JP2003216930A - Method and apparatus for inspecting discoloration - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method and apparatus for inspecting discoloration enabling efficient inspection for discoloration while reducing the number of parameters used in determining discoloration errors. <P>SOLUTION: The discoloration inspection apparatus comprises an illuminating means 1 for applying white light to a workpiece 7, an image pickup means 2 consisting of a color television, a determining process part 3 consisting of a computer, an output part 4 consisting of a CRT outputting the result of an inspection process, and the like. The determining process part 3 includes a comparison operation part 9 for judging whether the workpiece is good or bad according to the object feature quantities, a reference feature quantity, and a threshold r, and an automatic threshold setting part 11 for automatically setting the threshold r. The automatic threshold setting part 11 has a threshold setting part 13 which, using images obtained from good and bad products determined by visual inspection, sets the threshold r according to two feature quantities with a great degree of separation between the good and bad products, e.g. the feature quantities extracted by a feature quantity extracting part 12 and a feature quantity extracting part 12 which automatically extract G/B (green/blue) values and R/G (red/green) values. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a color change inspection method and a color change inspection apparatus. Specifically, the present invention relates to a discoloration inspection method and a discoloration inspection apparatus that determine the degree of discoloration of a product by image processing when determining the quality of the product.

[0002]

2. Description of the Related Art An image inspection apparatus has been developed which can perform various visual inspections that have hitherto relied on human vision accurately, at high speed and automatically. As this image inspection apparatus, for example, Japanese Patent Application Laid-Open No. 62-127617 discloses an apparatus which recognizes the shape of an inspection object and determines the quality of the product.

However, in a terminal inspection of a relay, which is an electronic component, for example, a discoloration failure of a contact spring may occur due to human intervention in the manufacturing process, and it may be processed as a product complaint. When such a color change abnormality occurs, the quality of the product cannot be determined by the image processing using the monochrome image as described above.

On the other hand, in recent years, there is an increasing need for inspection using a color image such as color discrimination of an inspection object and extraction of a specific color, and image processing for extracting the inspection object from the color image is performed. It is supposed to be. For example, in Japanese Patent Laid-Open No. 11-110552, in addition to the R value, G value, and B value of each pixel in the imaging region, the difference value between the R value and the G value, the difference value between the B value and the R value, the B value, and the G value. Of the difference value, the R value, the G value, and the B value corresponding to the pixel to be extracted.
Attempts have been made to set thresholds for values and B values separately and to extract pixels within the thresholds as the same color as the area to be extracted.

However, in this method, R for each pixel is
It is necessary to set a threshold value based on six parameters, namely, a value, a G value, a B value, a difference value between the R value and the G value, a difference value between the B value and the R value, and a difference value between the B value and the G value. When applied to the defect determination, it is difficult to understand which parameter is most effective for the determination. In addition, there is a problem that the threshold value cannot be set efficiently at the time of color change determination.

[0006]

DISCLOSURE OF THE INVENTION The present invention has been made in view of the above problems of the prior art, and a color change inspection method capable of efficiently performing a color change inspection by reducing the number of parameters used in determining a color change defect. And a discoloration inspection device.

[0007]

An image obtained by illuminating an illuminating means for emitting white light and picking up an image by a color image pick-up means arranged almost directly above a work. Among the multiple feature quantities of the combination of red, green, and blue components, the separation value between good and defective products is a certain value or more 1
Alternatively, it is characterized in that the color difference is determined by using two or more feature amounts.

A discoloration inspection method according to a second aspect is the discoloration inspection method according to the first aspect, wherein the characteristic amount is
The ratio of red / green / blue components: red / green value, green / blue value, red /
It is characterized by using the blue value.

A discoloration inspection method according to a third aspect is characterized in that, in the discoloration inspection method according to the second aspect, a feature amount having a large degree of separation obtained by using the following expression 1 is used.

[0010]

[Formula 1]

A discoloration inspection method according to a fourth aspect is the discoloration inspection method according to any one of the first to third aspects, in which two characteristic quantities selected from the characteristic quantities are used to determine whether the quality is good or bad. It is characterized by setting a separation value.

The discoloration inspection method according to claim 5 is the discoloration inspection method according to any one of claims 1 to 4, wherein the separation value is a defective product coordinate closest to the good product average coordinate and the good product average coordinate. The feature is that it is half the distance between them.

