JPH08145903A - Inspection method of soldering and of vacancy and slip-off of lead - Google Patents

Abstract

PURPOSE: To enable systematic, quick and accurate inspection of a vacancy and slip-off of a lead by a method wherein a soldered part is illuminated with all illumination patterns necessary for a plurality of inspection items, images thereof are picked up and judgement of the quality is made in the sequence of the inspection items registered beforehand. CONSTITUTION: When inspection of soldering is started, an illumination controller 16 switches power sources 10 of an illuminating part 5 on the basis of illumination patterns registered beforehand and thereby sets upper illumination, lower illumination, all-stage illumination and angle illumination. A soldered part 3 is illuminated properly in accordance with the contents of inspection and an image of the soldered part 3 is picked up in each condition of illumination by a TV camera 4 and stored in an image memory 14. CPU 15 executes inspection of a vacancy (a tunnel or a blow hole), slip-off of a lead, bridging and others sequentially for each land 9 on the basis of inspection items, an execution sequence, quality determination steps, reference data, a form of packaging, etc., registered in an image processing part 17 and a teaching table 18, confirms completion of inspection of all lands 9 and ends the inspection of soldering.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention, after fixing the lead of an electronic component to a land portion of a printed circuit board by soldering,
The present invention relates to a method for inspecting a soldering portion, particularly a method for detecting a hole, that is, a tunnel or a blowhole as a defect in the soldering portion by image processing, and a method for inspecting a lead missing state at the soldering portion by image processing.

The tunnel is a through hole formed in a soldering portion at a through hole portion into which a lead is inserted, and the blow hole is formed on the surface of the soldering portion. It is a hollow hole.

[0003]

2. Description of the Related Art JP-A-60-25405 and Japanese Patent Publication No. 5-
Japanese Patent No. 5281 discloses an image processing technique for this type of inspection. As shown in the cross-sectional view of the illuminating device in FIG. 1, a dome-shaped (hemispherical) illumination is arranged above the soldering portion and all the lights are turned on. When picked up, as shown in the image of FIG. 2, the soldered portion appears white, and the hole, that is, the tunnel or blow hole appears as a black dot in the area corresponding to the soldered portion, so that it can be distinguished from the smooth soldering surface. .

However, there are two drawbacks. (1) The brightness (luminance) of the illumination of the illumination device changes with temperature and changes over time. Therefore, for the binarization process, a pixel in the image is compared with a constant density level (binarization threshold value), and white (ordinary soldering surface) and black (hole defect in the soldering part) are compared. It becomes difficult to divide into. In addition, the surface of the soldering portion cannot be divided into black and white at a certain density level because the reflectance of the surface of the soldering portion changes due to the soldering process, the temperature change in the solder bath, and the increase / decrease in the amount of flux applied.

(2) In the case of dip-soldering an electronic component, in the soldering portion of a straight (straight) lead component, the vicinity of the lead looks dark as shown in FIG. This phenomenon occurs because reflected light cannot be obtained from the surface of the soldering part having a tilt angle of 45 degrees or more, that is, π / 4 or more, when the hemispherical illumination is used. As a result, this soldered portion may be confused with a tunnel or blowhole.

FIG. 4 shows the state of incidence and reflection of illumination light on the surface of the soldered portion. There is a relationship of the equation φ = (π / 4) − (θ / 2) between the incident angle θ of the illumination light with respect to the upper surface of the printed board and the inclination angle φ of the surface of the soldered portion. Because of the magnitude relationship of 0 <θ, only the inclined surface with φ <π / 4 returns the reflected light toward the television camera by the illumination.

On the other hand, in Japanese Patent Application Laid-Open No. 60-25405, only by illuminating the printed circuit board by a plurality of light sources at the same time, the defect of the soldered portion is determined from the image characteristics. However, the algorithm for the discrimination is complicated.

Further, in Japanese Patent Publication No. 5-5281, since the lighting means can be switched only in units of steps, clinch mounting (insertion into a through hole of a printed circuit board with leads of electronic components being bent) is performed. It is difficult to make a significant difference in the quality judgment of the soldered part.
Further, in order to determine the characteristics from the binarized image, the solder is used in the vicinity of the leads, such as the soldering part of the straight mounting (the state where the leads of the electronic component are linearly inserted into the through holes of the printed circuit board). It is not suitable for the one that is characterized by the presence or absence of a black area that occurs (when there is a lead, the inclination of the solder surface near the lead looks black) because the black area area varies greatly depending on the setting of the binarization level. .

According to Japanese Patent Application No. 5-58297,
A lead missing inspection method using angle illumination has already been proposed, but according to the inspection method, a straight mounting quality determination step and a clinch mounting quality determination step are performed in series, and each step is performed. Since the quality of the soldered portion is determined after the illumination is switched to and the image is captured, it takes time to determine the quality.

[0010]

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to solve the above-mentioned drawbacks, and in the process of inspecting the holes of the soldering portion by image processing, the hole of the soldering portion, that is, the tunnel or the blow hole is accurately detected. It is to be able to detect.

Another object of the present invention is to provide an appropriate and efficient lead drop inspection method according to the mounting form of electronic parts, that is, straight mounting and clinch mounting.

[0012]

According to the above object, a soldering inspection method of the present invention is a soldering inspection method for determining the quality of soldering by image-processing an image of a soldering portion of a printed circuit board with a television camera and performing image processing. An illumination unit is provided which is provided around the optical axis of the television camera and is composed of a plurality of light sources, and the illumination unit includes a plurality of types of divided illumination and total illumination in which the lighting range of each light source is divided into a plurality of types. A plurality of types of illumination patterns are executable, at least a plurality of inspection items to be inspected, their execution order, and a pass / fail judgment step in each inspection item, and a plurality of types of reference value data used as a determination reference in each determination step, and further The type of illumination pattern required at this time is registered in advance, and at the time of inspection, the soldering part is illuminated with all the illumination patterns required for the above-mentioned multiple inspection items to be inspected, and the The soldering portion is captured by the Bikamera,
After storing the image for each illumination pattern, each pass / fail judgment step is executed in the order of the registered inspection items.

