JP2003196641A - Article recognition device and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To accurately recognize the existence, the position, the angle, the direction or the inclination of an article even if a print of a character string or the illuminance of an illumination has dispersion, by recognizing the character string printed on the article. <P>SOLUTION: This device has an imaging device 11 for photographing an article image, a storage part for storing a model image, and a recognition processing part 14 for performing pattern matching by comparing the image of the character string included in the article image with the model image, and detecting the existence, the position, the angle, the direction or the inclination of the article. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to an article recognition apparatus and method.

[0002]

2. Description of the Related Art Conventionally, in an electronic device manufacturing line, it is inspected whether a chip-shaped electronic component such as a capacitor mounted on a substrate such as a printed wiring board is soldered in a predetermined direction. In order to do so, an article recognition device utilizing image processing is adopted, and the direction of the electronic component is determined based on the polarity mark formed on the electronic component.

In this case, an electronic component such as a capacitor has positive and negative terminals and is mounted by soldering the terminals to mounting terminals (lands) on the board. . Therefore, if the direction in which the electronic component is facing, that is, if the direction is wrong, the positive and negative polarities will be reversed, and the electronic device composed of the board on which the electronic component is mounted will not operate correctly. I will end up.

In general, a chip-shaped electronic component is extremely small, and it is difficult to determine the polarity of the terminal from the outer shape.
A polarity mark for indicating the polarity of the terminal is printed on the end of the surface of the electronic component near the positive or negative terminal.
Therefore, the conventional article recognition device is configured to recognize the polarity mark printed on the surface by image processing and determine the orientation of the electronic component. The board on which the electronic component having the wrong orientation is mounted is detected as a defective product by the article recognition device and is removed by an operator or the like, so that it cannot be sent to the subsequent process of the manufacturing line.

[0005]

However, in the conventional article recognition device, since the polarity mark is recognized to determine the orientation of the electronic component, the orientation of the electronic component is erroneously determined, and the electronic component is normal. In many cases, the board mounted on the board was erroneously detected as a defective product.

Generally, since the chip-shaped electronic component is extremely small, the polarity mark printed on the end of the surface of the electronic component near the positive or negative terminal becomes extremely small. Moreover, the polar mark has an extremely simple shape. Therefore, if the print is slightly unclear or the illuminance of the illumination for illuminating the electronic component varies even slightly, the polarity mark cannot be correctly recognized.

As described above, when the orientation of the electronic component is erroneously determined, the substrate on which the electronic component having the incorrect orientation is mounted is sent to the subsequent process of the manufacturing line, or all the electronic components are normally mounted. In addition, the defective board may be erroneously detected as a defective product, resulting in an increase in the defective product rate in the manufacturing line or a decrease in throughput.

The present invention solves the problems of the conventional article recognition device and recognizes a character string printed on an article, so that even if the character string is printed or the illuminance of illumination varies, It is an object of the present invention to provide an article recognition device and method capable of correctly recognizing the presence / absence, position, angle, direction, or inclination of the article.

[0009]

Therefore, in the article recognition apparatus of the present invention, an image pickup apparatus for taking an image of an article, a storage section for storing a model image, and a character string included in the image of the article are provided. Performing pattern matching by comparing the image with the model image, presence or absence of the article, position, angle,
And a recognition processing unit that detects an orientation or an inclination.

In another article recognition apparatus of the present invention, the recognition processing unit further compares the image of the character string with one model image and performs pattern matching,
Pattern matching is performed by comparing the image of the character string with a plurality of model images.

In still another article recognition apparatus of the present invention, the recognition processing section changes the illumination illuminance and performs the pattern matching again.

In still another article recognition device of the present invention, the character string comprises a plurality of characters.

In still another article recognition device of the present invention, the article is an electronic component mounted on a substrate, and the character string is a character string printed on the electronic component.

In the article recognition method of the present invention, an image of an article is photographed, an image of a character string included in the image of the article is compared with a model image to perform pattern matching, and the presence / absence, position, and angle of the article are detected. , Direction, or inclination is detected.

In another article recognition method of the present invention, the image of the character string is further compared with one model image to perform pattern matching, and then the image of the character string is compared with a plurality of model images. Pattern matching is performed.

In still another article recognition method of the present invention, the illumination illuminance is changed and the pattern matching is performed again.

In still another article recognition method of the present invention, the character string further comprises a plurality of characters.

In still another article recognition method of the present invention, the article is an electronic component mounted on a substrate, and the character string is a character string printed on the electronic component.

[0019]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. The article recognition apparatus and method of the present invention recognizes articles of any shape and size, such as cardboard boxes and wooden boxes for housing electrical products and automobile parts, industrial products such as televisions and radios. However, for convenience of explanation, here, a case where the article is a chip-shaped electronic component such as a capacitor mounted on a substrate such as a printed wiring board will be described.

FIG. 2 is a perspective view showing an electronic component according to the embodiment of the present invention.