The color change inspection method according to claim 6 is
The discoloration inspection method according to claim 5 is characterized in that the feature amount obtained from the image of the inspection object product is arranged on the coordinates having the good product average coordinates as the origin.

According to a seventh aspect of the present invention, there is provided a discoloration inspection apparatus in which white light illuminating means for illuminating a work, color image pick-up means arranged almost directly above the work, and red, green, and green images of the obtained image are displayed.
Among a plurality of feature quantities based on the combination of blue components, a determination processing unit that determines a color difference using one or more feature quantities in which the separation value between the good product and the defective product is a certain value or more is provided. It has a feature.

The discoloration inspection apparatus according to claim 8 is the discoloration inspection apparatus according to claim 7, wherein the red / green value, the green / blue value, and the red / blue value, which are the ratios of the red, green, and blue components, are calculated. It is characterized by using.

A discoloration inspection apparatus according to a ninth aspect is characterized in that it is provided with a feature amount extraction means for selectively extracting one or more feature amounts having a large degree of separation obtained by using the following expression 1. There is.

Further, the discoloration inspecting apparatus according to claim 10 is the discoloration inspecting apparatus according to claim 9, wherein the feature quantity selected and extracted by the feature quantity extracting means is used to determine the good product average coordinates and the good product average. It is characterized by including a threshold value setting unit that calculates a distance between defective product coordinates closest to the coordinates and sets half of the distance between the two coordinates as a separation value.

Further, in the discoloration inspection device according to claim 11, in the discoloration inspection device according to claim 10, the determination processing section is arranged on a coordinate having the non-defective item average coordinate as an origin,
The feature is that the feature amount obtained from the image of the inspection object product is arranged and determined.

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a schematic configuration diagram of a color change inspection apparatus which is an embodiment of the present invention. The discoloration inspection apparatus includes an illumination unit 1 that irradiates a work 7 with white light, an image pickup unit 2 that is a color television, a determination processing unit 3 that is a computer, an output unit 4 that is a CRT or a printer that outputs an inspection processing result, and an image pickup. A monitor 5 for displaying an image picked up by the means 2 and an input means 6 including a keyboard and a mouse are provided.

The image pickup means 2 picks up an image of the target area from directly above the work 7 and sends the image data to the determination processing section 3. Figure 1
In this case, the contact terminal of the relay is an inspection target, and the image pickup means 2 is arranged directly above the contact terminal. Lighting means 1
Illuminates the object to be inspected and is arranged almost directly above the object means. The illuminating means 1 is arranged so that the inspection target does not have shading or color unevenness.

The determination processing unit 3 stores a reference feature amount and a threshold value determined as non-defective products, a storage unit 10 for calculating a target feature amount from image data, a target feature amount and a reference feature. A comparison calculation unit 9 is provided for judging whether the product is a good product or a defective product based on the amount and the threshold value.

In the present invention, R, G, and B values obtained as image data are used as the characteristic amounts to be compared.
The value is not used as it is, but it is a value obtained by combining these three component values, and especially the difference between the good product and the defective product (=
A combination value that increases the degree of separation is used. As this combination value, for example, the difference between the values (the difference between the R value and the G value, G
Difference between value and B value, difference between B value and R value, ratio of each value (ratio of R value and G value (R / G value or G / R value), G value and B
Value ratio (G / B value or B / G value), ratio of B value and R value (B
/ R value or R / B value)), product of each value (product of R value and G value (RG value), product of G value and B value (GB value), product of B value and R value (BR value) )), Etc., and an appropriate combination is selected depending on the inspection target product. In the present invention, since it is used for the discoloration inspection of the contact terminal, the combination of increasing the degree of separation is the ratio of each value, more preferably the following three ratios among the six ratios, namely R / G. Value, G / B value,
The R / B value is selected, and more preferably, any one or more of these three ratios is selected as the feature amount. As the characteristic amount, one which is empirically considered to be good when discriminating between a good product and a defective product is selected, and the above-mentioned R / G value, G / B value, and R / B value are selected for the discoloration of the contact terminal. It is a thing. Further, by using the ratio of each value as the feature amount, stable determination can be performed even with respect to the light amount variation. It is conceivable that any one of these three characteristic values may be selected to determine whether it is a good product or a defective product. However, in the terminal color change determination, at least two or more values are characteristic because of the certainty of the determination. Selected as quantity.

