JP4518494B2 - Land pattern inspection method and inspection apparatus - Google Patents

Description

  The present invention relates to a printed circuit board inspection method and inspection apparatus, and more particularly to a land pattern inspection method and inspection apparatus on a printed circuit board.

  On the surface of the printed circuit board on which electronic components and the like are mounted, a conductor wiring necessary for constituting a predetermined circuit is formed in a pattern. As one of these patterns, there is a pattern called land for use in mounting and connecting parts. As defects generated in the land, there are a “land break” (see FIG. 16A) and a “land defect” (see FIG. 16B). If such a defect exists, conduction may be incomplete, and a product using the substrate may malfunction. For this reason, detection of land defects as described above is one of the important inspection items in printed circuit board inspection.

As a method generally used for inspecting a land defect as described above, there is a method of measuring the diameter of a through hole using a length measuring element (for example, Patent Document 1). FIG. 17 is a diagram showing a defect inspection method by measuring the diameter of the through hole. As shown in FIG. 17A, the diameter of a circle is measured using a radial length measuring element 171. If all the diameters are equal (FIG. 17A), it is assumed that the land has no defect. On the other hand, as shown in FIG. 17B, when the land has a cutout, the length of the length measuring element applied to the cutout portion is longer than the other diameters. That is, since all the diameters are not equal, in this case, it is assumed that the land is defective.
JP 2001-267722 A

  However, the method disclosed in Patent Document 1 as described above has a problem that it cannot be detected when a land defect is fine. For example, as shown in FIG. 17C, in the case where the land is finely cut, the length measuring element 171 does not reach the defective portion, and as a result, all the diameters become equal. As a result, it is determined to be normal although there is a defect. In order to avoid such a problem, there is a method of increasing the number of length measuring elements 171. However, even in this case, as shown in FIG. It becomes. The same can be said for the case of land loss.

  SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a land pattern inspection method and inspection apparatus capable of detecting a fine defect in a land that cannot be detected even by a method of measuring the diameter of a through hole with a radial length gauge.

  In order to achieve the above object, the present invention employs the following configuration.

  A first invention is a land pattern inspection method for inspecting a land pattern shape of an inspection object by comparing an object image obtained by imaging a land pattern formed on the inspection object with a master image of the land pattern. An object image creation step for creating an object image obtained by binarizing an image obtained by imaging a land pattern formed on an inspection object, and measuring a diameter of a through hole formed in the land pattern from the object image A hole recognizing step, a hole filling circle image generating step for generating a hole filling circle image that is an image of a circle larger than the land diameter of the land pattern based on the measured diameter, and a shrinking process is predetermined for the hole filling circle image. The filling circle image shrinking step executed repeatedly more than once, and the filling circle image and object after shrinking A composite image generating step for generating a composite image by combining the image and a land pattern inspection step for comparing the master image with the composite image and inspecting a land pattern shape of the inspection object In the hole-filling circle image shrinking step, the object image and the hole-filling circle image are aligned on the same coordinate axis, and it is determined that the point of interest that is the pixel to be shrunk corresponds to the land portion in the object image. In this case, the land pattern inspection method is characterized in that the contraction is not performed on the attention point.

  The second invention is characterized in that, in the first invention, the number of times the shrinking process is repeated in the hole-filling circle image shrinking step is at least the same number of times that the hole-filling circle image is smaller than the through-hole. Note that the number of times equal to smaller than the through-hole here means that the filled circle image becomes an image smaller than the through-hole when the shrinking process is continued even when the target point corresponds to the land. Say the number of times.

  The third invention is a land pattern inspection method for inspecting a land pattern shape of the inspection object by comparing an object image obtained by imaging a land pattern formed on the inspection object with a master image of the land pattern. An object image creation step for creating an object image obtained by binarizing an image obtained by imaging a land pattern formed on an inspection object, and measuring a diameter of a through hole formed in the land pattern from the object image A hole recognizing step, a hole filling circle image generating step for generating a hole filling circle image that is an image of a circle larger than the land diameter of the land pattern based on the measured diameter, and a shrinking process is predetermined for the hole filling circle image. The filling circle image shrinking step and the filling circle image shrinking step are executed repeatedly. A cutout circle image generation step for masking the shrunken hole-filled circle image with an object image and cutting out a region corresponding to the land portion of the object image from the hole-filled circle image to generate a cutout circle image; A cut-out circle image contraction / expansion step that is contracted by the predetermined number of times, and then expands the predetermined number of times, a composite image generation step that generates a composite image by combining the cut-out circle image after expansion and the object image, and a master image And a synthesized pattern and a land pattern inspection step for inspecting the pattern shape of the land of the object to be inspected, and in the filling circle image shrinking step, the object image and the filling circle image on the same coordinate axis. The point of interest that is the pixel to be contracted is the object image When it is determined that corresponds to the definitive land portion is characterized in that it does not perform the contraction for the point of interest, a land pattern inspection method.

  According to a fourth aspect, in the third aspect, the number of times the shrinking process is repeated in the hole-filling circle image shrinking step is at least equal to the number of times that the hole-filling circle image is smaller than the through-hole.