In the figure, reference numeral 21 indicates, for example, a vehicle navigation device, an automobile engine control device, a radio,
A substrate such as a printed wiring board incorporated in an electronic device such as a mobile terminal device or a computer, and 22 are electronic components fixed and mounted on the substrate 21 by means such as solder or a conductive adhesive. Here, the electronic component 22 is, for example,
Although it is an active element such as an IC or LSI, a passive element such as a resistor, a coil, or a capacitor, it may be of any type. In the present embodiment, for convenience of description, the circuit formed on the substrate 21 and a large number of mounted electronic components are omitted, and the electronic components 22 include the aluminum electrolytic capacitor 22a and the aluminum electrolytic capacitor 22a. An aluminum electrolytic capacitor 22b, which is smaller than the electrolytic capacitor 22a,
Also, a tantalum capacitor 22c is shown.

Then, the aluminum electrolytic capacitor 22
The a, 22b and the tantalum capacitor 22c are, for example, positive electrode terminals 23a, 23b and 23c, respectively.
And the electrode terminals 23a, 23b and 23c are provided on the substrate 2
1 is electrically connected to a land (not shown), which is a mounting terminal in a circuit formed on the surface of 1, by solder, a conductive adhesive, or the like. The aluminum electrolytic capacitors 22a and 22b and the tantalum capacitor 2 are used.
2c is the electrode terminals 23a, 23b and 2 respectively.
On the side opposite to 3c, a negative electrode terminal (not shown) having an opposite polarity is provided.

Further, the aluminum electrolytic capacitor 22a,
Character strings 24a, 24b, and 24c indicating a product name, a part number, a type, a standard, a performance, and the like are printed on the upper end surface of the 22b and the tantalum capacitor 22c. In the present embodiment, the character string is made up of a plurality of characters, and the characters are not only characters such as hiragana, katakana, alphabets, and kanji, but also Roman numerals, Arabic numerals, and Chinese numerals. Numbers such as +,-, x, √
Academic symbols such as, unit symbols such as%, mm, cm ,? ,! ,
It also includes various symbols such as descriptive symbols such as ~ and general symbols such as @ and 〒.

Further, for example, polarity marks 25a, 25b and 25c are printed on the upper end surface to indicate the side on which an electrode terminal (not shown) having negative polarity is located.
As a result, the aluminum electrolytic capacitors 22a, 22
It is possible to determine the direction in which b and the tantalum capacitor 22c are attached.

Next, the configuration of the article recognition device according to the embodiment of the present invention will be described.

FIG. 1 is a conceptual diagram showing the structure of an article recognition device according to an embodiment of the present invention.

In the figure, 23 is a mounting table on which the substrate 21 on which the electronic component 22 is mounted is mounted, and 11 is an image pickup device disposed above the mounting table 23. here,
The image pickup device 11 is, for example, an industrial television camera, but has a CCD (Charge Coupled Device).
e), any device may be used as long as it has an image pickup means such as an image pickup tube and can take an image of the electronic component 22 or the like. Further, the image pickup device 11 can be moved relative to the mounting table 23, and sequentially captures each part of the substrate 21 mounted on the mounting table 23 to detect each electronic component 22. Images of the upper end surface can be sequentially taken. In this case, the imaging device 11 may be moved, or the mounting table 23 may be moved.

Reference numeral 10 is a control device as an article recognition device for recognizing an article based on an image taken by the image pickup device 11, and is a storage unit such as a CPU, an MPU or the like, a semiconductor memory, a magnetic disk or the like. Means, keyboard, input means such as mouse, CRT, display means such as liquid crystal display, communication interface and the like. Then, the control device 10 stores an inspection image storage unit 12 that stores an image captured by the imaging device 11, a model image storage unit 13 that stores a model image that serves as a reference for recognizing an article, A recognition processing unit 14 that executes a process of recognizing an article, and a lighting device 16 described later.
The lighting control unit 15 for controlling the. Here, the recognition processing unit 14 compares the image captured by the imaging device 11 stored in the inspection image storage unit 12 with the model image stored in the model image storage unit 13,
The degree of similarity is calculated by the normalized correlation calculation, that is, pattern matching is performed by the normalized correlation. The pattern matching by the normalized correlation is a well-known technique (see JP-A-63-98070 and JP-A-5-189570).

The lighting device 16 is disposed above the mounting table 23. Here, the lighting device 16 is, for example,
It may be an incandescent lamp, a fluorescent lamp, a halogen lamp, or the like, but an LED (Light that has excellent responsiveness in illuminance control)
It is desirable that the ring-shaped illumination units 16-1 and 16-2 are formed of a ring-shaped (ring-shaped) light emitting surface and are formed of a t emitting diode. In this case, the outer diameter of the upper ring-shaped lighting unit 16-1 is smaller than the inner diameter of the lower ring-shaped lighting unit 16-2. In addition, the ring-shaped lighting unit 16-
The color of the illumination light of the ring-shaped illumination unit 16-2 of the first and the lower stage is the same. As a result, the upper end surface of the electronic component 22 is illuminated by the illumination light incident at different angles, so that the minute inclination of the upper end surface, the minute scratches, the cuts and the like existing on the upper end surface are not recognized. . In addition,
The ring-shaped illumination unit may have three or more stages. Further, one ring-shaped lighting unit may be a pseudo coaxial lighting unit. Moreover, it is desirable that the relative positional relationship between the illumination device 16 and the imaging device 11 does not change.

