JP2011153893A - Soldered region visual inspection device, and soldered region visual inspection method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To determine simply and accurately a spike defect and a bridge defect regardless of the thickness of solder. <P>SOLUTION: An image input means 12 acquires an RGB color image of a soldered region of a printed board imaged by a camera 1, and an image conversion means 13 converts it into an HSV color space image. A soldered domain extraction means 14 extracts a soldered domain by comparing the RGB color image with the HSV color space image, and a land domain extraction means 15 performs pattern matching of the RGB color image with model data generated by a model data generation means 11 for pattern matching to specify a land domain. A defect determination means 16 performs quality determination based on a difference domain between the soldered domain and the land domain. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a soldering part appearance inspection apparatus and method for judging whether or not a soldering part is good by image processing using a captured image.

  Conventionally, as a method for inspecting the quality of a soldered part on a printed circuit board, there is a method for determining the quality by determining the thickness of a solder fillet by a light cutting method using a line light source as disclosed in Patent Document 1, for example. In this method, first, a linear slit light is projected onto a soldering site and imaged, and by image processing, a light-cut image drawn with slit light is detected on the surface of the soldering site, and the soldering site is slit. The cross-sectional shape cut along the light is estimated. Subsequently, the estimated cross-sectional shape characteristics (fillet thickness, etc.) of the soldered part are counted, and compared with a predetermined determination threshold value, the quality of the shape is determined.

Japanese Patent Publication No. 63-9602

Since the conventional soldering part appearance inspection method is configured as described above, the thickness of the soldering part may be thin if there is a minute horn defect or a bridging defect. In this case, pass / fail judgment is made using an optical cutting method. Even if it went, there was a problem that the thickness could not be measured and the quality could not be determined accurately.
In addition, in order to detect a minute horn defect, it is necessary to estimate a plurality of cross-sectional shapes by finely moving a line light source or a printed circuit board, and thus there is a problem that it takes time and labor for inspection.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to easily and accurately determine horn defects and bridge defects regardless of the thickness of the solder.

  The soldering part appearance inspection apparatus according to claim 1 of the present invention converts an RGB color image into an HSV color space image using RGB color images obtained by imaging printed circuit boards including soldering parts to be inspected. An image converting means for comparing the RGB color image with the HSV color space image, extracting a soldering portion from the image and setting it as a solder area, and a land of each printed circuit board prepared in advance. A land area extracting means for identifying a land portion from an image and setting the land area as a land area based on position information of the area; a solder area set by the solder area extracting means; and a difference area between the land area set by the land area extracting means And a failure determination unit that determines whether the difference is good or bad according to the feature amount of the difference area.

  In the soldering site appearance inspection apparatus according to claim 2 of the present invention, the defect determination means calculates the width of the rectangular area circumscribing the difference area as the feature quantity of the difference area, and the width of the rectangular area is the threshold for determination. If it is larger, it is judged as defective.

  According to a third aspect of the present invention, there is provided a soldering site appearance inspection apparatus that further determines a bridging defect and a horn defect according to a feature amount with respect to a difference area that is determined to be defective by a defect determination means.

  In the soldering site appearance inspection apparatus according to claim 4 of the present invention, the defect determination means calculates the width of the rectangular area circumscribing the difference area as the feature quantity of the difference area, and the difference area determined to be defective is rectangular. When the width of the region is equal to or greater than the determination threshold value, it is determined that the bridge is defective, and when the width is less than that, the horn defect is determined.

  According to a fifth aspect of the present invention, there is provided a soldering part appearance inspection apparatus that is formed at an equal angle with each side starting from the intersection of one side of the substantially L-shaped soldering part to be inspected and the other side. An illumination device that is arranged on the dividing line and irradiates light toward the abutting part between the printed circuit boards, and a camera that is arranged on the bisector and that images the soldering part in a state of irradiating light from the illumination device, The image conversion means uses an RGB color image captured by the camera. The polarization filter is attached to the lens of the camera and blocks light reflected from the surface of each printed board.

