JPH1091785A - Check method and device using pattern matching - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately decide a mask part by setting a mask pattern after deciding the pixels of areas where the variance value of lightness data is larger than the prescribed threshold, excluding the pixels forming the mask pattern, and executing the pattern matching based on a reference pattern and the photographed image of a check object. SOLUTION: The pixels 2e are decided at the areas where the calculated variance value of lightness is larger than the prescribed threshold, and the pixels 2e are masked. Thus, the parts 2e, i.e., the periphery 2c of lustrous parts 2d which vary at random according to the illumination state in regard to their sizes and positions are easily and accurately decided as the pixels of the areas where the variance value of lightness data is larger than the prescribed threshold. Then the parts 2e are masked. A mask pattern 8 consisting of the pixels 2e are set. The pixels 2e of the parts painted in black are masked and excluded out of pixels 7 of the pattern 8 in a pattern matching mode.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for checking whether or not a gap (gap) between metal terminals (pins) of a connector is proper by pattern matching.

[0002]

2. Description of the Related Art When inspecting whether or not an interval (gap) between these metal terminals is appropriate on an image of a large number of images of the same shape such as a metal terminal (pin) of a connector, an image is taken. Inspection is generally performed using a pattern matching technique. In this case, it is necessary to determine a reference pattern in pattern matching. The determination of the reference pattern is performed as follows.

That is, as shown in the top view of FIG. 5A, a pair of metal terminals considered to be the most standard among a large number of metal terminals in an image of the upper surface of the connector body 1 taken by an operator. 2R and 2L are selected by surrounding them with a rectangle, and registered as reference patterns 6R and 6L (see FIG. 5C). Then, the reference patterns 6R, 6L
Is performed on the XY plane to match with the other metal terminals 2. That is, the reference pattern 6R,
6L is moved on the XY plane, and the reference patterns 6R and 6R are moved.
The similarity (degree of coincidence) between L and the pair of metal terminals to be inspected is calculated as a score, and it is determined that the pair of metal terminals 2R and 2L to be inspected exists at a location where the score has a peak. . Thus, the positions of the pair of metal terminals 2R and 2L to be inspected are detected, and the pair of metal terminals 2R and 2L are detected.
The gap d between R and 2L (see FIG. 5B) is measured to check whether the gap d is appropriate.

However, the pair of metal terminals 2R, 2L
Is imaged under epi-illumination as shown in FIG. For this reason, the end surface 2a of the metal terminal 2 shines. On the imaging screen, as shown in FIG. 6 (b), it is observed as a glossy portion 2d having higher brightness than the surroundings.

[0005] The distal end surface 2a of the metal terminal 2 is usually a cut surface, so that its shape is physically unstable. For this reason, as shown in FIG. 7, the position and size of the glossy portion 2d change randomly on the image. Such random changes in the position and area of the glossy portion 2d cause the following inconveniences.

[0006]

Due to the random change in the position and area of the glossy portion 2d, even if the metal terminal 2 to be inspected is a non-defective product, the score of similarity in pattern matching is reduced. Would be. For this reason, in order to prevent a false report that the actual position of the metal terminal is erroneously detected, the score threshold for determining whether or not the shapes match should be set lower. However, in the case of a connector, the background often has a complicated pattern. If the score threshold value is too low, there is a risk that the background pattern is erroneously searched.

In the pattern matching, as described above, a place where the reference pattern most matches the shape of the inspection object to be detected is searched based on the peak of the score. However, since the shape of the glossy portion 2d of the metal terminal 2 to be inspected changes randomly, if the shape of the metal terminal 2 changes randomly, the positioning accuracy deteriorates.

To avoid this, as shown in FIG. 10, a portion 2e of the periphery 2c of the glossy portion 2d whose size and position change randomly, that is, a pixel of a portion having a large change in brightness is masked. However, it is conceivable to perform pattern matching after excluding this.

However, in practice, it is difficult for a human eye to accurately judge a portion 2e of a pixel whose brightness changes randomly. Further, the work of masking countless pixels one by one is extremely complicated.

The present invention has been made in view of such circumstances, and when masking a portion around a glossy portion whose position and size change randomly, the mask portion can be accurately and easily determined. In addition, it is an object of the present invention to prevent an erroneous search from being caused by a decrease in a score at the time of pattern matching, and to improve positioning detection accuracy at the time of pattern matching.