5th invention is the land pattern inspection apparatus which test | inspects the land pattern shape of to-be-inspected object, Comprising: The imaging part which images the image of the land pattern of to-be-inspected object, The memory | storage part which memorize | stores the master image of a land pattern Based on the measured diameter, an object image creation unit that creates an object image obtained by binarizing the image captured by the imaging unit, a hole recognition unit that measures the diameter of the through hole formed in the land pattern from the object image, A hole-filled circle image generating unit that generates a hole-filled circle image that is an image of a circle larger than the land diameter of the land pattern, and a hole-filled circle image shrinkage that repeatedly executes shrinkage processing on the hole-filled circle image a predetermined number of times or more a Department, the composite image generation unit which synthesizes the filling circular image and the object image after the contraction to produce a composite image, read from the storage unit A land pattern inspection unit that compares the master image with the synthesized image and inspects the pattern shape of the land of the object to be inspected, and the object image and the hole-filled circle image on the same coordinate axis in the hole-filled circle image contraction unit And when it is determined that the target point that is a pixel to be contracted corresponds to a land portion in the object image, the target point is not contracted. This is a land pattern inspection apparatus.

  The sixth invention is characterized in that, in the fifth invention, the number of times the shrinking process is repeated in the hole-filling circle image shrinking portion is at least equal to the number of times that the hole-filling circle image is smaller than the through-hole.

  A seventh invention is a land pattern inspection apparatus for inspecting a land pattern shape of an object to be inspected, an image pickup unit for picking up an image of a land pattern of the object to be inspected, a storage unit for storing a master image of the land pattern, Based on the measured diameter, an object image creation unit that creates an object image obtained by binarizing the image captured by the imaging unit, a hole recognition unit that measures the diameter of the through hole formed in the land pattern from the object image, A hole-filling circle image generation unit that generates a hole-filling circle image that is an image of a circle larger than the land diameter of the land pattern, and a hole-filling circle image contraction unit that repeatedly executes contraction processing on the hole-filling circle image a predetermined number of times or more. And masking the hole-filled circle image shrunk by the hole-filled circle image shrinking part with an object image, and A cutout circle image generator that generates a cutout circle image by cutting out a region corresponding to a land portion of the object image, and a cutout circle image shrinkage that shrinks the cutout circle image a predetermined number of times and then expands the predetermined number of times The inflating unit, the composite image generating unit that generates the composite image by compositing the clipped circle image after expansion and the object image, the master image read from the storage unit and the composite image are compared, and the inspection object A land pattern inspection unit for inspecting the pattern shape of the land, and in the hole-filled circle image contraction unit, the object image and the hole-filled circle image are aligned on the same coordinate axis, and the pixel that is the target of contraction When it is determined that the point corresponds to the land portion in the object image, the point of interest is not contracted. The symptom is a land pattern inspection apparatus.

  The eighth invention is characterized in that, in the seventh invention, the number of times the shrinking process is repeated in the hole-filling circle image shrinking portion is at least equal to the number of times that the hole-filling circle image is smaller than the through-hole.

  According to the first aspect of the present invention, a circular image larger than the land is shrunk except for a portion overlapping with the land, thereby generating an image in which the hole of the through hole is filled, and comparison inspection with the master image can be performed. Thereby, it is possible to detect fine land defects that could not be detected by the conventional inspection method using the length measuring element. Further, it is not necessary to increase the number of length measuring elements in order to detect minute defects. For this reason, it is possible to prevent a load of calculation processing due to an increase in the length measuring element, and hence a decrease in processing speed. Furthermore, since there is no need to increase the length measuring element, it is possible to configure an inspection apparatus with high defect detection accuracy while suppressing the cost of hardware.

  According to the second aspect of the invention, by performing the shrinkage process up to the same number of times that the hole-filled circle image becomes smaller than the through-hole, it is possible to detect the land defect portion more prominently.

  According to the third aspect of the invention, the post-shrink hole filling circle image is cut out by masking with the object image, and once shrunk, it is expanded. Then, by synthesizing the expanded image with the object image, it is possible to generate an image of a land in which only the through hole is filled. Therefore, if the image and the master image are comparatively inspected, it is possible to detect a land defect (land defect) that cannot be detected even in the first aspect of the invention and that is not missing on the outer periphery of the land. As a result, it is possible to further improve the accuracy of land defect detection.

  According to the fourth aspect, the same effect as in the second aspect can be obtained.

  Further, according to the land pattern inspection apparatus of the present invention, the same effect as the land pattern inspection method of the present invention described above can be obtained.

(First embodiment)
A first embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram schematically illustrating the overall configuration of the optical appearance inspection apparatus 1. 1A is a top view of the optical appearance inspection apparatus, and FIG. 1B is a side view of the optical appearance inspection apparatus. In FIG. 1, the optical appearance inspection apparatus 1 includes a stage unit 11, a stage support unit 12, a stage drive mechanism 13, a base unit 14, an imaging camera 15, a support member 16, a camera support unit 17, and a camera drive mechanism 18. ing.

  The stage unit 11 constitutes a horizontal stage surface on the top surface. The printed wiring board S that is the object to be inspected is placed on the stage surface of the stage unit 11. The lower part of the stage unit 11 is supported by a stage support unit 12. The stage support 12 is fixed on the upper surface of the stage drive mechanism 13. The base portion 14 is parallel to the stage surface and extends and fixed in the illustrated Y-axis direction (main scanning direction). The stage drive mechanism 13 is slidably installed along a guide provided on the upper surface of the base portion 14 so as to extend in the Y-axis direction. That is, the stage drive mechanism 13 and the stage support part 12 and the stage part 11 fixed on the stage drive mechanism 13 are movable in the Y-axis direction.