  According to a sixth aspect of the present invention, a soldered part appearance inspection method converts an RGB color image into an HSV color space image using RGB color images obtained by imaging printed circuit boards including soldered parts to be inspected. An image conversion step for comparing the RGB color image with the HSV color space image, extracting a soldering portion from the image, and setting the solder region as a solder region; A land area extraction step for identifying a land portion from an image and setting it as a land area based on position information of the area, and a difference area between the solder area set in the solder area extraction step and the land area set in the land area extraction step And a failure determination step of determining pass / fail according to the feature amount of the difference area.

  In the soldering part appearance inspection method according to claim 7 of the present invention, in the defect determination step, the width of the rectangular area circumscribing the difference area is calculated as the characteristic amount of the difference area, and the width of the rectangular area is determined as a determination threshold value. If it is larger, it is judged as defective.

  The soldering part appearance inspection method according to claim 8 of the present invention further includes a horn / bridge defect determination step for determining a bridge defect and a horn defect according to the feature amount for the difference area determined to be defective in the defect determination step. Is.

  In the soldered part appearance inspection method according to claim 9 of the present invention, the width of the rectangular area circumscribing the difference area is calculated as the feature quantity of the difference area in the horn / bridge defect determination step, and the defect determination step determines that the The determined difference area is determined as a bridging defect when the width of the rectangular area is equal to or greater than the determination threshold, and as a horn defect when the width is less than the determination threshold.

  According to the first, second, fifth, sixth and seventh aspects of the present invention, the RGB color image is compared with the HSV color space image, and the soldering portion is extracted from the image and set in the solder region. Based on the position information of the land part, the land part is specified from the image and set as the land area, and the quality is determined according to the feature amount of the difference area between the solder area and the land area. When inspecting a soldering part, it is possible to extract a defective part regardless of the thickness of the solder and determine whether it is good or not, and therefore, it is possible to easily and accurately determine the defect of the soldering part.

  According to the third, fourth, eighth, and ninth aspects of the present invention, since the difference area determined to be defective is further discriminated from the bridge defect and the horn defect according to the feature amount, the cause of the defect can be accurately determined. Can be determined.

It is the schematic which shows the structure of the soldering site | part external appearance inspection apparatus which concerns on Embodiment 1 of this invention. FIG. 2A shows an RGB color image taken by the soldering site appearance inspection apparatus shown in FIG. 1, FIG. 2A shows a case where the polarizing filter 3 is not used, and FIG. 2B shows a case where the polarizing filter 3 is used. It is a block diagram which shows the structure of the image processing apparatus 5 shown in FIG. 4 is a flowchart for explaining the operation of the pattern matching model data creating unit 11 according to the first embodiment. An RGB color image of the unsoldered printed boards 6a and 7a used by the pattern matching model data creating means 11 for creating model data is shown. 4 is a flowchart for explaining the operation of the image processing apparatus 5 according to the first embodiment. 3 is an example of an RGB color image that is a detection target of the image processing apparatus 5; 4 is an RGB color image showing an example of a solder area extracted by a solder area extracting unit 14. 4 is an RGB color image showing the positional relationship between the wiring pattern 141 and the land area 142 determined by the solder area extracting means 14. 4 is an RGB color image showing an example of difference areas 151 and 152 calculated by the defect determination means 16. It is an RGB color image explaining how the defect determination means 16 determines whether a defect area is a horn defect or a bridge defect. It is a RGB color image which shows the example joined to the state which the printed circuit boards 6 and 7 shifted | deviated to the horizontal direction.

Embodiment 1 FIG.
1 is a schematic diagram showing the configuration of a soldered part appearance inspection apparatus according to Embodiment 1 of the present invention. This soldering site appearance inspection apparatus includes a camera 1, a lens 2, a polarizing filter 3, an illumination device 4, and an image processing device 5, and pulls a land portion at an abutting portion against which printed circuit boards 6 and 7 are abutted by soldering. This is to detect horns and bridging defects at the soldered portion having the L-shaped cross section that is soldered. The solder may be lead-free solder. In the following description, the color of the printed circuit boards 6 and 7 is green.