According to the above-mentioned inspection method, even if there is a change in the brightness of the illumination or a change in the reflectance due to the surface condition of the soldered portion, it is not affected, so that the quality of the soldered portion can be accurately judged. .

In the soldering inspection method, at least a hole inspection and a lead drop inspection are performed as inspection items.

The perforation inspection uses images of three illumination patterns of upper illumination and lower illumination as full illumination and divided illumination. First, the density value within a predetermined range of the image of all illumination is the reference value data. If it is larger than the density reference value, it is determined that there is no hole, and if there are pixels that are less than or equal to the density reference value, the first pass / fail judgment step for determining a hole candidate is executed. If it is determined in step 1 that it is a perforated candidate, then the density value of the image obtained by the upward illumination is subtracted from the density value of the image obtained by the downward illumination corresponding to the pixel having the density reference value or less in the first determination step. If the density difference is larger than the density difference reference value that is the reference value data, it is determined that there is no hole, and if there are pixels that are less than or equal to the density difference reference value, it is determined that there is a hole. .

According to this perforation inspection method, it is possible to stably detect the state of the tunnel or the blowhole as the defective soldering, and it is possible to prevent the vicinity of the straight lead from being erroneously determined as the blowhole or the tunnel.

Further, in the lead drop inspection, the soldering mounting form and the clinch direction data are registered in advance, and the total illumination, the upper illumination as the split illumination, and the angle illumination corresponding to the clinch direction as the split illumination are used. First, a straight mounting inspection is performed to determine whether or not there is a lead drop based on the mounting features of the soldering part, using the images captured by the two illumination patterns, and combining the image features of the above-mentioned illumination and the image features of all illumination. Step, and then selectively execute one of the stored clinching mounting inspection step that determines whether or not there is a lead drop by combining the stored image characteristics of the upward illumination and the image characteristics of the angle illumination. To do.

According to the lead drop inspection method, appropriate image processing can be efficiently performed according to the mounting form, and the inspection reliability is also improved.

[0019]

EXAMPLE FIG. 5 shows an outline of the soldering inspection apparatus 1. The soldering inspection apparatus 1 has a television camera 4, an illumination unit 5, and an image input control processing unit 6 which are aimed at the soldering unit 3 of the printed circuit board 2 to be inspected.

The printed circuit board 2 is placed on the XY table 7, and the movement of the XY table 7 causes the soldered portions 3 of the printed circuit board 2 to be in the visual field of the television camera 4 in a predetermined order. Sent out. Here, the soldering part 3
Is attached to the lead 8 of the electronic component and the land 9 as a part of the circuit pattern from the back side of the electronic component, and has a Mt. Fuji shape or a hemispherical shape. The television camera 4 is directed to the soldering portion 3 to be photographed above the soldering portion 3 in order to image the outer shape of the soldering portion 3.

The illumination section 5 is for irradiating the soldering section 3 to be inspected under an appropriate illumination condition according to the inspection content, and includes a plurality of point light sources or linear light sources. It is composed of a light source 10. These light sources 10
Are, for example, LEDs, and a plurality of them are arranged in the latitudinal direction above the soldering portion 3 by, for example, a dome-shaped holder 11 formed around the optical axis of the television camera 4 as a center. The holder 11 forms a window 12 for imaging at the top portion. Although the illumination unit 5 has eight stages here, it may have any number of stages as long as it has two or more stages. Incidentally, although not shown, the holder 11 may have a cylindrical shape or a conical shape.

As described above, the illuminating section 5 comprises a plurality of light sources 10 arranged in the dome-shaped holder 11 in the longitude direction which is the circumferential direction and in a plurality of stages in the latitude direction. It is possible to execute a plurality of types of illumination patterns such as stage-divided illumination in which the lighting range of is divided by the latitude line, divided illumination such as angle-divided illumination in the meridian line, and total illumination. The illumination unit 5 is not limited to the dome shape, and may be formed in a cylindrical shape or a conical shape centered on the optical axis of the television camera 4.

Next, the image input control processing unit 6 includes an A / D converter 13 for converting image information of the television camera 4, an image memory 14 for storing image data, and 1.
A memory (buffer) 31 for storing the inspection result (defective flag) of one printed circuit board 2, a sequencer (controller) of a printed circuit board manufacturing line for transmitting / receiving the carry-in / out of the printed circuit board 2 and transmitting the quality of the test result.
Communication line 20 for 35, keyboard 21, mouse 2
2, a CRT 23, a printer 24, a floppy disk drive 25, a hard disk drive 26, and the like, and a CPU unit 15 for controlling the same.

Further, the image input control processing unit 6 is
An illumination controller 16 for controlling the blinking state of the light source 10, an image processing unit 17 for processing image data, a teaching table 18 for inputting various data, and X.
It is assembled by a Y board 7, a board clamper (fixing section) 28, a printed board control section 19 for controlling the movements of the board feeding section (loading / unloading section) 29, and the like.

The teaching table 18 includes a board name, a board size, a land No. The land position from the reference position of the printed circuit board 2, the land diameter, the position correction mark position, the number of lands, the mounting form, that is, the straight mounting (code No. 0) or the clinch mounting or the clinch direction (code Nos. 1 to 8), etc. Original information for the printed circuit board 2 (input from a mouse / keyboard, etc. during teaching work before inspection), and visual field information (total number of visual fields, visual field positions, total number of lands within visual field, visual field) created by a visual field dividing program in this teaching work. Inner land N
o. ) Or an illumination pattern in each determination step and reference value data to be a determination reference value in each determination step.

Next, FIG. 6 shows a cross section of the soldered portion. The leads 8 of the discrete component project from the through holes 30 of the printed circuit board 2 and are connected to the conductive lands 9 of the circuit pattern by the soldering portions 3. As indicated by the dotted line, the tunnel 32 as the defective soldering is formed in the through hole 30 in a state of penetrating from one surface of the soldering portion 3 to the other surface. In addition, the blow hole 33 as a defective soldering is
As shown by a solid line, the surface of the soldering portion 3 is formed as a depression toward the center of the soldering portion 3.