[0011]

Therefore, in the main invention of the present invention, an inspection object is compared with a reference pattern of the inspection object, a degree of matching is calculated, and a value of the calculated degree of matching is calculated. In the inspection method by pattern matching for inspecting the shape of the inspection object according to the imaging process, an imaging step of imaging each of the plurality of inspection objects, and one inspection object among the plurality of imaging objects to be inspected is used as a reference. A reference pattern setting process to be set as a pattern, a calculation process of comparing a captured image of the reference pattern with a captured image of another inspection object, and calculating a variation in brightness data for each pixel corresponding to the same part of the object, A pixel at a portion where the calculated value of the variation in the brightness data is equal to or greater than a predetermined threshold value is determined, and a mask pattern including the determined pixel is determined. A mask pattern setting step of setting a pattern and a step of performing pattern matching by matching a captured image of the reference pattern with a captured image of another inspection object after excluding pixels constituting the mask pattern. I have it.

According to the present invention, a plurality of inspection objects are respectively imaged. Next, one of the plurality of imaged inspection objects is set as a reference pattern.

Next, the picked-up image of the reference pattern and the picked-up image of another inspection object are matched with each other, and the variation of the brightness data is calculated for each pixel corresponding to the same part of the object. Next, a pixel of a portion where the value of the variation of the calculated brightness data is equal to or more than a predetermined threshold is obtained,
A mask pattern including the obtained pixels is set.
In this manner, the peripheral portion of the glossy portion where the position and size change randomly is accurately and easily determined as a pixel of a portion where the value of the variation in the brightness data is equal to or greater than a predetermined threshold, and the portion is masked. You.

Next, after removing the pixels constituting the mask pattern, the captured image of the reference pattern and the captured image of another inspection object are matched to perform pattern matching. As a result, an erroneous search is prevented from being caused by a decrease in the score at the time of pattern matching, and the positioning detection accuracy at the time of pattern matching is enhanced.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an embodiment of an inspection method and an inspection apparatus by pattern matching according to the present invention will be described with reference to the drawings.

FIG. 3 is a perspective view showing an inspection target of this embodiment, and shows a plurality of metal terminals (pins) 2, 2,... In the present embodiment, by performing pattern matching, as shown in FIGS. 5A and 5B described above,
It is assumed that the position of the pair of metal terminals 2R and 2L in the inside is detected, the gap d between the metal terminals 2R and 2L is measured, and whether or not the gap d is appropriate is inspected.

FIG. 4 is a side view showing the configuration of the inspection apparatus according to the present embodiment. A connector main body 1 containing a plurality of metal terminals 2 to be inspected is placed on an inspection table 5.
Above the metal terminal 2, a ring light source 4, which is an imaging illumination for irradiating light toward the tip end surface 2 a of the metal terminal 2, is provided. Above the ring light source 4, a camera 3 that captures images of the tip surfaces 2a of the plurality of metal terminals 2 in the connector main body 1 is provided.

FIG. 1 is a flowchart for explaining the inspection processing according to the present embodiment. Hereinafter, description will be made with reference to this flowchart.

First, as shown in FIG.
Of the plurality of metal terminals 2 in the image (see FIG.
The operator selects one of 1, 2, 23, 24, 25, 26,... By surrounding the metal terminal 24, which is considered to be the most standard, with a rectangle. Here, the distal end surface 2a of each metal terminal is imaged under epi-illumination of the ring light source 4 (see FIG. 6A). For this reason, the end surface 2a of the metal terminal 2 shines. On the imaging screen, the portion is observed as a glossy portion 2d having higher brightness than the surroundings (see FIG. 6B). The distal end surface 2a of the metal terminal 2 is usually a cut surface, so that its shape is physically unstable. For this reason, as shown in FIG. 8, the position and size of the glossy portion 2d vary from one metal terminal to another. As an operator, the position and size of the glossy portion 2d are considered to be the most standard among the various metal terminals.
Set a rectangular area for. Then, the selected rectangular area is registered as a model 6 which is a reference pattern (step 101).

Next, as shown in FIG. 9, the model 6 is divided into metal terminals 21, 22, 23,... (Including its own metal terminal 24).
Is executed on the XY plane (step 102).