  In the soldering site appearance inspection apparatus shown in FIG. 1, the printed circuit boards 6 and 7 soldered are held by a holding member (not shown), and the two printed circuit boards 6 and 7 are placed on a bisector that is equiangular. The illumination device 4 and the camera 1 are arranged. The illumination device 4 uses ring illumination (white light source) with a diffusion plate to illuminate the inspection range of the printed boards 6 and 7. Further, the polarization direction of the polarizing filter 3 is adjusted so that the light reflected from the substrate surfaces of the printed circuit boards 6 and 7 does not appear in the captured image, and is installed in front of the lens 2, and the printed circuit boards 6 and 7 are imaged by the camera 1. To do. As the camera 1, a color camera is used, and RGB color images (red, green, and blue components) are output.

FIG. 2A shows an RGB color image captured without using the polarizing filter 3, and FIG. 2B shows an RGB color image captured using the polarizing filter 3. Although the drawing is a black and white image, the color of the substrate is green on an actual RGB color image. However, it is not the presence or absence of the polarizing filter 3 that the soldering part 101 of the printed circuit boards 6 and 7 in FIG. 2A and the soldering part 103 of the printed circuit boards 6 and 7 in FIG. This is because it is an RGB color image obtained by imaging different printed circuit boards.
When the solder to which the flux is applied is soldered to the land portion of the printed circuit board, the solder flux remains on the surface of the printed circuit board, so that the light from the illumination device 4 may be reflected by the flux and reflected in the captured image. Since the reflection 102 shown in FIG. 2A becomes noise in image processing by the solder region extraction means 14 described later, the reflected light is adjusted by adjusting the polarization direction of the polarizing filter 3 during imaging, and S / N is improved.

  FIG. 3 is a block diagram showing an internal configuration of the image processing apparatus 5 shown in FIG. The image processing device 5 includes a pattern matching model data creating unit 11 that creates model data (position information) used for pattern matching described later in advance, an image input unit 12 that acquires an RGB color image captured by the camera 1, The image conversion means 13 for converting the RGB color image into an HSV color space image (hue, saturation, brightness component) and the RGB color image and the HSV color space image are compared, and the soldered portions of the printed boards 6 and 7 are displayed on the image. A solder area extracting means 14 for extracting as a solder area, a land area extracting means 15 for specifying a land portion of the printed circuit boards 6 and 7 on the RGB color image using the pattern matching model data, and setting the land area as a land area; And a defect determination means 16 for detecting horn and bridge defects from the solder area and the land area.

Next, a method for creating pattern matching model data by the pattern matching model data creating means 11 will be described using the flowchart shown in FIG. FIG. 5 is an RGB color image obtained by imaging the printed boards 6 and 7 (hereinafter referred to as unsoldered printed boards 6a and 7a) that are not soldered with the camera 1 of the soldered portion appearance inspection apparatus shown in FIG.
The pattern matching model data creating means 11 sets a template for specifying the position of the land portion on the printed circuit board 6 for the RGB color image of FIG. 5 (step ST1). In the example of FIG. 5, a wiring pattern in the vicinity of the land portion is set in the template 111. Subsequently, the pattern matching model data creating means 11 sets a plurality of land portions of the unsoldered printed circuit board 6a in the land area 112 (step ST2), and further sets the positional relationship between the template 111 and the land area 112 (step ST2). ST3).

The pattern matching model data creating means 11 sets a template for specifying the position of the land portion on the printed circuit board 7 for the RGB color image of the unsoldered printed circuit board 7a as described above (step ST4). The land portion of the unsoldered printed circuit board 7a is set as a land area (step ST5), and the positional relationship between the template and the land area is set (step ST6).
The pattern matching model data creation means 11 generates model data from the template set for the printed circuit boards 6a and 7a and the setting information of the land area and stores it (step ST7).

Next, a method for detecting horns and bridging defects by the image processing apparatus 5 will be described using the flowchart shown in FIG.
First, the image input means 12 acquires RGB color images of the printed boards 6 and 7 captured by the camera 1 (step ST11). Subsequently, the image conversion unit 13 converts the RGB color image acquired by the image input unit 12 into an HSV color space image (step ST12, image conversion step).