FIG. 7 shows a flow chart of the soldering inspection system. The sequential operation of this flowchart is executed by the image input control processing unit 6.
First, after the start of the program, the printed board 2 to be inspected is sent to the inspection station by the board feeding unit 29 when the board is loaded, the printed board 2 is fixed by the board clamper 28, and then in the board inspection step, the printed board is printed. The soldering inspection of No. 2 is executed, and in the next step of determining the inspection result, based on the inspection result stored in the memory 31 of the CPU unit 15, the non-defective product and the defective product of the soldering portion 3 of the printed circuit board 2 are determined. To be judged.

Here, in the case of a defective product, the contents of the defect are displayed on the CRT 23, and the printer 24 displays
The content of the defect is printed, and the NG signal is sent to the CPU unit 1.
After being output to the line sequencer 35 by 5, the defective product is further discharged to the next process. During ejection, the substrate clamper 28 is disengaged and ejected by the substrate feeding section 29.

Thereafter, the defective printed circuit board 2 is stored in the NG stocker by the line sequencer 35. In the case of a non-defective product, an OK signal is output by the CPU unit 15 to the line sequencer 35 and then discharged to the next step. The above operation is repeated until the inspection is completed, and when the inspection is completed for all the printed circuit boards 2, the program is completed.

In addition to the case where the NG signal is simply NG for each board, the NG signal is corrected by the solder correction machine in the next step, or the liquid level of the solder tank in the previous step is adjusted (for example, the tunnel 32 is provided at a specific portion of the printed board 2). If the number of bridges is large, the molten solder liquid level in the solder bath may be lowered. No. And NG items may be sent.

Next, FIG. 8 shows a flow chart of the board inspection of FIG. After the start of the inspection, as a position correction operation of the printed board 2 that has been carried in, the position of a position correction mark 36, which will be described later, which has been taught in advance is recognized by image processing, and the amount of deviation from the normal position is detected, Then X
-The Y table 7 is driven to move the table while performing the position correction for the displacement amount, and the soldering portion 3 to be inspected is placed in one visual field, and the soldering inspection of one visual field is performed in the visual field. The procedure is repeated and repeated until the inspection for all fields of view is completed. The data to be stored in this field of view depends on the field of view information created by the field of view program during teaching work.

The position correction is an operation of reading the position correction mark 27 on the printed circuit board 2 and detecting the amount of displacement and the inclination of the printed circuit board 2 as shown in FIG. One visual field for inspection is set in accordance with the visual field information, and the position is corrected for the amount of positional deviation and the inclination. This visual field information is created by a visual field division program executed in the teaching work.

Next, FIG. 10 shows a flowchart of the soldering inspection for one field of view of FIG. After the start of the soldering inspection, in the imaging step, the illumination controller 16 performs the illumination switching operation by switching the light source 10 of the illumination unit 5, and sets the upper illumination, the lower illumination, the full-stage illumination, and the angle illumination. The TV camera 4 images the soldering part 3 for each of these illumination conditions. In this way, the required number of images are captured by the television camera 4 and stored in the image memory 14 as image data.

Here, the upper illumination means the light source 1 of the illumination unit 5.
Only the first stage of 0, or the first and second stages or 1
The third to third stages are turned on, and the downward illumination is an illumination state in which all stages of the light source 10 of the illumination unit 5 except the upper illumination are turned on, and the total illumination is Is an illumination state in which all the light sources 10 are turned on.

Further, the angular illumination means a line that divides the illumination unit 5 into eight equal parts in a plane from the illumination center (four of them are parallel or perpendicular to the XY axis), and each meridian is the center. Among the light sources 10 within the range of ± 45 degrees, the one excluding the upper illumination is an angle illumination block, and is an illumination state in which one block or two selected blocks are turned on.

FIG. 11 shows each illumination state of the illumination unit 5.
The angle illumination is set to 8 blocks because the clinching direction of the lead 8 is normally 8 directions. When one block is turned on, each block takes charge of each clinching direction, and it is necessary to take a screen image with a maximum of eight times of angle illumination. The lighting in 2 block units is for shortening the imaging time, and 2 blocks are in charge of 2 clinch directions at the same time.
Timed angle illumination screen imaging is required. In FIG. 11, (1) illustrates the point light source 10 and (2)
1 exemplifies a light source 10 'having a line (arc) shape.
The light source 10 may be any of (1) and (2).

The pattern of the angle illumination lighting is determined based on the field information within one field of view and the above-mentioned original information of the printed circuit board 2 depending on how many kinds of clinch directions are within one field of view. (Pattern) may be calculated in advance and calculated in a visual field division program for teaching work and input to the teaching table 18, or may be calculated by calculation before lighting, or
You may always get 4 or 8 angle illumination screens for all clinch directions.

After completion of the above-mentioned imaging step, 1
In the inspection step for each land, a hole inspection, a lead drop inspection, and a bridge inspection (which means a soldering failure in which the solder is bridged between adjacent soldering portions 3 and leads 8). Inspections and the like are sequentially performed, and the completion of all lands 9 is confirmed, and the soldering inspection is completed.

Next, FIGS. 12 and 13 show flow charts of the perforation inspection of the present invention. In the flow charts in both figures, the parts with the same symbols of the circled children C1 and C2 are connected. The illumination controller 16 and the TV camera 4 are all illuminations, that is, images by the light sources 10 on all latitude lines, upper illuminations, that is, top images, for example, images by the light sources 10 of the upper four steps (upper four latitude lines), and lower illuminations, that is, skirt sides Images from the light sources 10 in four stages (lower four latitude lines) are separately captured and stored in the image memory 14. After that, the image processing unit 17
Based on the flowcharts of FIGS. 12 and 13, necessary image processing is performed by the following steps.

FIG. 14 shows the frequency distribution of each pixel constituting the hole 32, the blow hole 33, or the non-defective soldering section 3 as a soldering failure based on the entire illumination image. Here, for convenience, the density level of the image is 256 gradations, and the tunnel 32 and the blowhole 3
3. The total frequency of the soldering portion 3 is normalized so as to take a constant value. As can be seen, it can be seen that the tunnel 32 and the blow hole 33, which indicate defective soldering, are composed of pixels having a density smaller than the predetermined density level A as the density reference value.