Next, the operation of the article recognition apparatus having the above-mentioned structure will be described.

FIG. 3 is a diagram showing an image photographed by the image pickup device according to the embodiment of the present invention, FIG. 4 is a diagram showing a method for removing noise of luminance of the image according to the embodiment of the present invention, and FIG. FIG. 1 is a first diagram showing an example of a model image in the embodiment of the present invention, FIG. 6 is a second diagram showing an example of the model image in the embodiment of the present invention, and FIG. 7 is an article in the embodiment of the present invention. It is a flowchart which shows a recognition process.

Here, the article recognition device according to the present embodiment discriminates the orientation of the electronic component 22 mounted on the substrate 21 in the manufacturing line of the electronic device or the like, and the electronic component 22 having the wrong orientation is mounted. A case will be described in which it is used in the inspection process for detecting the defective substrate 21 as a defective product.

First, through a process such as solder reflow, a large number of electronic components 22 are fixed, and the mounted substrate 21 is mounted on the mounting table 23. Then, the imaging device 11 and the illumination device 16 are moved onto one of the multiple electronic components 22 existing on the substrate 21 by a moving means such as a robot hand or an XY moving device (not shown). . Here, a case will be described in which the mounting table 23 does not move, and the imaging device 11 to which the illumination device 16 is integrally attached is moved by the moving unit with respect to the mounting table 23.

Then, the image pickup device 11 takes an image of the upper end surface of the electronic component 22 from above. As a result, an image as shown in FIG. 3 is captured and stored in the inspection image storage unit 12 as an inspection image to be subjected to pattern matching. Here, the image shown in FIG. 3 is an example when the electronic component 22 is an aluminum electrolytic capacitor,
The character string 24 and the polarity mark 25 are included. In this case, in order to reduce the variation in the brightness of the inspection image, a spatial filter is first used to perform the smoothing process on the inspection image. As a result, the image becomes blurred and minute noise components are removed from the image.

Subsequently, the recognition processing section 14 carries out a background cutting process for cutting luminance components above and below a predetermined range. As a result, it is possible to remove the noise of the luminance outside the range and further reduce the variation in the luminance of the inspection image. In the present embodiment, pattern matching based on normalized correlation is performed.
And in the case of pattern matching by normalized correlation,
When the variation in luminance in the inspection image is small, that is, when the luminance noise is small, a stable correlation value can be obtained, and a stable pattern matching result can be obtained. Therefore, as shown in FIG. 4, a predetermined range, for example, a luminance component having a luminance of 200 or more and 80 or less is cut.

Here, the brightness distribution of the inspection image before the background cutting process is, for example, as shown in FIG. 4 (a). Note that, in FIG. 4A, the horizontal axis represents the brightness of the inspection image captured by the imaging device 11, but the numerical values of the brightness are relative and the maximum value is 255 and the minimum value is 255. The value is set to 0 and is set in 256 steps. In this case, the CCD in the image pickup device 11,
It is usual to set the maximum brightness obtained by the sensitivity of the image pickup means such as the image pickup tube and the amplification degree of the amplifier to 255. However, the predetermined brightness can be set as 255. For example, the brightness of 80% of the maximum brightness may be set as the maximum value 255, and all the brightness higher than the maximum brightness may be represented as 255. The same applies to the minimum value, and it is usual to set 0 as the lowest luminance obtained by the sensitivity of the image pickup means, the amplification degree of the amplifier, etc., but it can be set appropriately. Further, in FIG. 4A, the vertical axis indicates the degree of frequent occurrence of pixels of each luminance in the inspection image captured by the imaging device 11, that is, the frequency. The frequency is
It may be the ratio of pixels of each luminance to the total number of pixels, or the number of pixels of each luminance.

Then, FIG. 4A shows an example of the luminance distribution in the inspection image as shown in FIG.
Here, the distribution range 27a indicates the luminance distribution of the pixels corresponding to the main body of the electronic component 22 as the background, and the distribution range 2a
7b shows the luminance distribution of the pixels corresponding to the character string 24 and the polarity mark 25. When the electronic component 22 is an aluminum electrolytic capacitor or the like, the background main body portion is an aluminum alloy base metal, and therefore the reflectance of the illumination light from the illumination device 16 is high, so that the brightness is high. Since the character string 24 and the polar mark 25 are printed on the aluminum alloy ingot, the reflectance of the illumination light from the lighting device 16 is lower than that of the aluminum alloy ingot.
The brightness becomes low.