FIG. 7 shows an example of an RGB color image to be detected by the image processing apparatus 5. As shown in FIG. 7, each land portion at the abutting portion of the printed circuit boards 6 and 7 is in a state in which each soldering portion 121 is formed by being pulled by soldering in the lateral direction.
The surface of the soldering part 121 is composed of an irregular reflection surface and a specular reflection surface. Among these, the irregular reflection surface has a large value in the R (red component) image of the RGB color image, and the luminance value of the red component is large. On the other hand, the S (saturation) image of the HSV color space image has a small value and no vividness. Also, the specular reflection surface has a small value in both the R image of the RGB color image and the S image of the HSV color space image. Incidentally, the horn and the bridge are mainly diffuse reflection surfaces.
On the other hand, since the boards of the printed boards 6 and 7 are green, the value is very small in the red image that is complementary to the green color, that is, the R image of the RGB color image. This is equivalent to lowering the luminance value when a red light source is irradiated to green. On the other hand, the S image of the HSV color space image has a large value and is vivid.

  Based on the above characteristics, the solder region extracting means 14 compares the R image of the RGB color image and the S image of the HSV color space image, and the region of the S image that is darker than the R image, that is, the irregular reflection surface of the soldering site. An image region having features and an image region having features of a specular reflection surface are extracted and set as solder regions (step ST13, solder region extraction step). FIG. 8 is an RGB color image showing an example of the solder region 131 extracted by the solder region extracting means 14. In FIG. 8, the solder region 131 is hatched for clarity.

The following steps ST14 to ST17 are detection processes for the printed circuit board 6, and steps ST18 to ST21 are detection processes for the printed circuit board 7. FIG. 9 is an RGB color image showing the positional relationship between the wiring pattern 141 and the land area 142 determined by the solder area extracting means 14.
First, detection processing for the printed circuit board 6 is performed. The land area extraction unit 15 performs pattern matching with the RGB color image using the template for the unsoldered printed circuit board 6a out of the model data generated by the pattern matching model data generation unit 11 in advance. The position of the wiring pattern 141 is determined (step ST14).
Subsequently, the land area extraction means 15 determines the position of the land area 142 of the printed circuit board 6 by applying the positional relationship between the template and the land area for the unsoldered printed circuit board 6 a in the model data to the position of the wiring pattern 141. (Step ST15).
Steps ST14 and ST15 are land area extraction steps.

Subsequently, the defect determining means 16 subtracts the land area 142 set by the land area extracting means 15 from the solder area 131 of the RGB color image set by the solder area extracting means 14 to set the set-theoretic difference area (hereinafter referred to as difference area). Is obtained (step ST16). FIG. 10 is an RGB color image showing an example of the difference areas 151 and 152 calculated by the defect determination means 16. In FIG. 10, the difference areas 151 and 152 are hatched for clarity.
Subsequently, the defect determination means 16 calculates a rectangular area circumscribing each of the difference areas 151 and 152, and if the feature amount (width, area, etc.) of the rectangular area is larger than a determination threshold value, a defective area where a horn or a bridge has occurred. (Step ST17).
Steps ST16 and ST17 are defect determination steps.

  Next, detection processing for the printed circuit board 7 is performed. As in the case of the printed circuit board 6, the land area extraction unit 15 determines the position of the land area of the printed circuit board 7 by pattern matching using the model data (steps ST18 and ST19, land area extraction step), and the defect determination unit 16 Is a defective area by determining the difference area between the solder area and the land area (steps ST20, ST21, defect determination step).

  Finally, the defect determination means 16 determines whether the defect area is a horn defect or a bridge defect (step ST22, horn / bridge defect determination step). FIG. 11 is an RGB color image for explaining a method by which the defect determination means 16 determines whether a defect area is a horn defect or a bridge defect. A threshold value for determination is calculated in advance based on the width between adjacent land portions and set in the defect determination means 16 in advance. The defect determination unit 16 calculates the maximum widths 161 and 162 of the defective area in the drawing solder direction, and determines that a bridge defect is detected if the maximum widths 161 and 162 are equal to or greater than the threshold value, and a horn defect if the maximum width 161 or 162 is less than the threshold value. In the example of FIG. 11, since the maximum width 161 is equal to or greater than the threshold value, the defective area 163 is determined as a bridging defect, and since the maximum width 162 is less than the threshold value, the defective area 164 is determined as a horn defect.