In this way, the shape of the model 6 and the shape of each metal terminal are matched, and at this time, for each pixel corresponding to the same location (the same portion of the metal terminal) of the model 6 and each metal terminal 21, 22,. The variation of the brightness data is calculated. Then, a pixel 2e in a portion where the calculated value of the variation in the brightness data is equal to or larger than a predetermined threshold value is obtained, and the pixel 2e is masked (see FIG. 10). As described above, the portion 2e, which is the periphery 2c of the glossy portion 2d and whose position and size randomly change depending on the lighting condition.
Is accurately and easily determined as a pixel of a portion where the value of the variation in the brightness data is equal to or greater than a predetermined threshold, and that portion is masked. Here, the “value of the variation in brightness data” can be obtained as a difference in brightness or as a variance value of brightness. Then, the mask pattern 8 including the obtained pixels 2e is set as shown in FIG.

Of the pixels 7 of the mask pattern 8 shown in FIG. 11, the pixel 2e of the black portion is masked during pattern matching and is excluded (step 103).

However, even with the mask pattern created in this way, the masked portion (the portion painted black) like the mask pattern 8 'in FIG.
May contain important positional information, that is, positional information such as the origin P and the outer peripheral edge portion 2f required for obtaining the gap d (the origin P in FIG. 5C).
R, see PL).

Therefore, as shown in FIG. 14, a mask release pattern 9 for releasing the mask of the area 2'f corresponding to the origin P and the edge portion 2f is created as shown in FIG. Alternatively, a mask release pattern 10 for releasing the mask of the area 2h around the origin P may be created as shown in FIG.

Of the pixels 7 of the mask release patterns 9 and 10 of FIGS. 14 and 15, the pixels 2'f and 2h in the blackened area are excluded from the mask pattern (step 104). .

The mask release pattern 9 thus created
When the pixel 2'f in the area corresponding to 2f is released from the mask pattern 8 'in FIG. 12 using FIG. 12, only the randomly changing portion 2e around the glossy portion 2d is finally formed as shown in FIG. Mask pattern 8 serving as a mask area
Is generated, and it is avoided that important position information (around the origin P) is masked.

The mask release pattern 10 is used to change the mask pattern 8 'of FIG.
When h is released, finally, as shown in FIG. 13, a mask pattern in which a randomly changing portion 2e around the glossy portion 2d and an edge portion 2g excluding the edge portion around the origin P are mask regions. 8 is generated, and it is avoided that important position information (around the origin P) is masked (step 105).

Then, the pixels 2e (or 2e and 2e and 2e) of the mask region constituting the finally obtained mask pattern 8 are formed.
After g) is excluded, pattern matching in which the model 6 and the metal terminals 21, 22,... are matched in the captured image is executed. As a result, an erroneous search is prevented from being caused by a decrease in the score at the time of pattern matching, and the positioning detection accuracy at the time of pattern matching is enhanced.

In the processing shown in FIG.
Can be set to an optimal value.

That is, as shown in FIG.
I indicating the number of times of the repetition processing is initialized to 1, and thereafter, the same processing 1, processing 2, processing 3, processing 4 and processing 5 as steps 101 to 105 in FIG. 1 are sequentially executed (steps 202 to 206). In step 204, the threshold value of the value of the variation in the brightness data is provisionally set to a predetermined set value Th1.

Then, using the mask pattern 8, pattern matching is performed in which the model 6 and each of the metal terminals 21, 22,. That is,
The similarity (degree of coincidence) between the model 6 and each of the metal terminals 21, 22,... Is calculated as a score, and it is determined that the metal terminals 21, 22,. (Step 207). Then, it is determined whether or not the value of the variation (for example, the variance) of the peak value of each score for each of the metal terminals 21, 22,... Has become smaller than a predetermined set value Th2 (step 20).
8).

Here, when it is determined that the value of the variation of the peak value of each score is smaller than the predetermined set value Th2, it is determined that the score distribution is stable and the pattern matching is performed with high accuracy. And the above step 20
The value Th1 provisionally set in 4 is adopted as a threshold value for generating a mask pattern.

However, when it is determined that the variation of the peak value of each score is equal to or greater than the predetermined set value Th2, it is determined that the score distribution is not stable and pattern matching is not performed accurately. Then, after confirming that the current number i of repetitive processing is equal to or smaller than the limit value n (determination YES in step 209), i is incremented by +1 (step 210), and the value currently set in step 204 is set. Th1 is reset so as to be smaller by a predetermined amount (step 211).