Various criteria can be considered as the criteria for determining the degree of separation to be selected. For example, the one having a large difference between the non-defective product and the defective product or the separation factor calculated by the following equation 1 is used.

[0024]

[Formula 1]

The characteristic amount calculating section 8 calculates the characteristic amount of the inspection object product from the obtained image data. For the calculation, the image data of the imaging region set in advance by the input means 6 is used, and the feature amount selected above is calculated based on the image data. For the imaging region, a region in which it is empirically recognized that discoloration occurs is partially selected (indicated by a dashed circle in FIG. 1). In addition, R used to calculate the feature amount
As the value, the B value, and the G value, the total value of the values in each pixel or the average value of each pixel is used.

When the comparison calculation unit 9 compares the calculated target feature amount with the reference feature amount stored in the storage unit 10 and determines that the difference is within a certain threshold value r (= separation value). Is judged as a non-defective product, and when it is judged that the threshold value r is exceeded, it is judged as a defective product and a non-defective judgment result is output to the output unit 4. As the reference feature amount and the threshold value r, a value calculated by using a product which is previously judged to be a non-defective product by visual confirmation is used, and is input from the input means 6, for example. FIG. 2 shows a determination conceptual diagram when two feature amounts are used. In FIG. 2, among the above feature amounts, two feature amounts having a high degree of separation between a good product and a defective product, G / B value and R / G value, are used, and the threshold value r is mainly the reference feature amount that is a good product. Is drawn as the radius of the circle. That is, in this method, the two feature amounts used for the determination are not individually judged to be large and small, but the distance between the reference feature amount (represented by coordinates) and the target feature amount (represented by coordinates) is used. The determination is made based on whether the distance is larger or smaller than a fixed value (= threshold r). The two feature quantities can be represented on a plane coordinate, and the distance R is used to judge whether the product is good or bad.
It is judged by one value. Therefore, the number of comparison operations can be smaller than that in the case of making a judgment using individual values. Further, in FIG. 2, since two feature quantities are used, they are represented two-dimensionally, but of course, the same determination can be made using three feature quantities. In this case, the threshold value r is represented as the radius of a sphere centered on the reference feature amount, and more reliable color change determination can be performed.

Next, the inspection method will be described with reference to the flow chart shown in FIG. 3. When the inspection is started, an image is picked up by the image pickup means 2 and a certain area of the contact terminal as the inspection object is taken in as image data. (S21).
It is assumed that the reference feature amount and the threshold value r are preset in the storage unit 10 from the input unit 6. Next, the feature amount calculation unit 8 calculates two values, for example, R / G value and G / B value, as target feature amounts (S22). Then, the comparison calculation unit 9 calculates the distance R between the calculated target feature amount (coordinates) and the reference feature amount (coordinates) and compares it with the threshold value r (S23). As a result, if the distance R between the two exceeds the threshold value r, it is determined as a defective product, and if it is smaller than that (or the same), it is determined as a good product. In this way, the determination of non-defective products and defective products can be made reliably and efficiently with a few steps. In particular, by selecting a feature amount having a high degree of separation, defective product inspection can be performed with high accuracy, and reliable discoloration inspection can be performed using only a few two feature amounts.

FIG. 4 is a schematic configuration diagram of the inspection apparatus according to the second embodiment. The inspection apparatus includes a threshold automatic setting unit 11 that automatically sets the threshold r in the determination processing unit 3. The automatic setting unit 11 includes a feature amount extraction unit 12 that selectively extracts a feature amount that is a determination target using image data obtained from a good product and a defective product that are determined by visual inspection, and the feature amount extraction unit 12 A threshold value setting unit 13 that calculates a threshold value r based on the feature amount selected and extracted in step S1 is provided.

A combination of three components used for calculation, for example, in the above example, a ratio of R, G and B values (R value / G value, G value / B value, R value / B) is applied to the feature amount extraction unit 12. It is preset from the input means 6 to use (value) as a feature amount. The feature amount extraction unit 12 selectively extracts two feature amounts from which the non-defective product / defective product can be reliably determined. For this criterion, for example, the degree of separation obtained by the above equation 1 is used. This selective extraction is performed by imaging and image-processing an inspection target product that has been previously determined as a non-defective product and a defective product by visual confirmation.