  When there is a soldering part in the abutting portion where the two printed circuit boards 6 and 7 are abutted, for example, as shown in FIG. 12, the land portion of the printed circuit board 6 and the land portion of the printed circuit board 7 are shifted in the lateral direction. Even if they are joined, pattern matching is performed for each of the printed boards 6 and 7 to set the position of each land portion, so that horns and bridging defects can be detected.

  Further, in the above description, since there is a soldering portion at the abutting portion where the printed circuit boards 6 and 7 are abutted, the detection process of steps ST14 to ST17 is performed on the printed circuit board 6, and ST18 to ST21 is performed on the printed circuit board 7. However, it goes without saying that the detection process only needs to be performed once if only one of the printed circuit boards has a soldered portion.

As described above, according to the first embodiment, the solder having a substantially L-shaped cross section is formed in the abutting portion where the printed circuit boards 6 and 7 are abutted to each other and electrically connects the land portions of the printed circuit boards 6 and 7. The soldering part visual inspection apparatus for performing the visual inspection of the soldering part is divided into two equal parts at the same angle with each side, starting from the intersection of one side of the substantially L shape of the soldering part to be inspected and the other side. An illuminating device 4 that is arranged on a line and irradiates light toward the abutting portion of the printed circuit boards 6 and 7; a camera 1 that images a soldered portion in a state where the light from the illuminating device 4 is irradiated; A polarizing filter 3 that is attached to the lens 2 and blocks light reflected from the substrate surface of each printed circuit board 6, 7, an image input unit 12 that acquires an RGB color image captured by the camera 1, and an RGB color image In HSV color space image An image converting means 13 for converting, a solder area extracting means 14 for comparing the RGB color image and the HSV color space image, extracting a soldering portion from the image, and setting the solder area, and a wiring pattern of the printed boards 6 and 7 Based on the pattern matching model data creation means 11 for creating pattern matching model data indicating the positional relationship between the land and the land portion, the land portion is specified from the image based on the model data created in advance and set in the land area. The land area extraction unit 15 calculates a difference area between the solder area and the land area, calculates a width of a rectangular area circumscribing the difference area, and determines a defect when the width of the rectangular area is larger than a determination threshold. The determination means 16 is provided.
Therefore, each area to be soldered can be imaged at once by the area-type camera 1, so that it is not time-consuming and the inspection time can be shortened. Further, noise removal can be performed by using the polarizing filter 3 at the time of imaging.
Also, since the RGB color image and the HSV color space image are compared and the solder area is extracted to determine whether it is acceptable or not, it is possible to accurately detect minute horn defects and bridge defects regardless of the thickness of the soldered part. .

  Further, according to the first embodiment, when the defect determination unit 16 determines that the difference area (defect area) determined to be defective is less than the determination threshold when the width of the rectangular area is equal to or greater than the determination threshold, It was configured to discriminate horn defects. For this reason, it is possible to determine whether the defect of the soldering site is a horn defect or a bridge defect, and it is possible to accurately determine the cause of the defect.

  In the first embodiment, the L-shaped soldered portion of the abutting portion of the printed circuit boards 6 and 7 is the object of the pass / fail judgment, but the present invention is not limited to this. It is also possible to detect a horn defect and a bridge defect in a soldered portion where lead wires and connector wires are drawn and soldered using lead-free solder. In this case, the camera 1 and the illuminating device 4 are arranged on the printed circuit board so that the angle formed by the imaging direction of the camera 1 and the illuminating direction of the illuminating device 4 is about 45 degrees. It is necessary to remove the reflection.

