Step S8). 2 here
The binarization process is, for example, a process in which a defective pixel is set to 1 and other non-defective pixels are set to 0. As described above, by performing the density determination using the threshold based on the absolute density for each local area, when the density of the captured image is at least partially extremely different from the reference density data, the part is determined to be defective. It can be determined that there is. It should be noted that by performing the determination in units of local regions before the determination in units of pixels (step S7) described later, it is possible to reduce the calculation time of the entire defect detection process.

On the other hand, if the average density G (X, Y) satisfies the above equation, it is determined whether or not the density data g (x, y) of all pixels in the local area satisfies the following equation. The determination is repeated (step S6 → S7 (corresponding to the first determination step) → S9 → S6 ...). (1-a / 100) × G (X, Y) ≦ g (x, y) ≦ (1-b / 100) × G (X, Y) where a and b are shown in FIG. As described above, the value (%) is used to set a predetermined density width based on the average density G (X, Y) in each local region, and is set in advance based on the reference density data. The values of a and b need not necessarily be set for each local area.
Usually, for example, it may be changed for each lot. here,
The range of the above expression defined by the above a and b is an example of the first concentration range. In step S7, the g
Pixels determined that (x, y) do not satisfy the above expression (for example, in the histogram shown in FIG.
Pixels that are out of the range of -a% and + b% based on (X, Y)) are determined to be defective pixels, and are subjected to the above-described binarization processing (step S8). As described above, when the density data of each pixel is significantly different from the average density in the local area, the pixel is determined to be a defective pixel, and the relative defect determination is performed. Even when there is a variation in the density or a variation in the density for each production lot, it is possible to stably detect the defect without being affected by the variation.

When the above processing is completed for one local area, the processing from step S4 is repeated for the next local area. When the above processing is completed for all the local areas (step S10), the inspection calculation processing unit 6 determines that the defect is a defect, performs further labeling processing on the binarized pixels, and based on the labeling result, The final judgment of non-defective / defective of the inspection object 0 is made. Usually, a defect often appears in a group of several to several hundred pixels. However, a process of assigning the same label to the group of the same defective pixel and enabling the above-described group of the defective pixel to be determined is performed here. Is the labeling process. As described above, by labeling the defective pixels, the defective pixels can be regarded as a unit (defect area), so that the defect determination can be performed easily and accurately. The positions of some defective areas extracted by the labeling process can be specified by calculating, for example, the center of gravity or the like, and their areas can be obtained by calculating the number of pixels belonging to the defective areas. In the final judgment of non-defective / defective products, for example, if there is a defect having an area of 10 or more, it is judged to be defective, and other than that, it is judged to be non-defective, and an appropriate judgment method according to the standard of the inspection target product is applied. It is possible.

As described above, in the defect inspection method according to the present embodiment, in step S7, if the density data of each pixel is far from the average density in the local area, the pixel is determined as a defective pixel. In the case where there is a variation in the density depending on the location of the inspection object or a variation in the density for each production lot, etc., the relative defect judgment is performed stably without being affected by the variation. Defect detection can be performed. Further, in step S5, by performing the density determination using the threshold based on the absolute density for each local area, if the density of the captured image is at least partially extremely different from the reference density data, the partial Can be determined to be defective. Note that even if the threshold based on the absolute density is set to be slightly wider, the defect is surely recognized by the relative determination in step S7, and there is no fear that the accuracy of the defect determination is reduced. Further, by performing the determination in units of local areas (step S5) before the determination in units of pixels (step S7), it is possible to reduce the calculation time of the entire defect detection process. Further, since the defect determination is performed after removing the density data other than the inspection area by the masking processing, high-speed processing without waste can be performed. Also,
By performing the labeling process on the defective pixels, the defective pixels can be perceived as a unit. Therefore, it is possible to easily and accurately determine a defect which usually appears in a certain unit.

[0013]

DESCRIPTION OF THE PREFERRED EMBODIMENTS When the inspection area included in the local area is small and the number of inspection target pixels p (x, y) in the local area is small, the average density calculation itself is easily affected by noise. In some cases, the result of the expression is NG.
In order to prevent such a situation, when the number of pixels to be inspected in the local area is smaller than a certain value, the above-described step S4 is performed.
Pixels in that area can be processed as non-defective products without performing the processing of S10. In addition, g ₁ (X, Y) and g ₂ (X, Y) in the above equation, and a and b in the above equation can be finely adjusted in accordance with, for example, each time a production lot or the like changes. Is desirable to maintain high. It is also desirable to set the local area to an appropriate size according to the inspection object to be inspected.