Then, first, as the first pass / fail judgment step of the perforated inspection, all illumination pixels lower than a certain density level A in the predetermined range, that is, the range of the soldering section 3 for all illumination images are tunneled or blown. It is extracted as a candidate for the hole 33. In this step, the density level A
If the pixels have the following density levels, even a normal soldering portion 3 may be a candidate for defective soldering. Here, the defective candidate pixel address a1 is stored, and the number of pixels is also counted.
If it is 1 or more, it is considered as a defect candidate.

The soldering portion 3 which is a candidate for this soldering failure is a dark soldering portion (a portion which appears dark because reflected light does not return near the lead when the soldering portion 3 shown in FIG. 15 has an inclination angle of 45 degrees or more. ).

In the dark portion of the soldering portion 3 having an inclination angle of 45 degrees or more, if it is specular reflection, no reflected light is returned to the television camera 4, but normally the soldering surface is coated with flux or a matting agent. Since it is diffusely reflected, some reflected light is returned. Therefore, the inclination angle of the soldering portion 3 is π / 4 (φ = 4
The distribution of the reflection intensity for each illumination angle θ at the point (5 °) is obtained as shown in FIG. When this distribution characteristic is used, as shown in FIG. 17, in the case of 45 ° ≦ φ ≦ 90 °, in any pixel of the solder dark part (tilt angle π / 2 or more), [brightness of image of upper illumination <lower illumination Brightness] is established.

Next, FIG. 18 shows an image by upward illumination,
The density map of each corresponding pixel of the image by downward illumination is shown. In the solder dark portion, [(density of image of lower illumination-density of image of upper illumination)> density difference level B] is satisfied.

The second pass / fail judgment step of the perforated inspection is
For the pixel (pixel address a1) that is a candidate for the hole 32, that is, the tunnel 32 or the blowhole 33 in the first pass / fail judgment step, the corresponding pixel by the downward illumination of the position on the image corresponding thereto and the image by the upward illumination are 2 The difference in brightness (density value) of the two corresponding pixels is compared with a predetermined density difference level B as a density difference reference value, and the difference (lower illumination-
Only pixels whose upper illumination) is smaller than the density difference level B are judged to be holes, that is, the pixels forming the tunnel 32 or the blow hole 33. Also here, the number of pixels determined to be perforated is counted, and when the counted number is the reference number S2 or more, it is determined to be a perforated defective candidate.

In the third pass / fail judgment step, with respect to the pixel (pixel address a1 ') judged to be a hole, that is, a candidate for the tunnel 32 or the blow hole 33, the S / N ratio between this pixel and the peripheral pixel of the corresponding pixel by all the illuminations. The ratio is calculated, and only the pixels having the S / N ratio larger than a predetermined threshold value E are regarded as the pixels forming the tunnel 32 or the blowhole 33. Also here, the number of pixels determined to be perforated is counted, and when the counted number is the reference value S3 or more,
Finally, it is decided that there is a hole.

By performing this process, the solder flux that looks dark will not be erroneously recognized as the tunnel 32 or the blow hole 33. That is, the first of image processing
The pixels determined to be the tunnel 32 or the blow hole 33 in the quality determination step and the second quality determination step may include those in which the flux is thickly adhered to the surface of the soldering portion 3.

FIG. 19 shows the state in which the thick flux 34 is attached to the surface of the soldering portion 3 and the corresponding relationship between the images. With such a flux 34,
First of all, in the first pass / fail judgment step, even if all lights are illuminated,
Since it looks dark, it becomes a candidate for the tunnel 32 or the blowhole 33, and in the second pass / fail judgment step, the portion of the flux 34 absorbs both the illumination light of the upper illumination and the illumination of the lower illumination. Since there is almost no difference in the density with the upper illumination, the difference becomes smaller than the density difference level B, and the flux 34 is determined as the tunnel 32 or the blow hole 33 as it is.

Therefore, the significant difference between the tunnel 32 or the blowhole 33 and the flux 34 is that the dip slope rate from the bright portion (soldered portion) to the dark portion (hole or flux) in the image of all illumination is different. Focus on it. Figure 2
0 is the threshold value E which is the slope ratio reference value which is the relationship between the density of the screen and the dip slope ratio, that is, the S / N ratio of the screen of all-stage illumination corresponding to the blowhole 33 on the soldering surface and the thick flux 34 on the soldering surface. It shows with.

The dip slope rate (S / N ratio) is defined as follows. Here, f is the density value of the pixel of interest of the image by full illumination, and f ′ is the density value of the pixels in the vicinity of the pixel of interest.

S / N ratio = | f−f ′ | / min (f, f ′) ... (1)

As shown in FIG. 21, the above equation (1)
The relative position and the set number of neighboring pixels to be inspected are
If one of the predetermined S / N ratios with the neighboring pixels is larger than the threshold value E, the pixel is regarded as one of the pixels forming the tunnel 32 or the blowhole 33.

In the above equation (1), the denominator is min (f,
f '), and if there are holes, the S / N ratio is more emphasized. Note that this denominator originally has a meaning for dealing with a change in the amount of light, and the following formula may be used instead of the formula (1).

S / N ratio = | f−f ′ | / max (f, f ′) ... (1) -1 S / N ratio = | f−f ′ | / f・ ・ ・ ・ ・ ・ ・ ・ ・ (1) -2 S / N ratio = | f-f '| / f' ・ ・ ・ ・ ・ ・ ・ ・ ・ (1) -3 S / N ratio = | f−f ′ | / (f + f ′) (1) -4 Since this S / N ratio involves division and takes time to calculate, Instead, a difference value such as | f−f ′ | or a differential value may be used instead.

Further, the following measures are required against the reflectance of the soldering surface, the temperature change of the light source, and the secular change.
FIG. 22 is a graph of the dimming characteristic, showing what the density value of a pixel of the soldering part 3 becomes when the light is dimmed. From a certain value (v, v) It is indicated by a radially extending line segment. This density value v is a density value that the image memory 14 that has taken a picture has when all the light sources 10 are turned off.