In this case, the upper limit line 28a corresponding to a luminance of 200 and the lower limit line 28b corresponding to a luminance of 80
Is set to cut the range above the upper limit line 28a and the range below the lower limit line 28b, and the upper limit line 2
The range between 8a and the lower limit line 28b is extracted. Then, by adjusting the scale of the horizontal axis, the upper limit line 28a
Is set to 255, and the lower limit line 28
By setting the brightness corresponding to b to 0, it is possible to obtain an image subjected to background cut processing. Here, the luminance distribution of the inspection image subjected to the background cut processing is shown in FIG.
As shown in (b).

The corresponding luminance values of the upper limit line 28a and the lower limit line 28b can be set appropriately. For example, in the example shown in FIG. 4A, when the distribution range 27b is closer to the vertical axis, that is, the brightness distribution of the pixels corresponding to the character string 24 and the polarity mark 25 is in the lower brightness range. In this case, it is desirable that the lower limit line 28b corresponds to a lower brightness value. In this case, it is desirable to set the corresponding luminance value of the lower limit line 28b so that the lightest printed portion of the character string 24 can be detected.

Subsequently, the recognition processing section 14 refers to one model image and performs a first pattern matching process in order to find out at which position in the inspection image the model image is located. Here, a model image serving as a reference for recognizing an article is created in advance for each article and stored in the model image storage unit 13. It is desirable that a plurality of the model images be further created for each article. For example, for the electronic component 22 having the image of the upper end surface as shown in FIG. 3, four model images having different thicknesses of characters forming the character string are created in advance as shown in FIG. , Which are stored in the model image storage unit 13. The model image shown in FIG. 5 includes a contour line and a polarity mark on the upper end surface of the electronic component 22 in addition to the character string, but the contour line and the polarity mark on the upper end surface of the electronic component 22 are omitted. can do. further,
Not all of the character strings but only a part of the character strings may be included in the model image. In a pattern matching process described later, only a necessary character string portion may be extracted from various images included in the model image and used as a model image for pattern matching.

If the electronic component 22 is a tantalum capacitor, as shown in FIG. 6, two models in which the characters forming the character string have different thicknesses are created in advance, and the model image storage unit 13 is created. It is stored in. As with the image shown in FIG. 5, the model image shown in FIG. 6 also includes the outline of the upper end surface of the electronic component 22 and the polarity mark, in addition to the character string.
The contour line and the polarity mark on the upper end surface of 2 can be omitted. Further, not all of the character strings but only a part of the character strings may be included in the model image.

In the first pattern matching process, one model image is selected from a plurality of model images created for each article, and the
A pattern matching process is performed with reference to one model image to find out where the model image is located in the inspection image. Here, the case where the model image shown in FIG. 5A is selected will be described. in this case,
A case where a predetermined range 26 in the character string is extracted and used as a model image for pattern matching will be described. The range 26 may be set to include all character strings, or may be set to include only a character string composed of a small number of characters, for example, “16v”. It should be noted that it is desirable to avoid using a character string consisting of characters of extremely simple shapes as the model image for the pattern matching, such as the Arabic numeral 1 and the Chinese numeral 1. However, even if the shape of each character is simple, if the shape of the entire character string is not so simple, it is suitable to be selected as the model image for the pattern matching.

Subsequently, the recognition processing unit 14 refers to the character string included in the range 26 of the model image shown in FIG. 5A as a model image for pattern matching,
The inspection image as shown in FIG. In this case, the recognition processing unit 14 performs pattern matching by the normalized correlation, calculates the correlation value as the similarity, and finds out in which position of the inspection image the character string included in the range 26 exists. The position coordinates and the rotation angle of the character string included in the range 26 in the inspection image are calculated. The first pattern matching process is performed in order to preliminarily determine the position coordinates and the rotation angle of the character string included in the range 26 in the inspection image.
The threshold value of the correlation value in the pattern matching process is set to a low value. This reduces the processing load on the recognition processing unit 14 and shortens the processing time of the first pattern matching processing.

Subsequently, the recognition processing section 14 carries out the second pattern matching processing in order to confirm the position coordinates and the rotation angle in the inspection image of the character string included in the range 26 calculated by the first pattern matching processing.
In this case, the pattern matching is performed with reference to all the model images created for each article and stored in the model image storage unit 13. Therefore,
The recognition processing unit 14 sequentially refers to the character string included in the range 26 of the model image shown in FIGS. 5A to 5D as a model image for pattern matching, and the inspection image as shown in FIG. Compare with.

In the case of electronic parts, the character string for displaying the product name, part number, type, standard, performance, etc. printed on the front surface is slightly different in color and line width depending on the printing state. However, they are often different. for that reason,
As shown in FIGS. 5A to 5D, the model image storage unit 13 stores four model images in which the line widths of the character strings are slightly different.