The threshold value setting unit 13 uses the two feature quantities selected by the feature quantity extraction unit 12 to determine a threshold value r used as a criterion.
Is stored in the storage unit 10. Various methods can be considered as a method of setting the threshold value r. For example, a method of setting a threshold value r from the average value of good products and its variance, a method of arbitrarily setting the threshold value r from the distance between the average value (coordinates) of good products and the average value (coordinates) of defective products, or FIG. As shown, there may be a method of setting the threshold value r from the distance R between the average value of non-defective products (non-defective product average coordinates) and the value of the defective product closest to the average value (nearest neighbor defective product coordinates). According to the example of FIG. 5, half of the distance (2r) between the average value of non-defective products and the nearest neighbor value is set as the threshold r.

The inspection method in this inspection apparatus will be described with reference to the flow chart shown in FIG. 6. First, image data is obtained from the image pickup means 2 by using some non-defective products and defective products (preferably N = 10 or more). Capture (S31). Next, three feature amounts of the selected R / G value, G / B value, and R / B value are calculated (S32). Of the three feature quantities, two feature quantities having a high degree of separation, for example, R / G value and G / B value, are selectively extracted by the feature quantity extracting unit 12 (S33). The average coordinates of non-defective products are obtained based on these two feature quantities (S34). Next, on the coordinates, the nearest point closest to the average value of the good products among the defective products is obtained (S35),
The threshold r is set by the threshold setting unit 13 (S3
6). Then, the threshold value setting unit 13 moves the entire coordinate system so that the average coordinates of non-defective products are the origin (S37). As a result, as shown in FIG. 7, the area judged to be non-defective is within the circle of radius r centering on the origin, and the distance R outside the origin is R.
The inspection target product at the position of is judged to be a defective product.

In this way, when the threshold value r is stored in the storage unit 10, the color images of the inspection object are sequentially captured (S3).
8) The non-defective product / defective product is determined by comparing the calculated feature amount (coordinate position) with the threshold value r (S39, S).
40). In the flow chart of FIG. 6, only one inspection target product is shown.

In this inspection apparatus, the selection and extraction of the characteristic amount used for the determination of non-defective product / defective product and the setting of the threshold value r can be automatically performed, so that the parameter setting can be easily performed. Also,
Since the average coordinates of non-defective products are moved to the origin on the coordinates, the amount of calculation required for determining the quality of the product to be inspected is reduced, high-speed data processing can be performed, and the inspection speed is improved.

As described above, since the non-defective product is determined to be a defective product by using the characteristic amount obtained by combining one or two or more color components, effective parameter setting becomes easy. In addition, since the inspection target value can be represented on a two-dimensional coordinate by using two feature quantities having a large degree of separation,
The single judgment parameter simplifies the judgment process. Further, by using a feature amount having a high degree of separation between non-defective products and defective products, highly accurate inspection can be realized and the inspection speed can be increased.

Further, in the above embodiment, the color change inspection of the contact terminals of the relay has been described, but the present invention can be applied to all products requiring the color change inspection such as not only the contact terminals of the relay but also other electronic parts. Needless to say. In this case, the feature amount may be selected according to the inspection target and the threshold value r may be set.
By using 1, the feature amount can be easily selected and the threshold value r can be set by using the few samples that have been determined to be non-defective and defective.

[0036]

According to the discoloration inspection method of the present invention, among a plurality of feature quantities of combinations of red, green and blue components of the obtained image,
The color difference is determined by using one or two or more feature amounts in which the separation value between the good product and the defective product is a certain value or more. For this reason, it is possible to make a judgment with one separation value as a judgment parameter, and it becomes easy to set effective parameters. Also, the determination process is simplified.

At this time, according to the method of claim 2, as the feature amount, red / green values, green / blue values, and red / blue values, which are ratios of red, green, and blue components, are used. Therefore, there is little influence of fluctuations in light quantity, and stable judgment can be performed.

Further, in the method according to claim 3, as shown in the equation 1, the characteristic amount having a large degree of separation obtained by using the dispersion in the good product and the dispersion in the defective product and the dispersion between the good product and the defective product is obtained. Since it is intended to be used, more accurate inspection can be performed.