  The support member 16 is installed in the upper space of the stage unit 11. On the support member 16, there is provided a camera drive mechanism 18 that extends in the illustrated X-axis direction (sub-scanning direction) parallel to the stage surface and perpendicular to the Y-axis direction. The camera support unit 17 is connected to the camera drive mechanism 18 and is disposed so as to be movable along the camera drive mechanism 18. The imaging camera 15 is supported by the camera support 17 so that the imaging direction is vertically downward (downward in the Z axis in the drawing). The imaging camera 15 is constituted by a CCD camera, for example, and converts incident light into an electrical signal indicating its color and intensity to generate an image of the captured printed wiring board S.

  The optical appearance inspection apparatus 1 captures an image of the printed wiring board S by the imaging camera 15 and acquires a land image. At this time, in order to acquire all the land images of the printed wiring board S, the optical appearance inspection apparatus 1 moves the stage unit 11 in the Y-axis direction and moves the imaging camera 15 in the X-axis direction. Specifically, main scanning is performed by moving the stage unit 11 in the Y-axis direction while the position of the imaging camera 15 in the X-axis direction is fixed. Here, every time the main scanning from one end to the other end of the inspection region of the substrate S is completed, the imaging camera 15 moves by a predetermined distance along the sub-scanning direction (X-axis direction). As a result, an image of the land for the entire inspection area of the printed wiring board S can be obtained by the imaging camera 15.

  FIG. 2 is a block diagram showing a functional configuration of the optical appearance inspection apparatus 1. In FIG. 2, the optical appearance inspection apparatus 1 includes an imaging camera 15, a control unit 21, an object image generation unit 22, a hole recognition unit 23, a filled circle image generation unit 24, an image contraction unit 25, an image composition unit 26, and a comparative inspection. Unit 27 and storage unit 28.

  The control unit 21 is configured by, for example, a CPU board. The control unit 21 is connected to each component described below. In addition, the control unit 21 controls the entire inspection process according to the present embodiment, such as reading of the land coordinate file, operation control of the imaging camera 15, and various image processes.

  The object image generation unit 22 binarizes the image captured by the imaging camera 15. Here, among the pixels constituting the image captured by the imaging camera 15, a pixel having a density higher than a predetermined threshold is set to 1, and the remaining pixels are set to 0 to be binarized. Of course, a pixel having a density higher than a predetermined threshold may be set to 0 and the remaining pixels may be set to 1. Then, an object image that is an image of a land portion is generated and output to the hole recognition unit 23. Here, it is assumed that the number of pixels of the object image is equal to the number of pixels of the hole-filled circle image described below. Further, the object image generation unit 22 generates a 1-pixel partial coordinate file 31 in which the coordinates of the portion of the pixel having a value of 1 (hereinafter referred to as 1-pixel) in the object image is extracted, and is stored in the storage unit 28. Store. In the drawing, 1-pixel is expressed in black and 0-pixel (a pixel having a value of 0) is expressed in white.

  The hole recognizing unit 23 measures the diameter of a land hole (hereinafter referred to as a through hole) with respect to the object image using a radial length measuring element (see FIG. 17). Then, the measured diameter is output to the hole-filling circle image generation unit 24.

  The hole-filling circle image generation unit 24 refers to the hole-filling circle pattern table 29 and reads a circle image corresponding to the diameter of the through hole measured by the hole recognition unit 23 (hereinafter referred to as a hole-filling circle image). Then, the read hole-filled circle image 292 is output to the image contraction unit 25.

  The image contraction unit 25 performs 8-neighbor contraction processing on a predetermined image. That is, a process (shrinkage process) for thinning 1-pixel by one layer at the boundary between 1-pixel and 0-pixel in a binarized image, and if there is even one 0-pixel in the vicinity of 8, Processing for setting the pixel to 0-pixel is performed. Further, the image contraction unit 25 outputs the contracted image to the image composition unit 26.

  The image composition unit 26 composes a predetermined image and an object image to generate a composite image. The composite image is an image used by the comparative inspection unit 27 for comparative inspection with the master image 302. The image composition unit 26 outputs the composite image to the comparison inspection unit 27.

  The comparison inspection unit 27 compares the composite image with the master image 302 read from the land coordinate file 30. If there is a certain difference or more, it is determined that the land portion has a defect.

  The storage unit 28 is a storage medium such as a semiconductor memory or a hard disk, and stores a hole filling circle pattern table 29, a land coordinate file 30, a 1-pixel partial coordinate file 31, and other data necessary for processing.

  Next, various data used in the present embodiment will be described. FIG. 3 shows the data structure of the hole-filling circle pattern table 29. As shown in FIG. In FIG. 3, the hole-filling circle pattern table 29 includes a set of sets of a diameter value 291 and a hole-filling circle image 292. The diameter value 291 corresponds to the diameter of the through hole measured by the hole recognition unit 23. The hole-filled circle image 292 is a circle image (binarized) that is slightly larger than the land having the through hole. For example, as shown in FIG. 4, when the land has a diameter of 1 mm and the through hole has a diameter of 0.8 mm, circular image data having a diameter of 1.2 mm is obtained as a hole-filled circle image 292 corresponding to the diameter of 0.8 mm. Stored (see FIG. 4). In addition, the number of pixels of the hole-filled circle image 292 matches the number of pixels of the object image.