Then, the recognition processing section 14 is shown in FIG.
The character string included in the range 26 of the model image shown in (d) is referred to as a model image for pattern matching, and sequentially compared with the inspection image as shown in FIG. In this case, the recognition processing unit 14 performs pattern matching by the normalized correlation to calculate the correlation value as the similarity, and selects the model image having the highest correlation value. Since the position coordinates and the rotation angle of the character string included in the range 26 in the inspection image are preliminarily determined by the first pattern matching process, even if the pattern matching by the normalized correlation is performed, the recognition processing unit The processing load of 14 does not increase so much, and the processing time does not increase. Here, it is assumed that the highest correlation value can be obtained when the character string included in the range 26 of the model image shown in FIG. 5A is used as the model image for pattern matching.

Subsequently, the recognition processing unit 14 determines whether or not the correlation value of the model image selected by the second pattern matching processing, that is, the highest correlation value is equal to or more than a preset threshold value. Perform value determination processing. In this case, the threshold value is set to a value higher than the correlation value threshold value set in the first pattern matching process. Then, when the correlation value is equal to or more than the threshold value, the recognition processing unit 14 determines that the electronic component 22 is correctly recognized, and based on the position coordinates and the rotation angle of the character string included in the range 26 in the inspection image. The electronic component 2
2 is confirmed, the position, angle, direction, inclination, etc. of the electronic component 22 are calculated and output. Thereby, the control device 10 can determine the orientation of the electronic component 22 and determine whether the electronic component 22 is correctly mounted on the substrate 21.

On the other hand, when the correlation value is less than the threshold value,
The recognition processing unit 14 performs an illumination illuminance changing process. here,
The line width of the character string in the inspection image captured by the imaging device 11 is the illuminance of the illumination light from the illumination device 16, that is,
The line width of a character string printed in a dark color on a light background changes depending on the illumination intensity, and becomes narrower as the illumination intensity increases. This is because the illumination light reflected from the background surrounding the line of the character string and incident on the image pickup device 11 becomes strong and erodes the line width of the character string. Therefore, even if the line width of the character string in the inspection image photographed at a predetermined illumination illuminance is thick as shown in FIG. 5D, if the illumination illuminance is increased, for example, FIG. As shown in, it becomes thinner.

Further, the line width of a character string printed in a bright color on a dark background becomes thicker as the illumination illuminance becomes stronger.
In this case, even if the line width of the character string in the inspection image taken at a predetermined illumination illuminance is thin as shown in FIG. 6A, if the illumination illuminance is increased, for example, as shown in FIG. ), It becomes thicker.

Therefore, the recognition processing section 14 includes the illumination control section 1
5, the lighting device 16 is controlled to change the illumination illuminance by a preset change amount, and again,
The image pickup device 11 is caused to take an image of the upper end surface of the electronic component 22. Then, the above-described inspection image smoothing process, background cutting process, first pattern matching process, second pattern matching process, and correlation value determination process are performed.

Then, in the correlation value judgment processing, the second
If it is determined that the highest correlation value calculated by the pattern matching process is equal to or greater than the preset threshold value, as described above, the position of the electronic component 22,
The angle, direction, inclination, etc. are calculated and output. On the other hand, when the correlation value is less than the threshold value, the illumination illuminance changing process may be performed again, or the electronic component 22.
It is also possible to end the processing of (1) and confirm that the electronic component 22 cannot be detected, and output the electronic component 22. Although it is possible to set the illumination illuminance changing process to be performed many times, it is desirable to set it to be performed once or twice because the throughput of the inspection process of the substrate 21 is reduced.

In this way, the control device 10 determines the orientation of the electronic component 22, and the electronic component 22 determines that the substrate 2
When it is determined whether or not it is correctly mounted on the board 1, the imaging device 11 is moved onto the other electronic component 22 on the board 21 and the inspection image is smoothed with respect to the other electronic component 22. Conversion processing, background cutting processing, first pattern matching processing, second pattern matching processing, and correlation value determination processing. Then, the orientations of all electronic components 22 on the substrate 21 are determined, and the electronic components 22 are
It is determined whether or not it has been correctly mounted on 1.

Next, the flow chart will be described.
Step S1 In order to reduce the variation in the brightness of the inspection image, a smoothing process is performed using a spatial filter to remove minute noise components. Step S2 In order to reduce the variation in the brightness of the inspection image, the brightness component above or below a predetermined range is cut, and the noise of the brightness outside the range is removed. Step S3 Referring to one model, the position of the model in the inspection image is searched, and the position coordinates and the rotation angle are calculated. Step S4
From the calculated position coordinates and rotation angles, the model having the highest correlation value among the plurality of models is selected. Step S5
Correlation value determination processing is performed to determine whether the correlation value of the selected model is equal to or greater than the threshold value. Step S6 If the correlation value is less than the threshold value, the illumination illuminance changing process is performed. Step S7: If the correlation value is greater than or equal to the threshold value, the electronic component 2
2 is correctly recognized, and if the correlation value is less than the threshold value even after performing the illumination illuminance changing process again, the electronic component 2
The process ends as 2 is not recognized.