[0014]

As described above, the present invention relates to a defect inspection method for inspecting a defect of an inspection object based on density data for each pixel obtained from a captured image of the inspection object. An area dividing step of dividing an imaging range of an object into predetermined small areas; an average density calculating step of calculating an average density of density data of the inspection object for each of the small areas obtained in the area dividing step; The density data for each pixel of the inspection object is a first preset density based on the average density in each small area obtained in the average density calculation step.
The first determination step is to determine whether or not the pixel is within the density range, and if not, the pixel is determined to be a defective pixel. And a final determination step of performing a final defect determination of the inspection object based on the determination result for each pixel. Even when there is a variation in density depending on the location or a variation in density for each production lot, it is possible to detect defects stably without being affected by the variation. It is desirable to change the first concentration range to an appropriate value for each production lot, for example, in order to maintain high inspection accuracy. Further, it is determined whether or not the average density for each of the small areas calculated in the average density calculation step is included in a second density range preset for each of the small areas based on predetermined reference density data. A second determining step of determining that all the pixels in the small area are defective pixels when the average density is not included in the range, and in the final determining step, If the final defect judgment of the inspection object is performed based on the judgment results obtained in the judgment step and the second judgment step, the density of the picked-up image is at least partially and extremely reduced to the reference density. If the data is far from the data, it can be determined that the part is defective. Even if the second density range is set to be slightly wider, the defect is reliably recognized by the relative determination in the first determination step, and there is no fear that the accuracy of the defect determination is reduced. It is desirable to change the second concentration range to an appropriate value for each production lot, for example, in order to maintain high inspection accuracy. As for the order of the first and second determination steps, the second determination step is performed first, and the first determination step is performed only on pixels in a small area that are not determined to be defective. By performing the process by the process, it is possible to reduce the calculation time of the entire defect detection process. Furthermore, if the binarized defective pixels are further labeled, and a final defect determination is performed based on the labeling result, the defective pixels can be regarded as a unit, so that the defective pixels usually appear to some extent. It is possible to easily and accurately determine a defect that often occurs. In this final determination, for example, only when the number of defective pixels of the same label obtained by the above labeling exceeds a predetermined threshold, it is conceivable to determine a defective pixel relating to the label as defective. Furthermore, by performing, for example, a masking process on the density data of the inspection object and performing the defect determination based on only the density data in a predetermined inspection target area, high-speed processing without waste can be performed.

[Brief description of the drawings]

FIG. 1 is a flowchart illustrating an example of a processing procedure of a defect inspection method according to an embodiment of the present invention.

FIG. 2 shows a defect inspection apparatus A1 according to an embodiment of the present invention.
FIG. 2 is a block diagram showing a schematic configuration of FIG.

FIG. 3 is an explanatory diagram showing an example of setting a local area and an inspection area on an inspection object.

FIG. 4 shows an average density G (X, Y) and threshold values g ₁ , g ₂ ,
5 is a density histogram showing an example of the relationship between a and b.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Moving stage 2 ... Lighting 3 ... CCD camera 4 ... Image detection processing part 5 ... Image memory 6 ... Inspection calculation processing part (Average density calculation means, 1st determination means, 2nd determination means, and final determination means Example) 7: Drive control unit 8: Master CPU

Claims (10)

[Claims] In a defect inspection method for inspecting a defect of an inspection object based on density data for each pixel obtained from a captured image of the inspection object, an imaging range of the inspection object is set to a predetermined small area. An area dividing step of dividing; an average density calculating step of calculating an average density of density data of the inspection object for each of the small areas obtained in the area dividing step;
It is determined whether or not the density data of each pixel of the inspection object is included in a first density range set in advance based on the average density in each small area obtained in the average density calculation step. Then, based on the first determination step of determining that the pixel is a defective pixel when the pixel is not included in the range, and the determination result for each pixel obtained in the first determination step, the inspection is performed. A final determination step of performing a final defect determination of the object. 2. The defect inspection method according to claim 1, wherein said first density range can be changed. 3. The method according to claim 1, wherein the average density for each of the small areas calculated in the average density calculating step is included in a second density range preset for each of the small areas based on predetermined reference density data. A second determination step of determining whether or not all the pixels in the small area are defective if the average density is not included in the range. 3. The defect inspection method according to claim 1, wherein a final defect determination of the inspection object is performed based on the determination results obtained in the first determination step and the second determination step. 4. The defect inspection method according to claim 3, wherein said second density range can be changed. 5. The defect inspection method according to claim 3, wherein the processing in the first determination step is performed only for pixels in a small area that has not been determined to be defective in the second determination step. 6. The defect inspection method according to claim 1, wherein, in the final determination step, the defective pixel is labeled, and a final defect determination is performed based on the labeling result. 7. A defective pixel associated with a label in a final defect determination is determined only when the number of defective pixels of the same label obtained by the labeling exceeds a predetermined threshold. Defect inspection method. 8. The defect inspection method according to claim 1, wherein the defect determination is performed based only on the density data in a predetermined inspection target area among the density data of the inspection target. 9. A defect inspection apparatus for inspecting a defect of the inspection object based on density data for each pixel obtained from a captured image of the inspection object, wherein an imaging range of the inspection object is set to a predetermined small area. Area dividing means for dividing, average density calculating means for calculating an average density of density data of the inspection object for each of the small areas obtained by the area dividing means,
It is determined whether or not the density data of each pixel of the inspection object is included in a first density range set in advance based on the average density in each small area obtained by the average density calculation means. Then, based on the first determination means for determining that the pixel is a defective pixel when the pixel is not included in the range, and the determination result for each pixel obtained by the first determination means, A defect inspection apparatus comprising: final determination means for performing final defect determination of an object. 10. A method according to claim 1, wherein the average density for each of the small areas calculated by the average density calculating means is included in a second density range preset for each of the small areas based on predetermined reference density data. A second determining means for determining whether or not the average density is not within the range, and determining that all pixels in the small area are defective pixels. 10. The defect inspection apparatus according to claim 9, wherein a final defect determination of the inspection object is performed based on the determination results obtained by the first determination unit and the second determination unit.