  FIG. 5 is a diagram showing the data structure of the land coordinate file 30. As shown in FIG. 5, the land coordinate file 30 includes a set of a set of land coordinates 301 and a master image 302 of the land. The land coordinates 301 are coordinates of each land on the printed circuit board to be inspected. For example, the center coordinates of each land previously obtained from CAD data or the like are applicable. The master image 302 is a land image created from CAD data. Since the master image 302 is created from CAD data, the master image 302 is an image of a land with no holes.

  FIG. 6 is a diagram showing the data structure of the 1-pixel partial coordinate file 31. As shown in FIG. 6, the 1-pixel partial coordinate file 31 includes a set of 1-pixel coordinates 311 indicating the positions of pixels that are 1-pixels on the object image.

  Next, an outline of the operation of the optical appearance inspection apparatus 1 according to this embodiment will be described. Here, a case where a land break is detected as shown in FIG. 16A will be described as an example. First, the control unit 21 sequentially accesses the land coordinate file 30 and reads the land coordinates 301 and the master image 302. Next, the imaging camera 15 is moved to the read land coordinates. Next, a land image is picked up and a binarized image (hereinafter referred to as an object image) is generated. At the same time, 1-pixel coordinates are extracted from the object image, and a 1-pixel partial coordinate file 31 is generated. Next, the diameter of the through hole is measured from the binarized image. Next, the filling circle image 292 corresponding to the diameter is read from the filling circle pattern table 29. Next, while comparing with the 1-pixel partial coordinate file 31, the 8-neighbor contraction process is performed a predetermined number of times on the hole-filled circle image 292. Then, by comparing the post-shrinkage hole-filled circle image with the read master image 302, the land is checked for the presence or absence of a break. This series of processing is repeated until there is no inspection object (until the end of the land coordinate file 30).

  Hereinafter, the detailed operation of the land pattern inspection process performed by the optical appearance inspection apparatus 1 will be described with reference to FIGS. FIG. 7 is a flowchart showing a land pattern inspection process according to the first embodiment of the present invention. First, the control unit 21 reads out the land coordinates 301 and the master image 302 from the land coordinate file 30 (step S1).

  Next, the control unit 21 moves the imaging camera 15 to the read land coordinates 301. And the control part 21 images a land image with the imaging camera 15 (step S2).

  When the land image is captured, the control unit 21 causes the object image generation unit 22 to generate an object image (step S3). In step S3, the object image generation unit 22 binarizes the land image captured by the imaging camera 15, and generates an object image. Furthermore, the object image generation unit 22 extracts a coordinate group of a 1-pixel portion in the object image, and generates a 1-pixel partial coordinate file 31 storing the coordinate group.

  Next, the control part 21 makes the hole recognition part 23 measure the diameter of a through hole (step S4). In step S4, the hole recognizing unit 23 measures the diameter of the through-hole with respect to the object image generated by the object image generating unit 22 using a radial length measuring element in the same manner as in the related art described above (see FIG. 17). .

  If the diameter of the through hole can be measured, the control unit 21 then causes the hole-filling circle image generation unit 24 to generate a hole-filling circle image 292 (step S5). In step S5, the hole-filling circle image generation unit 24 searches the hole-filling circle image 292 from the hole-filling circle pattern table 29 based on the measured diameter, and generates a hole-filling circle image 292 by reading it (step S5).

  Next, if the hole-filled circle image 292 can be generated, the control unit 21 next causes the image contraction unit 25 to perform the contraction process with land stop (step S6). In step S6, 8-neighbor contraction processing is performed on the hole-filled circle image 292. At this time, the 1-pixel partial coordinate file 31 is referred to before contracting each pixel (target point). Then, when the point of interest coincides with the coordinates of 1-pixel in the object image (that is, the land portion), processing for not performing the contraction process for the point of interest is performed.

  FIG. 8 is a flowchart showing details of the contraction process with land stop shown in step S6. Here, a contraction process (hereinafter referred to as a rounding process) is performed a predetermined number of times for pixels (hereinafter referred to as boundary pixels) for one round of the outer periphery of the hole-filled circle image. With respect to the number of times that the rounding process is performed, the number of times necessary and sufficient for at least the outer circumference of the hole-filled circle image 292 to reach a position corresponding to the inside of the through hole on the object image is determined in advance. In other words, even when the outer circumference of the hole-filled circle image reaches the outer circumference of the land during the process of contracting with the land stop, the ring is cut further with an annular ring (difference between the land diameter and the drill bit diameter). Beyond that, the number of rounding processes is determined in advance so that the outer circumference of the hole-filled circle image 292 reaches the inside of the through hole. Hereinafter, in order to make the explanation easy to understand, it is assumed that the circulation process is performed 10 times.

  In FIG. 8, first, the image contracting unit 25 extracts a coordinate group (hereinafter referred to as a boundary coordinate group) of boundary pixels of the hole-filled circle image 292 (step S11). Next, the image contraction unit 25 determines the coordinates of the pixel as the target point (hereinafter referred to as the target coordinate) from the extracted boundary coordinate group (step S12). Next, the image contraction unit 25 searches the 1-pixel partial coordinate file 31 based on the value of the target coordinate (step S13). Next, the image contraction unit 25 determines the search result (step S14). As a result, if there is a coordinate that matches the target coordinate in the 1-pixel partial coordinate file 31, the number of pixels of the object image and the hole-filled circle image 292 match. It corresponds to the upper land part. Therefore, the image contraction unit 25 does not perform the 8-neighbor contraction process for the target coordinate. That is, as a result of the determination in step S14, if there is no matching coordinate (NO in step S14), the image contraction unit 25 performs 8-neighbor contraction processing on the target coordinate (step S15). On the other hand, if there is a matching coordinate (YES in step S14), the image contraction unit 25 proceeds to step S16 without performing the process in step S15.