By the way, in order to recognize the shape of the character string so that the inspection image does not include noise, it is desirable that the difference between the brightness of the background and the brightness of the character string that is the printed portion is large. Therefore, in the present embodiment, the illumination illuminance is set in advance so that the difference between the background brightness and the character string, that is, the brightness of the printed portion is maximized.

FIG. 8 is a diagram showing a method of setting the illumination illuminance in the embodiment of the present invention, and FIG. 9 is a diagram showing an actual measurement value of the luminance change of the image with respect to the change of the illumination illuminance in the embodiment of the present invention. .

Here, when the illumination illuminance is changed, the brightness of the background and the brightness of the printed portion change as shown in FIG. In FIG. 8, the horizontal axis represents the illuminance of the illumination light from the illumination device 16, that is, the illumination illuminance, and the vertical axis represents the brightness of the inspection image captured by the imaging device 11. The numerical value of the luminance is relative, like the luminance in FIG. 4A, and the maximum value is 255 and the minimum value is 0.
Is set in 256 steps. In this case, it is usual to set the maximum luminance obtained by the sensitivity of the image pickup means such as the CCD and the image pickup tube of the image pickup device 11 and the amplification degree of the amplifier as 255, but set the predetermined luminance as 255. You can also For example, the brightness of 80% of the maximum brightness may be set as the maximum value 255, and all the brightness higher than the maximum brightness may be represented as 255. The same applies to the minimum value, and it is usual to set 0 as the lowest luminance obtained by the sensitivity of the image pickup means, the amplification degree of the amplifier, etc., but it can be set appropriately.

When the background is a bright color and the reflectance is high and the printed portion is a dark color and the reflectance is low as in the inspection image as shown in FIG. 3, when the illumination illuminance is changed, the luminance of the background is changed. And the brightness of the printed portion changes as shown by lines 31 and 32 in FIG. Here, the line 31 shows the change in the brightness of the background, and the line 32 shows the change in the brightness of the printed portion. Also, the line 33 has a maximum value of 25
5 is shown. And when the illumination illuminance is changed,
The difference between the background brightness and the printed part brightness is indicated by the length of the line segment 34.

In FIG. 8, it is understood that the line segment 34 has the longest length when the upper end of the line segment 34 is located at the intersection of the line 31 and the line 33. The value of the illumination illuminance corresponding to the intersection is indicated by arrow 35. From this, when the illumination illuminance is set so that the background brightness becomes maximum,
That is, it was found that when the value of the illumination illuminance was set to the value indicated by the arrow 35, the difference between the brightness of the background and the brightness of the printed portion was the maximum. Note that, as shown in FIG. 3, when the brightness of the printed portion is higher than the brightness of the background, the illumination illuminance is set so that the brightness of the printed portion is maximized.

Here, if the difference between the brightness of the background and the brightness of the printed portion is large, noise included in the image can be reliably removed. That is, a normal noise filter sets a threshold value sufficiently larger than the value of noise included in an image so as to remove all signals below the threshold value,
Noise can be removed. Even in this case, there is a large difference between the brightness of the background and the brightness of the printed portion, so that the background and the printed portion can be clearly identified.

On the other hand, when the difference between the brightness of the background and the brightness of the printed portion is small, the threshold value cannot be increased and noise cannot be reliably removed. Therefore, the background and the printed portion cannot be clearly identified. Further, if the threshold value is increased, noise is also removed, but since a signal indicating the background or the printed portion is also removed, the background and the printed portion cannot be clearly discriminated.

Therefore, the illumination control section 15 controls the illumination device 16 so that the higher luminance of the background or the printed portion in the inspection image photographed by the image pickup device 11 becomes the maximum, so that the normal illumination is performed. Even if the noise filter is used, it is possible to set the optimum illumination illuminance that can reliably remove the noise included in the image.

As shown in FIG. 6, when the inspection image has a dark background and a low reflectance and the printed portion has a bright color and a high reflectance, the luminance of the background in FIG. It shows a change like the line 32, and the brightness of the printed part is line 3
It shows a change like 1. Therefore, the illumination controller 15 controls the illumination device 16 so that the brightness of the printed portion in the inspection image captured by the imaging device 11 is maximized, so that the optimal illumination illuminance can be set.

FIG. 9 shows "illumination illuminance-CCD sensitivity characteristic" which represents the actual measurement value of the luminance change with respect to the illumination illuminance change. In FIG. 9, the horizontal axis represents the illuminance of the illumination light from the illumination device 16, that is, the illumination illuminance, and the vertical axis represents the CCD sensitivity as the brightness of the inspection image captured by the imaging device 11. .