From this graph, the average density value of the pixel representing the soldering portion 3 is obtained for the pixel as shown in FIG. 23, and the ratio D between that value and a predetermined constant density level C is obtained. You can see how faint it is. Strictly speaking, the level V is subtracted from each of the average density value and the level C in advance, and then the ratio is obtained. If the extinction ratio (ratio D) is known, all the pixels of the image by all the light sources 10 are dimmed due to the dimming characteristic. In consideration of this, the density level A and the density difference level B as the determination level are taken into consideration. If the ratio D is multiplied by and A × D and B × D are used, the inspection can be performed without being affected by the dimming. Then, this method can cope with a change in the reflectance of the soldering surface.

According to the soldering / perforation inspection as described above, even if there is a change in the brightness of the illuminating part 5 or a change in the brightness of the surface of the soldering part 3, they are affected in the inspection process. Therefore, the inspection of the blow hole 33 and the tunnel 32 of the soldering portion 3 can be detected stably. Moreover, since the vicinity of the straight lead 8 is not mistakenly determined as the blow hole 33 or the tunnel 32, the reliability of the inspection is improved.

The light sources 10 constituting the illumination unit 5 are different from each other, for example, R (red), G (green), B.
By using a multicolor light source capable of emitting three (blue) hue lights and a combined light thereof, and using a color camera as the television camera 4, two of the three R, G, and B light emitting elements are illuminated by upward illumination. By illuminating two of the three light emitting elements in a combination of the lower illumination and the upper illumination, the upper, lower and all illuminations can be simultaneously captured by the color camera in the R, G and B images.

The following table is an example of this combination method,
For example, in case (1), the upper illumination image can be input to the R image, the lower illumination image can be input to the B image, and the entire illumination image can be input to the G image.

[0060]

Next, FIG. 24 shows a flowchart of the lead missing inspection of the soldering portion. This flowchart is executed for each land 9 in one visual field. After the start, first, the land N in the visual field based on the visual field information
o. Based on the original information, the land No. Data such as the clinch direction, land position, and land diameter associated with each other are extracted, straight mounting and clinch mounting are identified based on the clinch direction, and the position and size of the inspection window are determined by the land position and land diameter.

FIG. 25 shows the clinch direction and the code No. for specifying this. 0-8 (Code No. 0-
(Straight mounting for 8: 0, clinch mounting for 1-8)
26 shows an example of the correspondence with the code No. in the clinch direction. The angle illumination block corresponding to is shown. Illumination is applied from the opposite direction to the clinch direction. When lighting is performed in units of two blocks, the code No. 1-2, 3
-4, 5-6, 7-8 and code N
o. There are combinations of 2-3, 4-5, 6-7, and 8-1.

27 and 28 show visual field information (image) and original information (code No. and land N in the clinch direction).
o. ) As the clinch direction within one field of view, land No. The angle illumination block which lights up corresponding to is shown. Figure 2
In No. 7, the light is turned on in 1 block units, and the code No. in the clinch direction is displayed. Three angle illumination images will be captured for every 5, 7, and 8. In FIG. 28, lighting is performed in units of 2 blocks, and the code No. in the clinch direction is displayed. One image was picked up in Nos. 7 and 8, and code No. One image is taken with 5, total 2
One angle illumination image will be picked up.

FIG. 29 shows defective products with missing leads in straight mounting and non-defective products of types 1, 2, and 3 together with an image of full illumination and an image of upward illumination with respect to the cross-sectional shape of the soldering portion 3. . Further, FIG. 30 shows defective lead missing products at the time of clinch mounting and non-defective products of types 1, 2, and 3 along with an image of upper illumination and an image of angle illumination in correspondence with the cross-sectional shape of the soldering portion 3. ing.

Next, FIG. 31 shows a flow chart of the lead drop inspection of straight mounting for one land. In the first pass / fail judgment step, highlight portion extraction and judgment are sequentially performed. The highlight part extraction utilizes the property that a highlight part (bright spot) is generated in the center part of the lead 8 in the image of the upper illumination when the lead mounting state is straight mounting.

Therefore, first, in the highlight portion extraction step, using the image of the upper illumination, an area higher than a certain binarization level in the teaching table in advance is used as a window (window: land, center or land in circle or square). The size is set in proportion to the diameter.), And the area S of the highlight portion there is calculated, and the ratio V1 to the land area S0 is calculated from the calculation formula V1 = S / S0.

Then, in the determination step, by using the first area ratio reference value C1 registered in advance as reference value data and comparing V1 and C1 with each other, if V1 <C1, there is no lead drop. Processed as non-defective, but V1 ≧
If it is C1, it is determined as a defective product candidate with missing lead, and the process proceeds to the next second pass / fail judgment step. Here, type 1 of FIG. 29,
3 is determined to be non-defective.

In the second quality judgment step, the absolute value of the difference value is calculated and judged sequentially. The straight mounted good product (type 2 in FIG. 29) that is determined as the lead dropout candidate in the first pass / fail determination step occurs in the case where the tip of the lead is illuminated by upward illumination. In this type 2, since the inclination angle of the soldering part 3 near the lead is large, and it looks dark as a soldering dark part under all lighting, as shown in FIG. 32, it is acceptable to use this brightness change pattern, that is, the density distribution. Identify Therefore, a pattern close to the similarity of the lead missing density pattern is determined as a defect candidate.

Therefore, in the process of calculating the absolute value of the difference value, the absolute value or the difference value of the difference values of the inter-pixel density values searched in two directions, four directions, or eight directions from the lead center using the image of all the illuminations is calculated. The sum V2 of absolute values is calculated, and it is compared with the difference value reference value C2 registered in advance as reference value data. When V2 ≧ C2, it is determined as a good product with no lead missing, and when V2 <C2, it is determined as a defective product with missing lead. Here, the type 2 shown in FIG. 29 is determined to be non-defective.

Since the absolute value of the difference value or the sum V2 of the absolute values of the difference values increases in proportion to the large land area, V2 / S0 divided by the land area S0 is compared with the difference value reference value C2. You may. As a result of the determination, the content determined to be a non-defective product or a defective product is written in a defective flag described below and stored in the memory 31.