Then, a mask pattern is created using the threshold value T'h1 thus reset.
12) Further, a mask release pattern is created (step 213), and pattern matching is executed again using the mask pattern (after the mask release area is excluded by the mask release pattern) (step 207).

Thereafter, until the score distribution as a result of the pattern matching is stabilized (YES in step 208).
A similar process is repeatedly executed while the threshold value Th1 for creating a mask pattern is sequentially set to be smaller (steps 209 to 213 and 207). Then, when it is determined that the value of the variation of the peak value of each score is smaller than the predetermined set value Th2 (step 2).
08), it is determined that the score distribution is stable and the pattern matching is performed with high accuracy, and the value T′h1 that is currently reset is the threshold value for generating the mask pattern. Will eventually be adopted. When the number of repetitive processes i becomes larger than the limit value n (NO in step 209), it is determined that an error has occurred and the entire process is terminated.

In the above-described embodiment, the object to be inspected is the metal terminal of the connector. However, the present invention is not limited to this, and can be applied to any object to be inspected such as an IC lead.

The following cases can be considered as modes for implementing the above-described embodiment.

That is, the processing contents of the above embodiment may be appropriately stored in a floppy disk and distributed and distributed to a user or an operator who performs an inspection by pattern matching. The storage media to be distributed and distributed include hard disks other than floppy disks, IC cards,
A storage medium such as a D-ROM may be used.

In addition to the modes of distributing portable storage media as described above, the modes of distribution include hardware that uses the software of the present embodiment, and development of the software of the present embodiment. A development computer containing the software may be communicably connected via a public line or a network, and may be distributed via the network or the like.

When the software is distributed and distributed, the software is installed in hardware (inspection apparatus using pattern matching) that uses the software.
The software may be distributed and distributed with an installer attached, or data obtained by compressing the capacity of the software may be distributed and distributed.

[Brief description of the drawings]

FIG. 1 is a flowchart showing a processing procedure in an embodiment of an inspection method or an inspection apparatus by pattern matching according to the present invention.

FIG. 2 is a flowchart showing another processing procedure in the embodiment of the inspection method or the inspection apparatus using the pattern matching according to the present invention.

FIG. 3 is a perspective view showing a plurality of metal terminals housed in a connector main body.

FIG. 4 is a side view showing an apparatus configuration of the embodiment.

FIGS. 5A, 5B, and 5C are a top view, a side view, and an enlarged view of a top surface of a metal terminal, respectively.

FIGS. 6A and 6B are a side view and a top view when illumination light is applied to a tip end surface of a metal terminal.

FIG. 7 is a diagram for explaining how the position and size of the glossy portion on the tip end surface of the metal terminal change.

FIG. 8 is a diagram used to explain the processing contents of FIG. 1;

FIG. 9 is a diagram used to explain the processing contents of FIG. 1;

FIG. 10 is a diagram used to explain the processing contents of FIG. 1;

FIG. 11 is a diagram used to explain the processing contents of FIG. 1;

FIG. 12 is a diagram used to explain the processing contents of FIG. 1;

FIG. 13 is a diagram used to explain the processing contents of FIG. 1;

FIG. 14 is a diagram used to explain the processing contents of FIG. 1;

FIG. 15 is a diagram used to explain the processing contents of FIG. 1;

[Explanation of symbols]

 2 Metal terminal (pin) 3 Camera 6 Model (reference pattern)

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Yoshihiko Nakagawaji 1200 Manda, Hiratsuka-shi, Kanagawa Prefecture Inside Komatsu Seisakusho Laboratory (72) Inventor Toru Inomoto 1200 Manda, Hiratsuka-shi, Kanagawa Prefecture Komatsu Seisakusho Research Inside

Claims (6)