Here, the numerical value of the illumination illuminance is a relative value, and the maximum value is 255 and the minimum value is 0.
The illumination illuminance when the output of the lighting device 16 is set to the maximum controllable output is set to 255 and the output of the lighting device 16 is set to the controllable minimum output. Is set as 0. Further, the CCD sensitivity is an output value of the CCD which is the image pickup means of the image pickup device 11, and is a value proportional to the luminance in the inspection image. The CCD sensitivity value is also a relative value, and the maximum value obtained is set to 255,
The lowest value is set as 0.

The line 36 in FIG. 9 indicates the lighting device 1.
6 shows CCD sensitivity characteristics of reflected light when the illumination light from 6 is incident on the base metal of the tantalum alloy on the upper end surface of the tantalum capacitor as the electronic component 22. Also, line 37
Shows the CCD sensitivity characteristic of the reflected light when the illumination light from the illuminating device 16 is incident on the aluminum alloy base metal on the upper end surface of the aluminum electrolytic capacitor as the electronic component 22. Further, the line 38 shows the CCD sensitivity characteristic of the reflected light when the illumination light from the illumination device 16 is made incident on the characters printed on the upper end surfaces of the aluminum electrolytic capacitor and the tantalum capacitor as the electronic component 22.

In each case, even if the illumination illuminance is 0, the CCD sensitivity is not 0. However, even if the output of the illumination device 16 is the minimum output that can be controlled, some illumination light will not be emitted. It is considered that this is because it is irradiated. Further, it can be seen that the reflected light from the base metal of the tantalum alloy of the tantalum capacitor does not reach the maximum value of CCD sensitivity even if the output of the illumination device 16 is the maximum output that can be controlled. Further, the CCD sensitivity characteristics of the reflected light from the aluminum alloy base metal on the upper end surface of the aluminum electrolytic capacitor and the reflected light from the printed characters are not clear near the upper limit.

However, since the values of the illumination illuminance and the values of the CCD sensitivity are relative and the scale can be adjusted, the values shown in FIG. 8 can be obtained by appropriately adjusting the scales on the vertical axis and the horizontal axis. It is possible to obtain a linear relationship as shown below.

The relationship between the illumination illuminance and the luminance is not limited to the hue and the surface roughness of the part that reflects the illumination light from the illumination device 16, but also the distance from the illumination device 16 and the image pickup device 11, that is, the illumination. It also changes depending on the optical path length of light. The height relationship between the mounting table 23 and the lighting device 16 and the image pickup device 11 is normally set so as not to change. Even if it has an end face, the relationship between illumination illuminance and brightness changes.

Therefore, for each electronic component 22, the value of the illumination illuminance when the difference between the brightness of the background and the brightness of the printed portion is maximized, that is, the value indicated by arrow 35 is set in advance, and the control device is set. It is desirable to store it in 10 storage means. In this case, since it is only necessary to set the value of the proof illuminance in the case where the higher brightness of the background or the printed part has the maximum brightness, the respective electronic components 22
With respect to, it is possible to easily set the value of the illumination illuminance. Then, when determining the orientation of the electronic component 22 mounted on the board 21, the imaging device 11 is moved onto the electronic component 22 on the board 21, and an image of the upper end surface of the electronic component 22 is captured. At some times, the illumination control unit 15 controls the illumination device 16 so that the illumination illuminance becomes the value set for the electronic component 22.

When the orientation of the electronic component 22 mounted on the board 21 is determined without presetting the illumination illuminance value when the difference between the background brightness and the printed part brightness is maximum. Can also be set. In this case, the imaging device 11 is moved onto the electronic component 22 on the substrate 21, and when the image of the upper end surface of each electronic component 22 is captured, the illumination illuminance is changed and the change in the background brightness is measured. To do. Then, since it is only necessary to set the value of the illumination illuminance in the case where the higher brightness of the background or the printed part has the maximum brightness for each electronic component 22, it is possible to easily and easily The value of the illumination illuminance can be set in a short time.

As described above, in the present embodiment, the recognition processing unit 14 of the article recognition device compares the image of the character string included in the image of the upper end surface of the electronic component 22 with the model image to perform pattern matching. The presence or absence of the electronic component 22, the position, the angle, the orientation, the inclination, etc. are calculated and output, that is, detected. In this case, since pattern matching is performed by using an image of a character string with a complicated shape, even if the character string is printed or the illumination illuminance varies,
The presence / absence, position, angle, direction, inclination, etc. of the electronic component 22 can be detected. Therefore, the control device 10 can determine the orientation of the electronic component 22 and reliably determine whether the electronic component 22 is correctly mounted on the substrate 21.

Further, since the recognition processing unit 14 performs the background cutting processing for cutting the components of the luminance above and below the predetermined range, the noise of the luminance outside the above range is removed, and the variation in the luminance of the inspection image is effective. Can be reduced.

Further, the recognition processing unit 14 refers to one model image, performs the first pattern matching process for finding out the position of the model image in the inspection image, and then performs the first pattern matching process. Based on the result of the pattern matching process, a more precise second pattern matching process is performed with reference to the plurality of model images. Therefore, the recognition processing unit 1
The processing load of 4 is reduced, and the processing time is also shortened.