DESCRIPTION OF SYMBOLS 1 Camera 2 Lens 3 Polarizing filter 4 Illumination device 5 Image processing device 6,7 Print board 6a, 7a Unsoldered print board 11 Pattern matching model data creation means 12 Image input means 13 Image conversion means 14 Solder area extraction means 15 Land area Extraction means 16 Defect determination means 101, 103 Soldering part 102 Reflection 111 Template 112 Land area 121 Soldering part 131 Solder area 141 Wiring pattern 142 Land area 151, 152 Difference area 161, 162 Maximum width 163, 164 Defect area

Claims (9)

A soldering part appearance inspection device for inspecting the appearance of a soldering part having a substantially L-shaped cross section, which is formed in abutting parts where the printed circuit boards are brought into contact with each other and electrically connects the land parts of the printed circuit boards. ,
Image conversion means for converting the RGB color image into an HSV color space image using the RGB color image obtained by imaging the printed boards including the soldering portion to be inspected;
A solder region extracting means for comparing the RGB color image with the HSV color space image, extracting the soldering portion from the image, and setting the solder region;
Land area extraction means for specifying the land part from the image and setting it as a land area based on position information of the land part of each printed circuit board created in advance,
And a failure determination means for calculating a difference area between the solder area set by the solder area extraction means and the land area set by the land area extraction means, and determining whether the difference is characteristic or not. Soldering part appearance inspection device.   The defect determination means calculates a width of a rectangular area circumscribing the difference area as a feature value of the difference area, and determines that the defect is defective when the width of the rectangular area is larger than a determination threshold. The soldering part appearance inspection apparatus as described.   2. The soldering site appearance inspection apparatus according to claim 1, wherein the defect determination means further determines a bridging defect and a horn defect according to a feature amount for the difference area determined to be defective.   The defect determination means calculates a width of a rectangular area circumscribing the difference area as a feature value of the difference area, and the difference area determined to be defective is less than a bridging defect when the width of the rectangular area is equal to or larger than a determination threshold. 4. The soldered part appearance inspection apparatus according to claim 3, wherein a horn defect is determined in the case of. Starting from the intersection of one side and the other side of the substantially L-shaped soldering part to be inspected, it is arranged on a bisector that is equiangular with each side, toward the abutting part between the printed boards An illumination device that emits light;
A camera that is arranged on the bisector and that images the soldered part in a state of irradiating light from the illumination device;
A polarizing filter attached to the lens of the camera and blocking light reflected from the surface of each printed circuit board;
The soldering part appearance inspection apparatus according to claim 1, wherein the image conversion unit uses an RGB color image captured by the camera. A soldering part appearance inspection method for inspecting the appearance of a soldering part having a substantially L-shaped cross section, which is formed in abutting parts where the printed circuit boards are brought into contact with each other and electrically connects the land parts of the respective printed circuit boards. ,
An image conversion step of converting the RGB color image into an HSV color space image using the RGB color image obtained by imaging the printed boards including the soldering site to be inspected;
A solder region extracting step of comparing the RGB color image with the HSV color space image, extracting the soldering site from the image, and setting the solder region as a solder region;
A land area extraction step for identifying and setting the land portion from the image based on position information of the land portion of each printed circuit board created in advance,
A defect determination step of calculating a difference area between the solder area set in the solder area extraction step and the land area set in the land area extraction step, and determining pass / fail according to a feature amount of the difference area. And soldering part appearance inspection method.   The defect determination step calculates a width of a rectangular area circumscribing the difference area as a feature quantity of the difference area, and determines that the defect is determined when the width of the rectangular area is larger than a determination threshold. The soldering part appearance inspection method as described.   The soldering part appearance inspection method according to claim 6, further comprising: a horn / bridge defect determination step for determining a bridge defect and a horn defect according to a feature amount for the difference area determined to be defective in the defect determination step. .   In the tsuno / bridge defect determining step, the width of the rectangular area circumscribing the difference area is calculated as a feature value of the difference area, and the width of the rectangular area is determined for the difference area determined to be defective in the defect determining step. 9. The soldering part appearance inspection method according to claim 8, wherein a bridging defect is determined in the above case, and a horn defect is determined in the case of less than that.

Images