This defective flag is used as follows.
First, all are cleared at the beginning of the inspection of one board. Next, in the inspection of one field of view, if defective items occur in the inspection land, they are stored. When the inspection of one field of view is over,
For lands whose defective bits (described later) are all off, all their defective flags are cleared (0). By doing so, it is possible to prevent the data amount of the defective flag from increasing.

When one board inspection is completed, referring to the defect flag (the number of defective lands) in the memory 31, which land No. You can see what kind of defect there was.

Here, the defect flag will be described. The defect flag is a number handled as a hexadecimal number, for example, and has a size of 32 bits, for example. The defective flag is the inspection land No. , Bridge opponent land No. , Composed of defective bits. Here, the inspection land No. 12 bits for the other land of the bridge. 12 bits for the inspection land and 8 bits for the defective bit. And bridge opponent land No. Can be expressed as 0 to 4095, and eight types of defective bits can be recorded, but the size may be increased according to the total number of lands and the number of inspection items.

Inspection land No. : Land N to be inspected
o. (1 to 4095) Bridge opponent land No. : Land No. which is the opponent of the bridge. (1 to 4095) When there is no bridge (the defective bit of the bridge is 0), 0 is entered. Defective bit: One bit is allocated for each inspection item,
Bit is bad if it is bad and bit is off if it is good.

Next, FIG. 33 shows a flow chart of lead missing inspection for clinch mounting for one land. In the first pass / fail judgment step here, highlight portion extraction and judgment are performed. This is the same as the first pass / fail judgment step in the straight mounting of FIG. In the highlight part extraction process, a region higher than a certain binarization level is obtained in the window using the image of the upper illumination, the area S of the highlight part is obtained, and the land area S
The ratio V1 to 0 is calculated from the calculation formula V1 = S / S0.

Then, in the next determination step, C1 is used as the area ratio reference value, and if V1 <C1 by comparing the magnitudes of V1 and C1, it is processed as a non-defective product with no lead drop, but V1 ≧ If it is C1, it is determined as a defective product candidate with missing lead, and the process proceeds to the next second pass / fail judgment step. In this first quality determination step, the type 1 shown in FIG. 30 is determined to be a good product.

The second pass / fail judgment step is a determination of the circularity of the highlighted portion, and as shown in FIGS. 34 and 35, it is a non-defective product for clinch mounting which is a lead dropout candidate in the first pass / fail judgment step. This corresponds to the processing for excluding type 3 from the lead missing candidates. The circularity V2 of the highlight portion of the image by the upward illumination used in the highlight portion extraction is close to a circle (smaller than the circularity reference value C2 registered in advance as reference value data. For example, C2 = 1.
If it is 5), it is a non-defective product with no lead missing, and if not, it is determined as a defective product candidate with lead missing.

Here, the circularity V2 of the highlight portion is obtained by the equation V2 = L1 / L2. L1 and L2 in the formula are
As shown in FIG. 36, it corresponds to the long axis and the short axis based on the clinch direction. In this determination, if V2 ≧ C2, it is determined as a non-defective product with no lead missing.

When V2 <C2, the process proceeds to the next third quality determining step, and the process proceeds to the step of counting and determining the area of the highlight portion by angle illumination. In the second pass / fail judgment step, the non-defective product of the clinch mounting determined to be the lead missing candidate is the type 2 in FIG. 30 where V2 <C2. The determination by the angle illumination is to detect the lead portion illuminated by illuminating the illumination light from the direction opposite to the clinch direction. As shown in FIG. 37, a lead missing candidate is a case where there is no shining area in a fan-shaped, triangular, or quadrangular window.

Therefore, an image of the angle illumination (an image in which the angle illumination block 180 degrees in the direction opposite to the clinching direction of the inspection land is lit. The lighting state is shown in FIGS.
As already exemplified in the above, there is a case of lighting by one block unit and a case of lighting by two block unit. ) In the step of counting the highlight area, the fan-shaped (clinched leads 8) having an opening angle of 45 to 90 degrees (stored in the teaching table) having a size set in proportion to the land diameter. (Set in a certain direction), or the area S2 of the highlight portion in the triangular or quadrangular window is obtained, and the ratio V3 to the land area S0 is calculated by the formula V3 = S2 /
It is calculated by S0.

Next, using the image of the angle illumination, V3 and the second area ratio reference value C registered in advance as reference value data.
3 is compared, and when V3 ≧ C3, it is processed as a non-defective product with no lead missing, but when V3 <C3, it is determined as a lead missing defective product. Finally Figure 30
Type 2 is also judged to be a good product with clinch mounting, and only the true lead missing is judged to be a defective product. Also in this case, the determination content is written in the defect flag and stored in the memory as in the case of the straight mounting.

[0082]

According to the soldering inspection method of the present invention, the judgment step for each inspection item, the judgment reference value such as the illumination pattern, the density reference value, the density ratio reference value, the clinch mounting,
Register the mounting form such as straight mounting in advance,
Before the soldering inspection, images are taken with various illumination patterns, and after each image is stored, inspection steps such as perforation inspection and lead drop inspection are performed, so the necessary inspection steps can be performed in an orderly and quick manner. .

According to the soldering / perforation inspection method of the present invention, even if there is a change in the brightness of the illumination or the brightness of the surface of the soldered portion, these are not affected in the inspection process. Stable detection of blowholes and tunnels in soldered parts. Further, since the vicinity of the straight lead is not mistakenly determined as a blow hole or a tunnel, the reliability of the inspection can be improved.

According to the lead missing inspection of the present invention, by switching the illumination direction of the illuminating section, it is possible to obtain an image that highlights the features of the lead missing defect of the straight mounting and the features of the lead missing defect of the clinch mounting. Moreover, since the straight mounting and the clinch mounting are separately inspected, the feature extraction can be efficiently inspected, and the comprehensive determination of the features of the two illumination images increases the ability to classify the quality of the soldered part, thus improving the inspection rate. improves.

[Brief description of drawings]

FIG. 1 is a side view of a conventional inspection state of a soldering portion.

FIG. 2 is an explanatory diagram of an image of a tunnel or a blowhole as a holed portion.