  In step S16, the image contraction unit 25 determines whether or not the processing from steps S12 to S15 has been performed on all the coordinates of the boundary coordinate group, that is, whether or not one round of processing has been completed. (Step S16). As a result, when one round process has not been completed yet (NO in step S16), the image contraction unit 25 returns to step S12 and repeats the process. On the other hand, when one round of processing is completed, the image contraction unit 25 advances the processing to the next step S17.

  In step S <b> 17, the image contraction unit 25 determines whether or not the circulation process has been performed for the designated number of times, that is, whether or not the circulation process has been performed 10 times. As a result, if the circulation process has not been performed 10 times (NO in step S17), the image contraction unit 25 returns to step S11 and repeats the process. On the other hand, if the circulation process has been performed 10 times (YES in step S17), the image contraction unit 25 ends the contraction process with the land stop.

  The flow of the contraction process with the land stop will be supplementarily described with reference to FIG. First, FIG. 9A shows an object image and a hole-filled circle image 292 before starting the contraction process with land stop. Assume that both images have the same number of pixels. Note that although the coordinates of the land portion of the object image are actually stored in the 1-pixel portion coordinate file 31 described above, the object image will be described as it is for easy understanding.

  Next, FIG. 9B is a diagram illustrating a processing target of the first round processing. As shown in the lower part of FIG. 9B, the boundary pixels (pixels corresponding to the outer periphery) of the hole-filled circle image 292 are the objects of the circulation process. Then, 8-neighbor contraction processing is performed on the boundary pixel. As a result, the land portion (1-pixel) does not exist in the area of the object image corresponding to the boundary pixel, and therefore, the entire boundary area is replaced with 0-pixel. That is, the hole-filled circle image 292 is slightly reduced.

  Next, FIG. 9C is a diagram showing a state at the end of the fourth round process. FIG. 9C shows a state in which the boundary pixel, that is, the outer periphery of the hole-filled circle image (the lower stage in FIG. 9C) and the outer periphery of the land image (the upper stage in FIG. 9C) match. That is, in the fourth contraction, the hole filling circle and the land have the same diameter.

  Next, FIG. 9D is a diagram showing a state at the start of the fifth round process. After the fifth rounding process, no further contraction is performed on the portion of the boundary pixel that is in contact with the land portion, and only the pixel K in the cut-off portion contracts. As a result, as shown in FIGS. 9E and 9F, only the cut-off portion K contracts. FIG. 9E shows a state at the start of the sixth contraction process, and FIG. 9F shows a state at the start of the eighth contraction process. Then, when the rounding process is finally performed up to 10 times, as shown in FIG. 9G, a hole-filled circle image 292 in a state in which only the spotted portion is contracted is formed. In addition, since the number of times of the rounding process is such that the outer circumference of the hole-filled circle image 292 reaches the inside of the through hole beyond the annular, the part corresponding to the through hole is contracted. As a result, the part of the seat is more prominent.

  Returning to FIG. 7, after the process of step S6, the control unit 21 causes the image composition unit 26 to generate a composite image (step S7). In step S7, the image composition unit 26 generates a composite image by compositing the hole-filled circle image 292 after the contraction processing and the object image (FIG. 9H).

  Next, the control unit 21 causes the comparison inspection unit 27 to perform a comparison inspection process (step S8). In step S8, the comparison inspection unit 27 compares the master image 302 read in step S1 with the synthesized image, as shown in FIG. As a result, if there is a difference equal to or greater than a predetermined threshold, it is determined that the land to be inspected is defective.

  When the process of step S8 is completed, the control unit 21 determines whether or not all the land coordinate files 30 have been read, that is, whether or not there are lands to be inspected (step S9). As a result, when the land to be inspected still remains (YES in step S9), the control unit 21 returns to step S1 and repeats the process. On the other hand, when no land to be inspected remains (NO in step S9), the control unit 21 ends the land pattern inspection process. The land pattern inspection process according to the first embodiment is thus completed.

  As described above, in the first embodiment, when a circular image larger than the land is prepared and the contraction process is performed, it is determined whether or not the attention point corresponds to the 1-pixel portion of the binarized object image. However, it contracts in the vicinity of 8. Then, by synthesizing the circle image and the object image, an image of the land with the through hole filled therein is generated. By comparing this with the master image 302, it is possible to detect minute land breaks that could not be detected by a conventional inspection method using a length measuring element. In addition, it is not necessary to increase the number of length measuring elements in order to detect the minute break. For this reason, it is possible to prevent a load of calculation processing due to an increase in the length measuring element, and hence a decrease in processing speed. Furthermore, since there is no need to increase the length measuring element, it is possible to configure an inspection apparatus with high defect detection accuracy while suppressing the cost of hardware.

  In the above-described embodiment, the 8-neighbor contraction process is used in the contraction process. However, the present invention is not limited to this, and a 4-neighbor contraction process or the like may be used. In addition, regarding the reading of the hole-filled circle image in the above step 5, instead of the hole-filled circle image 292 of the hole-filled circle pattern table 29, only the diameter value of the image is stored, and each time the image is based on the diameter. A circle may be generated and binarized to generate a hole-filled circle image. As a result, it is not necessary to store an image file that tends to increase the capacity required for storage, and it is possible to save the capacity of the storage unit 28 and reduce costs.