Further, the recognition processing section 14 performs the illumination illuminance changing process for changing the illumination illuminance and performing the pattern matching again. Therefore, even if the illumination illuminance varies, pattern matching can be appropriately performed.

Although the case where the article is the chip-shaped electronic component 22 such as the capacitor mounted on the substrate 21 such as the printed wiring board has been described in the present embodiment, the article recognition device and the article recognition apparatus of the present invention are described. The method can recognize any shape and size of an article such as a cardboard box or a wooden box for housing an electric product or an automobile part, an industrial product such as a television or a radio, and the presence or absence of the article. , Position, angle, direction, inclination, etc. can be accurately detected.

The present invention is not limited to the above-mentioned embodiments, but can be variously modified within the scope of the present invention, and they are not excluded from the scope of the present invention.

[0077]

As described in detail above, in the present invention, the article recognition apparatus includes an image pickup apparatus for taking an image of an article, a storage section for storing a model image, and a character string included in the image of the article. The image of is compared with the model image to perform pattern matching, the presence or absence of the article, the position, the angle,
And a recognition processing unit that detects an orientation or an inclination.

In the article recognition method, an image of an article is photographed, an image of a character string included in the image of the article is compared with a model image to perform pattern matching, and the presence / absence, position, angle, direction of the article are detected. Alternatively, the tilt is detected.

In this case, since the pattern matching is performed by the image of the character string having a complicated shape, the presence or absence of the article, the position, the angle, the direction, the inclination, etc. are detected even if the character string is printed or the illumination illuminance varies. be able to.

When the article is an electronic component mounted on a substrate and the character string is a character string printed on the electronic component, the orientation of the electronic component is determined and the electronic component is It is possible to surely determine whether or not the board is correctly mounted on the board.

[Brief description of drawings]

FIG. 1 is a conceptual diagram showing a configuration of an article recognition device according to an embodiment of the present invention.

FIG. 2 is a perspective view showing an electronic component according to an embodiment of the present invention.

FIG. 3 is a diagram showing an image captured by the image capturing apparatus according to the embodiment of the present invention.

FIG. 4 is a diagram showing a method for removing noise of luminance of an image according to the embodiment of the present invention.

FIG. 5 is a first diagram showing an example of a model image in the embodiment of the present invention.

FIG. 6 is a second diagram showing an example of a model image according to the embodiment of the present invention.

FIG. 7 is a flowchart showing an article recognition process in the embodiment of the present invention.

FIG. 8 is a diagram showing a method for setting illumination illuminance in the embodiment of the present invention.

FIG. 9 is a diagram showing an actually measured value of a change in image brightness with respect to a change in illumination illuminance in the embodiment of the present invention.

[Explanation of symbols]

10 Control device 11 Imaging device 13 Model image storage 14 Recognition processing unit 21 board 22 Electronic components 24, 24a, 24b, 24c Character string

   ─────────────────────────────────────────────────── ─── Continued front page    F term (reference) 2G051 AA65 AB11 AB14 BB01 BC01                       CA04 EB01 EC02                 5B057 AA03 BA02 CA12 CA16 DA01                       DB02 DC08 DC34                 5B064 AA05 BA01 DB07 DB10 DB18                       EA34                 5L096 BA03 BA17 FA67 HA07 JA03

Claims (10)

[Claims] 1. An image pickup device for photographing an image of an article, (b) a storage unit for storing a model image, and (c) an image of a character string included in the image of the article is compared with the model image. Pattern matching is performed, and the presence or absence of the article,
An article recognition device, comprising: a recognition processing unit that detects a position, an angle, a direction, or an inclination. 2. The recognition processing unit compares the image of the character string with one model image to perform pattern matching, and then compares the image of the character string with a plurality of model images to perform pattern matching. The article recognition device according to claim 1. 3. The article recognition device according to claim 1, wherein the recognition processing unit changes the illumination illuminance and performs the pattern matching again. 4. The article recognition device according to claim 1, wherein the character string includes a plurality of characters. 5. The article recognition device according to claim 1, wherein the article is an electronic component mounted on a substrate, and the character string is a character string printed on the electronic component. . 6. (a) capturing an image of an article, (b) comparing an image of a character string included in the image of the article with a model image to perform pattern matching,
An article recognition method characterized by detecting an angle, a direction, or an inclination. 7. The pattern matching according to claim 6, wherein the image of the character string is compared with one model image to perform pattern matching, and then the image of the character string is compared to a plurality of model images to perform pattern matching. Item recognition method. 8. The article recognition method according to claim 6, wherein the pattern matching is performed again by changing illumination illuminance. 9. The article recognition method according to claim 6, wherein the character string includes a plurality of characters. 10. The article recognition method according to claim 6, wherein the article is an electronic component mounted on a substrate, and the character string is a character string printed on the electronic component. .