FIG. 3 is an explanatory diagram of a correspondence relationship between a surface of a soldering portion having an inclination of 45 degrees or more and images corresponding to the surface.

FIG. 4 is an explanatory diagram of an incident / reflection state of illumination light.

FIG. 5 is a configuration diagram of a soldering inspection device.

FIG. 6 is an enlarged cross-sectional view of a soldering portion.

FIG. 7 is a flowchart of a soldering inspection system.

FIG. 8 is a flowchart of board inspection.

FIG. 9 is a plan view of the relationship between the land of the printed circuit board and the field of view.

FIG. 10 is a flow chart of a one-view soldering inspection.

FIG. 11 is an explanatory diagram of lighting states under full illumination, upper illumination, lower illumination, and angle illumination.

FIG. 12 is a flowchart of a third pass / fail judgment step of the perforated inspection and a defect flag bit set.

FIG. 13 is a flowchart of a method for detecting a soldered portion.

FIG. 14 is a graph of a density distribution of an image by illumination at all stages.

FIG. 15 is an explanatory diagram of a solder dark portion when the surface of the soldering portion is 45 degrees or more.

FIG. 16 is a graph showing a distribution of reflection intensity with respect to each illumination angle at a solder surface inclination angle of about 45 degrees in a soldering part.

FIG. 17 is a graph showing a distribution of reflection intensity with respect to each illumination angle when the inclination angle of the surface of the soldered portion is 45 degrees or more.

FIG. 18 is a graph of a density map of corresponding pixels in an image illuminated by upper illumination and an image illuminated by lower illumination.

FIG. 19 is an explanatory diagram of a flux adhesion state on a soldering surface and an image thereof.

FIG. 20 is an explanatory diagram of a concentration and a dip slope rate (S / N ratio) corresponding to a blow hole on the surface of a soldering portion.

FIG. 21 is an explanatory diagram of relative positions and set numbers of neighboring pixels with respect to a target pixel.

FIG. 22 is a graph of a dimming characteristic of a soldering surface.

FIG. 23 is an explanatory diagram of pixels for calculating an average density.

FIG. 24 is a flowchart of a lead drop inspection.

FIG. 25 shows a code No. in the clinch direction. FIG.

FIG. 26 is an explanatory diagram of an angle illumination block corresponding to a clinch direction.

FIG. 27 is an explanatory diagram of a clinch direction in one field of view when lighting in a block unit and a lighting angle illumination block corresponding to the clinch direction.

FIG. 28 is an explanatory diagram of a clinch direction in one visual field when lighting is performed in units of two blocks and a lighting angle illumination block corresponding thereto.

FIG. 29 is a cross-sectional shape of a soldering portion for straight mounting,
It is explanatory drawing of the image of all-stage illumination and the image of upper illumination.

FIG. 30 is an explanatory diagram of a cross-sectional shape of a soldering part of clinch mounting, an image of upward illumination, and an image of angle illumination.

FIG. 31 is a flowchart of a lead drop inspection for straight mounting.

FIG. 32 is an explanatory diagram of a density distribution based on images of full-stage illumination.

FIG. 33 is a flowchart of a lead missing inspection for clinch mounting.

FIG. 34 is an explanatory diagram of circularity based on an image of upward illumination.

FIG. 35 is an example of a frequency graph of circularity of a lead missing defect.

FIG. 36 is an explanatory diagram of calculation of circularity.

FIG. 37 is an explanatory diagram of a window setting position with respect to the soldering portion.

[Explanation of symbols]

 1 Soldering Inspection Device 2 Printed Circuit Board 3 Soldering Section 4 Television Camera 5 Illumination Section 6 Image Input Control Processing Unit 7 XY Table 8 Lead 9 Land 10 Light Source 11 Holder 12 Window 13 A / D Converter 14 Image Memory 15 CPU Unit 16 Lighting controller 17 Image processing unit 18 Teaching table 19 Printed circuit board control unit 20 Communication line 21 Keyboard 22 Mouse 23 CRT 24 Printer 25 Floppy disk drive 26 Hard disk drive 27 Position correction mark 28 Board clamper 29 Board feed section 30 Through hole 31 Memory 32 tunnel 33 blow hole 34 flux 35 line sequencer

Claims (10)