(Second Embodiment)
Next, a second embodiment of the present invention will be described with reference to FIGS. In the first embodiment described above, after the shrink process with land stop of the hole-filled circle image 292 is completed, a comparative inspection is performed using the hole-filled circle image after shrinkage. On the other hand, in the second embodiment, after the contraction process with the land stop is finished, a cutout circle image that is an image cut out by masking the hole-filled circle image 292 with the object image is generated. Further, image contraction → expansion processing is performed on the cutout circle image. Thereafter, a composite image is generated by combining the cutout circle image and the object image. Then, the composite image and the master image 302 are compared and inspected. Note that the optical appearance inspection apparatus 40 according to this embodiment includes a cut-out circle image generation unit 41 and an image dilation part in the functional configuration of the optical appearance inspection apparatus 1 described with reference to FIG. 2 in the first embodiment described above. The other components are the same as those in the first embodiment. Therefore, components other than the cutout circle image generation unit 41 and the image expansion unit 42 are denoted by the same reference numerals, and detailed description thereof is omitted.

  FIG. 11 is a block diagram showing a configuration of an optical appearance inspection apparatus 40 according to the second embodiment of the present invention. In FIG. 11, the cutout circle image generation unit 41 masks the object image on the hole-filled circle image 292 after the shrinkage process, and cuts out the hole-filled circle image 292. In other words, a cutout circle image corresponding to a through hole is generated in a manner of die cutting. Note that the cutout circle image is generated by being binarized. The image expansion unit 42 expands a predetermined image by performing 8-neighbor expansion processing. That is, a process (expansion process) of increasing the thickness of 1-pixel by one layer at the boundary between 1-pixel and 0-pixel in a binarized image, and if there is even one 1-pixel in the vicinity of 8 pixels, that pixel Is processed as 1-pixel. In addition, the image expansion unit 42 outputs the expanded image to the image composition unit 26.

  The details of the land pattern inspection process according to the second embodiment of the present invention will be described below with reference to FIGS. Here, a case where a land defect as shown in FIG. 16B is detected is taken as an example. FIG. 12 is a flowchart illustrating land defect detection processing performed by the optical appearance inspection apparatus 40 according to the second embodiment. In FIG. 12, the processing from step S21 to step S25 is the same as the processing from step S1 to step S5 described with reference to FIG. 7 in the first embodiment described above, and therefore detailed description thereof is omitted here. . Also, the processing from step S31 to step S32 is the same as the processing from step S8 to step S9 described with reference to FIG. 7 in the first embodiment, and therefore detailed description thereof is omitted here.

  Next, the process of steps S26 to S30 will be described with reference to FIGS. FIG. 13 is a diagram illustrating changes in the hole-filling circle image and the cutout circle image in steps S26 to S30. FIG. 14 is a diagram showing changes in the cutout circle image in steps S28 and S29. First, the control unit 21 causes the image contraction unit 25 to perform the contraction process with land stop of the hole-filled circle image 292 (step S26). In step S26, the image contraction unit 25 performs the contraction process with land stop on the hole-filled circle image 292 in the same manner as the process of step S6 described with reference to FIG. 7 in the first embodiment. As a result, for example, it is assumed that the hole-filled circle image 292 becomes the same size as the land diameter by the fourth contraction (FIG. 13A). Here, the defect of the land in this embodiment exists in the inner peripheral part of the land, and there is no missing part about the outer periphery. For this reason, even if the shrinkage is subsequently performed, the hole-filled circle image 292 does not shrink from the fourth state. As a result, even when the contraction process with the land stop in step S26 is completed, the size of the hole-filled circle image 292 is the same as the outer circumference circle of the land (FIG. 13A).

  Next, the control unit 21 causes the cutout circle image generation unit 41 to perform cutout circle image generation processing (step S27). In step S27, the cutout circle image generation unit 41 masks the object image on the hole-filled circle image 292 (FIG. 13B), and cuts out the hole-filled circle image 292, thereby cutting-out circle image (FIG. 13C). Is generated. As a result, the cut-out circle image is an image in which the land missing portion is present as a “projection” with respect to the through hole.

  Next, the control unit 21 performs contraction → expansion processing to remove the protrusions (steps S28 and S29). The contraction process performed here is performed without using the 1-pixel partial coordinate file 31, unlike the contraction process with land stop performed in step S26. That is, since the process is for erasing the “projection” of the cutout circle image, the 1-pixel and 0-pixel of the cutout circle image itself are simply determined, and the 8-neighbor contraction process is performed (hereinafter, referred to as “rejection”). In order to distinguish it from the contraction process with land stop, it is called the uniform contraction process).

  In step S <b> 28, the image contraction unit 25 performs the uniform contraction process on the cutout circular image a predetermined number of times, for example, once. 14, first, FIG. 14A shows a cut-out circle image before contraction (“□” corresponds to 1-pixel). In FIG. 14A, there are pixels 141 (protrusions) corresponding to the land missing portions (corresponding to FIG. 13C). If this image is subjected to 8-neighbor contraction, as shown in FIG. 14 (B), only the pixels with “•”, which are 1-pixel around 8 pixels, remain (FIG. 13 ( Equivalent to D)). Above, the process in step S28 is complete | finished.