[Claims] 1. A pattern matching for matching an inspection object with a reference pattern of the inspection object, calculating a degree of matching thereof, and inspecting the shape of the inspection object according to the calculated value of the degree of matching. An imaging step of imaging each of the plurality of inspection objects; a reference pattern setting step of setting one inspection object of the plurality of imaged inspection objects as a reference pattern; and imaging the reference pattern. A calculation step of comparing the image with a captured image of another inspection object to calculate the variation of the brightness data for each pixel corresponding to the same part of the object; and determining a value of the variation of the calculated brightness data by a predetermined threshold. A mask pattern setting step of determining a pixel of a portion having a value greater than or equal to the value, and setting a mask pattern including the determined pixel; On excluding the pixels constituting down, the reference pattern inspection method according to pattern matching comprising a step of performing a pattern matching against a captured image of the captured image and the other test objects. 2. When the pixels constituting the mask pattern set in the mask pattern setting step include a pixel of a specific portion of the object necessary for pattern matching, the pixel of the specific portion is 2. The inspection method according to claim 1, wherein a mask pattern including the removed pixels is newly set. 3. A pattern matching method for matching an inspection object with a reference pattern of the inspection object, calculating the degree of coincidence, and inspecting the quality of the shape of the inspection object according to the calculated value of the degree of coincidence. An imaging step of imaging each of the plurality of inspection objects; a reference pattern setting step of setting one inspection object of the plurality of imaged inspection objects as a reference pattern; and imaging the reference pattern. A calculation step of comparing the image with a captured image of another inspection object to calculate a variation in the brightness data for each pixel corresponding to the same part of the object;
A pixel of a portion that is equal to or larger than the threshold value is obtained, a mask pattern setting step of setting a mask pattern composed of the obtained pixels, and after excluding pixels constituting the mask pattern, a captured image of the reference pattern A pattern matching execution step of performing pattern matching for calculating the degree of coincidence by comparing the captured images of other inspection objects with each other, and the value of the variation of the degree of coincidence is equal to or greater than a predetermined second threshold value In this case, after setting the first threshold value to a smaller value, the mask pattern setting step and the pattern matching execution step are performed in such a manner that the value of the variation in the degree of coincidence is smaller than the second threshold value. An inspection method by pattern matching that includes a process of sequentially repeating until the size becomes smaller. 4. A pattern matching for comparing an inspection object with a reference pattern of the inspection object, calculating a degree of coincidence thereof, and inspecting the shape of the inspection object according to the calculated value of the degree of coincidence. Imaging means for imaging each of the plurality of inspection objects, reference pattern setting means for setting one inspection object of the plurality of imaged inspection objects as a reference pattern, imaging of the reference pattern Calculating means for comparing the image with the captured image of another inspection object to calculate the variation in the brightness data for each pixel corresponding to the same part of the object; and calculating the variation in the brightness data by a predetermined threshold value. A mask pattern setting means for determining a pixel of a portion having a value equal to or greater than the value, and setting a mask pattern including the determined pixel; On excluding the pixels constituting down, the reference pattern inspection apparatus according to pattern matching and means for performing pattern matching against a captured image of the captured image and the other test objects. 5. When the pixels constituting the mask pattern set by the mask pattern setting means include a pixel of a specific portion of an object necessary for pattern matching, the pixel of the specific portion is replaced with a pixel of the specific portion. 5. The inspection apparatus according to claim 4, wherein a mask pattern including the removed pixels is newly set. 6. A pattern matching for comparing an inspection object with a reference pattern of the inspection object, calculating a degree of matching thereof, and inspecting the shape of the inspection object according to the calculated value of the degree of matching. Imaging means for imaging each of the plurality of inspection objects, reference pattern setting means for setting one inspection object of the plurality of imaged inspection objects as a reference pattern, imaging of the reference pattern Computing means for comparing the image with a captured image of another inspection object to calculate a variation in the brightness data for each pixel corresponding to the same part of the object;
A pixel of a portion that is equal to or larger than the threshold value is obtained, a mask pattern setting means for setting a mask pattern composed of the obtained pixels, and after excluding pixels constituting the mask pattern, a captured image of the reference pattern Pattern matching executing means for performing pattern matching for calculating the degree of coincidence by comparing the captured images of the other inspection objects with each other, and the value of the variation in the degree of coincidence is equal to or greater than a predetermined second threshold value In this case, after setting the first threshold value to a smaller value, the mask pattern setting unit and the pattern matching execution unit are set to determine that the value of the variation in the degree of coincidence is smaller than the second threshold value. Inspection apparatus by pattern matching, comprising means for sequentially repeating until the size becomes smaller.