[Claims] 1. A soldering inspection method for determining the quality of soldering by image-processing a soldered portion of a printed circuit board with a television camera and performing image processing to determine a plurality of soldering portions provided around the optical axis of the television camera. An illumination unit composed of a light source is formed, and this illumination unit can execute a plurality of types of illumination patterns of divided illumination and total illumination in which the lighting range of each light source is divided into a plurality of types, and at least a plurality of inspection items to be inspected. And the order of execution, the quality judgment step in each inspection item, and a plurality of types of reference value data used as a determination criterion in each determination step, and the type of illumination pattern required at that time are registered in advance. After illuminating the soldering part with all the illumination patterns required for a plurality of inspection items to be inspected and imaging and storing the image for each illumination pattern, the above-mentioned registration is performed. A soldering inspection method, characterized in that each pass / fail judgment step is executed in the order of inspection items. 2. The soldering inspection method according to claim 1, wherein at least hole inspection and lead drop inspection are performed as inspection items. 3. The perforated inspection uses images of three illumination patterns of upper illumination and lower illumination as full illumination and divided illumination, and first, a density value within a predetermined range of the image obtained by the full illumination is the reference. If it is larger than the density reference value which is the value data, it is determined that there is no hole, and if there are pixels that are less than the density reference value, it is determined as a hole candidate. If it is determined in step 1 that it is a perforated candidate, then the density value of the image obtained by the upward illumination is subtracted from the density value of the image obtained by the downward illumination corresponding to the pixel having the density reference value or less in the first determination step. If the density difference is larger than the density difference reference value that is the reference value data, it is determined that there is no hole, and if there are pixels that are less than or equal to the density difference reference value, it is determined that there is a hole. Characterized contract The soldering inspection method according to claim 2. 4. In the lead drop inspection, soldering mounting form and clinch direction data are registered in advance, and all lighting,
Using images captured by three illumination patterns of upward illumination as divided illumination and angular illumination corresponding to the clinch direction as divided illumination, first, based on the mounting form of the soldering part, the characteristics of the image by the above illumination and the total A straight mounting inspection step that determines the presence / absence of a lead by combining the characteristics of the image by illumination, and the determination of the presence / absence of a lead by combining the stored image feature by the above-mentioned illumination and image feature by the angle illumination. 4. The soldering inspection method according to claim 2, wherein any one of the clinch mounting inspection step and the clinch mounting inspection step is performed. 5. A soldering inspection method for determining the quality of soldering by image processing of an image of a soldering portion of a printed circuit board by a television camera, wherein a plurality of soldering parts are provided around the optical axis of the television camera. An illumination unit composed of a light source is formed, and this illumination unit is capable of executing a plurality of types of illumination patterns of divided illumination and total illumination in which the lighting range of each light source is divided into a plurality of types, and at least a plurality of types for the quality judgment. The reference value data of the above is registered in advance, and at the time of inspection, the soldering portion is illuminated with three illumination patterns of the above-mentioned full illumination and divided illumination, that is, the upper illumination and the lower illumination, and the image is captured and the image for each illumination pattern is stored. After that, first, if the density value of the image by all the illuminations is larger than the density reference value which is the reference value data, it is determined that there is no hole, and if there are pixels below the density reference value, it is a hole candidate. When the first pass / fail judgment step for determination is executed and it is determined to be a perforated candidate in this first pass / fail determination step, it corresponds to the pixel which is below the density reference value in the above first pass / fail determination step. If the density difference obtained by subtracting the density value of the image by the upper illumination from the density value of the image by the lower illumination is larger than the density difference reference value that is the reference value data, it is determined that there is no hole, and the density difference is less than or equal to the reference value. If there is a pixel of, a second pass / fail judgment step for determining that there is a hole is executed. 6. If the second pass / fail judgment step determines that there is a hole, then the brightness of the image by all the illumination corresponding to the pixels that are below the density difference reference value in the second pass / fail judgment step. Slope ratio from dark area to dark area (S / N ratio)
If the inclination ratio (S / N ratio) is equal to or smaller than the inclination ratio reference value which is the reference value data, it is determined that there is no hole, and if there is a pixel larger than the inclination ratio reference value, the true value is calculated. 6. The perforation inspection method according to claim 5, wherein a third pass / fail judgment step for determining perforation is performed. 7. A lighting range of an upper illumination by using a multicolor light source capable of emitting three different hue lights and their combined light for each light source forming the illumination unit and using a color camera as a television camera. In the upper light source, two of the three hue lights are respectively lit up, and in the lower light source which is a lighting range of the downward illumination, two of the three hue lights are lit together in a different combination from the upper light source. 7. The method according to claim 5, wherein the color camera performs an image with three illumination patterns of upward illumination, downward illumination, and total illumination at one time, and simultaneously obtains three images with each illumination pattern. Hole inspection method. 8. A soldering inspection method for determining the quality of soldering by image-processing an image of a soldering portion of a printed circuit board by a television camera, and a dome comprising a plurality of light sources centering on an optical axis of the television camera. Form a lighting unit, and this lighting unit can execute multiple lighting patterns of divided lighting and total lighting in which the lighting range of each light source is divided into multiple types, and at least solder mounting form, clinch direction data, and pass / fail judgment For each inspection, a plurality of types of reference value data are registered in advance, and at the time of inspection, a soldering portion is formed by three illumination patterns of the above-mentioned all illumination, upper illumination as divided illumination, and angle illumination corresponding to the clinch direction as divided illumination. After illuminating and capturing and storing the image for each illumination pattern, first, based on the mounting form of the soldering part, the features of the image by the above illumination and total illumination A straight mounting inspection step for determining presence / absence of a lead drop by combining the features of an image due to light, and a determination of presence / absence of a lead by combining the image feature by the above-mentioned illumination and the image feature by the angle illumination. A lead drop inspection method characterized by selectively performing one of a clinch mounting inspection step and an inspection step. 9. In the straight mounting inspection step, an area of a highlight portion within a predetermined range of the image by the upper illumination is extracted, and an area ratio to the predetermined range is calculated, and the area ratio is the reference value data. If it is smaller than the first area ratio reference value that is
If it is equal to or larger than the area ratio reference value, the first pass / fail judgment step for determining a lead drop candidate is executed. If the lead drop candidate is determined in this first pass / fail judgment step, the next image from all illuminations is selected. The maximum absolute value of the difference values of the inter-pixel density values searched for in two or more directions from the read center or the sum of the absolute values of the difference values is calculated, and this value is greater than or equal to the difference value reference value which is the reference value data. 9. The lead drop inspection method according to claim 8, wherein if there is no lead drop, it is determined that there is no lead drop, and if it is smaller than the difference value reference value, the second pass / fail judgment step that determines that there is a lead drop is executed. 10. In the clinch mounting inspection step, the area of the highlight part within a predetermined range of the image by the above-mentioned illumination is extracted, and the area ratio to the predetermined range is calculated, and the area ratio is the reference value data. If it is smaller than the first area ratio reference value that is
If it is equal to or larger than the area ratio reference value, the first pass / fail judgment step for determining a lead drop candidate is executed, and if the lead drop candidate is determined in this first pass / fail judgment step, next, from the image by the above-mentioned upward illumination. The circularity of the highlight portion, that is, the width dimension / short width dimension is calculated. If this circularity is equal to or more than the circularity reference value which is the above reference value data, it is determined that no lead is missing, and the circularity is less than the circularity reference value. If there is, a second pass / fail judgment step for determining a lead missing candidate is executed.
If it is determined in the pass / fail judgment step that the lead missing candidate is present, then the area of the highlight portion within the predetermined range corresponding to the clinch portion is extracted from the image by the angle illumination, and the area ratio to the predetermined range is calculated. If the area ratio is equal to or larger than the second area ratio reference value which is the reference value data, it is determined that there is no lead drop, and if it is smaller than the second area ratio reference value, the lead drop is determined. 10. The lead missing inspection method according to claim 8 or 9, wherein a determination step is executed.