  Next, in step S29, the image expansion unit 42 performs expansion processing on the cutout circular image (FIG. 14B) uniformly contracted in step S28 as many times as contracted in step S28. That is, since the uniform contraction process is performed only once in step S28, the expansion process is performed only once here. As a result, as shown in FIG. 14C, a cutout circle image in which the “protrusions” corresponding to the land defect portions are deleted (corresponding to FIG. 13E) is obtained. In addition, since the number of contractions in step S28 is equal to the number of expansions in step S29, the cut-out circle image returns to the original size. As a result, an image in which only the through hole is filled without a gap can be formed.

  Returning to FIG. 12, next, the control unit 21 causes the image composition unit 26 to generate a composite image (step S30). In step S30, the image composition unit 26 generates a composite image by combining the clipped circle image and the object image after the expansion process, as shown in FIG.

  As described above, an image having the same size as the inner circle of the land can be obtained by cutting out the cutout circle image from the hole-filled circle image, performing the uniform shrinkage process, and then performing the expansion process. Then, a composite image is generated using this, and compared with the master image 302 as shown in FIG. 15 (step S31), it is possible to detect a land defect. Thus, the land pattern inspection process according to the second embodiment is completed.

  As described above, in the second embodiment, noise (protrusions) in the image corresponding to the land defect portion can be eliminated by cutting out the hole-filled circle image with the object image, and once expanding it after uniform shrinkage. Then, if this image is combined with the object image, a land image in which only the through hole is filled can be generated, and if the image is compared with the master image 302, defects that cannot be detected in the first embodiment can also be detected. . That is, in the first embodiment, a defect having a cut on the outer periphery of the land (land cutout: FIG. 16A) can be detected, but a defect having no cut on the outer periphery of the land (land defect: FIG. 16 (B)) cannot be detected. On the other hand, according to the second embodiment, a land defect can be detected in addition to a land cut. As a result, it is possible to further improve the accuracy of land defect detection.

  In addition, about the inspection method concerning this invention, it is desirable to carry out following the length measuring inspection performed conventionally. That is, it is desirable to avoid this step if a land break is found by the length sensor inspection, and to carry out the inspection method according to the present invention when no abnormality is found by the length measurement inspection.

  Moreover, in the said embodiment, it is specialized to inspect the land pattern shape, and it has been described that the target lands are sequentially imaged and inspected based on the land coordinates, but the entire pattern on the substrate is inspected. The land pattern may be inspected simultaneously in the comparative inspection. That is, all the substrate images are read in advance, and the object image near the target land coordinates is converted into an object image that is filled in so that a land cutout or a land defect can be identified by image processing according to the present invention, You may make it perform the comparative test | inspection of all the patterns containing a land. In the above case, as a master image to be compared, an image obtained by imaging a non-defective substrate may be used instead of design data such as CAD.

  INDUSTRIAL APPLICABILITY The land pattern inspection method and land pattern inspection apparatus according to the present invention can detect fine land defects that cannot be detected by a method using a length measuring element, and are useful for applications such as a printed circuit board inspection apparatus.

The figure which shows typically the whole structure of the optical inspection apparatus which concerns on embodiment of this invention. 1 is a functional block diagram showing an optical inspection apparatus according to a first embodiment of the present invention. The figure which showed an example of the data structure of the hole-filling circle pattern table 29 of FIG. Diagram showing the relationship between the diameter of the land, the diameter of the through hole, and the diameter of the circle on the filled circle image The figure which showed an example of the data structure of the land coordinate file 30 of FIG. The figure which showed an example of the data structure of the 1-pixel partial coordinate file 31 of FIG. The flowchart which shows the land pattern inspection process which concerns on the 1st Embodiment of this invention. The flowchart which showed the detail of the contraction process with a land stop shown by step S6 of FIG. The figure which shows the change of the hole-filling circle image 292 in the shrinkage | contraction process with a land stop shown by step S6 of FIG. Diagram showing comparative inspection process Functional block diagram showing an optical inspection apparatus according to a second embodiment of the present invention The flowchart which shows the land pattern inspection process which concerns on the 2nd Embodiment of this invention. The figure which shows the process shown by step S26-S30 of FIG. The figure which shows the process shown by step S28 of FIG. Diagram showing comparative inspection process Figure showing an example of land loss The figure which shows the land defect inspection method used conventionally

Explanation of symbols

DESCRIPTION OF SYMBOLS 1, 40 Optical appearance inspection apparatus 11 Stage part 12 Stage support part 13 Stage drive mechanism 14 Base part 15 Imaging camera 16 Support member 17 Camera support part 18 Camera drive mechanism 21 Control part 22 Object image generation part 23 Hole recognition part 24 Fill hole Circle image generation unit 25 Image contraction unit 26 Image composition unit 27 Comparison inspection unit 28 Storage unit 29 Filled circle pattern table 30 Land coordinate file 31 1-pixel partial coordinate file 41 Cutout circle image generation unit 42 Image expansion unit 121 Corresponding portion of pixel 151 Length element 291 Diameter value 292 Filled circle image 301 Land coordinate 302 Master image 311 1-pixel coordinate

Claims (8)

A land pattern inspection method for inspecting a land pattern shape of an inspection object by comparing an object image obtained by imaging a land pattern formed on the inspection object with a master image of the land pattern,
An object image creation step for creating an object image obtained by binarizing an image obtained by imaging a land pattern formed on the inspection object;
A hole recognition step of measuring the diameter of the through hole formed in the land pattern from the object image;
Based on the measured diameter, a filling circle image generating step for generating a filling circle image that is an image of a circle larger than the land diameter of the land pattern;
A hole-filling circle image shrinking step for repeatedly executing shrinkage processing for a predetermined number of times for the hole-filling circle image;
A combined image generation step of generating a combined image by combining the object image and the hole-filled circle image after contraction;
A land pattern inspection step for comparing the master image with the composite image and inspecting a land pattern shape of the inspection object;
In the hole-filling circle image shrinking step, the object image and the hole-filling circle image are aligned on the same coordinate axis, and a target point that is a pixel to be shrunk corresponds to a land portion in the object image. The land pattern inspection method according to claim 1, wherein the contraction is not performed on the attention point when it is determined that the target point is present.   2. The land pattern inspection method according to claim 1, wherein in the hole filling circle image shrinking step, the number of times the shrinking process is repeated is at least equal to the number of times that the hole filling circle image is smaller than the through hole. A land pattern inspection method for inspecting a land pattern shape of an inspection object by comparing an object image obtained by imaging a land pattern formed on the inspection object with a master image of the land pattern,
An object image creation step for creating an object image obtained by binarizing an image obtained by imaging a land pattern formed on the inspection object;
A hole recognition step of measuring the diameter of the through hole formed in the land pattern from the object image;
Based on the measured diameter, a filling circle image generating step for generating a filling circle image that is an image of a circle larger than the land diameter of the land pattern;
A hole-filling circle image shrinking step for repeatedly executing shrinkage processing for a predetermined number of times for the hole-filling circle image;
A cutout circle image generation step of masking the hole-filled circle image shrunk in the hole-filling circle image shrinking step with an object image and generating a cutout circle image by cutting out a region corresponding to the land portion of the object image from the hole-filled circle image. When,
A cutout circle image shrinking and expanding step of shrinking the cutout circle image a predetermined number of times and then expanding the predetermined number of times;
A combined image generation step of generating a combined image by combining the clipped circle image after expansion and the object image;
A land pattern inspection step for comparing the master image with the composite image and inspecting a land pattern shape of the inspection object;
In the hole-filling circle image shrinking step, the object image and the hole-filling circle image are aligned on the same coordinate axis, and a target point that is a pixel to be shrunk corresponds to a land portion in the object image. The land pattern inspection method according to claim 1, wherein the contraction is not performed on the attention point when it is determined that the target point is present.   4. The land pattern inspection method according to claim 3, wherein in the filling circle image shrinking step, the number of times the shrinking process is repeated is equal to at least the number of times that the filling circle image is smaller than the through hole. A land pattern inspection apparatus for inspecting a land pattern shape of an inspection object,
An imaging unit that captures an image of a land pattern of the inspection object;
A storage unit for storing a master image of the land pattern;
An object image creation unit that creates an object image obtained by binarizing the image captured by the imaging unit;
A hole recognition unit that measures the diameter of the through hole formed in the land pattern from the object image;
Based on the measured diameter, a hole-filling circle image generation unit that generates a hole-filling circle image that is an image of a circle larger than the land diameter of the land pattern;
For the hole-filled circle image, a hole-filling circle image shrinking unit that repeatedly executes shrinkage processing more than a predetermined number of times,
A composite image generation unit that generates a composite image by combining the object image and the hole-filled circle image after contraction;
Comparing the master image read from the storage unit and the composite image, and having a land pattern inspection unit that inspects the pattern shape of the land of the inspection object,
In the hole-filling circle image contraction unit, the object image and the hole-filling circle image are aligned on the same coordinate axis, and a point of interest that is a pixel to be contracted corresponds to a land portion in the object image. When it is determined that there is a land pattern inspection apparatus, the contraction is not performed on the attention point.   6. The land pattern inspection apparatus according to claim 5, wherein the number of times the shrinking process is repeated in the hole-filled circle image shrinking unit is at least equal to the number of times that the hole-filled circle image is smaller than the through-hole. A land pattern inspection apparatus for inspecting a land pattern shape of an inspection object,
An imaging unit for imaging a land pattern of the object to be inspected;
A storage unit for storing a master image of the land pattern;
An object image creation unit that creates an object image obtained by binarizing the image captured by the imaging unit;
A hole recognition unit that measures the diameter of the through hole formed in the land pattern from the object image;
Based on the measured diameter, a hole-filling circle image generation unit that generates a hole-filling circle image that is an image of a circle larger than the land diameter of the land pattern;
For the hole-filled circle image, a hole-filling circle image shrinking unit that repeatedly executes shrinkage processing more than a predetermined number of times,
A cutout circle image generating unit that masks the hole-filled circle image shrunk by the hole-filled circle image shrinking unit with an object image and generates a cutout circle image by cutting out a region corresponding to the land portion of the object image from the hole-filled circle image. When,
A cutout circle image shrinking and expanding portion that shrinks the cutout circle image a predetermined number of times and then expands the predetermined number of times;
A combined image generating unit that generates a combined image by combining the clipped circle image after expansion and the object image;
Comparing the master image read from the storage unit and the composite image, and having a land pattern inspection unit that inspects the pattern shape of the land of the inspection object,
In the hole-filling circle image contraction unit, the object image and the hole-filling circle image are aligned on the same coordinate axis, and a point of interest that is a pixel to be contracted corresponds to a land portion in the object image. When it is determined that there is a land pattern inspection apparatus, the contraction is not performed on the attention point. The land pattern inspection apparatus according to claim 7 , wherein the number of times the shrinking process is repeated in the hole-filled circle image shrinking unit is at least equal to the number of times that the hole-filled circle image is smaller than the through-hole.

Images