Further, in the method according to the fourth aspect, since the separation value between the good product and the defective product is set by using the two feature values selected from the feature values, the parameter for setting the separation value is set. Only two are required, and the number of calculations for parameter calculation is reduced, and the processing time is reduced. For example,
As the separation value, as described in claim 5, it can be set to half the distance between the good product average coordinate and the defective product coordinate closest to the good product average coordinate.

Further, in the method according to the sixth aspect, since the feature quantity to be inspected is arranged on the coordinates having the good product average coordinates as the origin, the feature quantity to be inspected is separated from the origin by the coordinate position. The value is calculated immediately and the number of operations can be reduced.

The discoloration inspection apparatus of the present invention uses a combination of white light illuminating means for illuminating the work, color image pickup means arranged almost directly above the work, and red, green and blue components of the obtained image. A feature is that a determination processing unit for determining a color difference is provided by using one or two or more feature amounts in which the separation value between the non-defective item and the defective item is a certain value or more among the plurality of feature amounts. In the inspection device,
Of the multiple feature quantities based on the combination of green and blue components,
Since the color difference is determined by using one or two or more feature amounts in which the separation value between defective products is a certain value or more, it is easy to set an effective determination parameter. Also, the inspection process is improved because the determination process is simplified.

At this time, in the discoloration inspection apparatus according to the eighth aspect, the red / green value, the green / blue value, and the red / blue value, which are the ratios of the red, green, and blue components, are used as the characteristic quantities. Because
Judgments can be made that are less affected by fluctuations in light intensity.

Further, in the discoloration inspection apparatus according to claim 9,
As shown in Expression 1, a feature quantity extraction means is provided for selectively extracting one or more feature quantities having a high degree of separation obtained by using the variance in the good product, the variance in the defective product, and the variance between the good product and the defective product. Therefore, it is not necessary to input the parameter setting in advance, and the automatic setting can be performed by applying the good product and the defective product which are visually judged by holding them to the inspection device.

Further, as in the discoloration inspection apparatus according to the tenth aspect, by using the feature amount selected and extracted by the feature amount extracting means, the good product average coordinate and the defective product coordinate distance closest to the good product average coordinate. By including a threshold value setting unit that calculates the distance and sets half of the distance between the two coordinates as a separation value, the separation value (threshold value) is also automatically set, and the discoloration inspection work can be almost automated.

Further, in the discoloration inspection apparatus according to the eleventh aspect, the determination processing section determines by arranging the feature amount to be inspected on the coordinates having the good product average coordinates as the origin. Therefore, the number of calculations required for the determination is further reduced, and the inspection can be further speeded up.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram of a color change inspection device according to an embodiment of the present invention.

FIG. 2 is a conceptual diagram of determination of non-defective / defective products in the color change inspection device of FIG.

3 is a flow chart showing an inspection method in the discoloration inspection apparatus of FIG.

FIG. 4 is a configuration diagram of a color change inspection device according to another embodiment of the present invention.

5 is an explanatory diagram showing a threshold value setting method in the color change inspection apparatus of FIG.

6 is a flow chart showing an inspection method in the discoloration inspection device of FIG.

7 is a conceptual diagram of determination of non-defective products / defective products in the color change inspection device of FIG.

[Explanation of symbols]

2 Imaging device 3 Judgment processing unit 7 work 8 Feature quantity calculator 9 Comparison operation part 11 Automatic threshold setting unit 12 Feature Extraction Unit 13 Threshold setting section

Continued front page    F-term (reference) 2G020 AA08 DA05 DA22 DA31 DA34                       DA35 DA66                 2G051 AA90 AB20 CA04 EA17 EB01                       EC02 ED21                 5B057 AA01 BA02 CA01 CA12 CA16                       CE16 DA03 DB02 DB06 DC25                       DC36                 5L096 AA02 BA03 CA02 FA15 FA33                       GA40 JA11

Claims (11)

[Claims] 1. A plurality of characteristic amounts of a combination of red, green, and blue components of an image obtained by turning on an illuminating unit that emits white light and picking up an image by a color image pick-up unit arranged almost directly above a work. Among them, a discoloration inspection method characterized by determining a color difference by using one or two or more feature amounts in which a separation value between a good product and a defective product is a certain value or more. 2. The discoloration inspection method according to claim 1, wherein red / green values, green / blue values, and red / blue values, which are ratios of red, green, and blue components, are used as the characteristic amounts. 3. The color change inspection method according to claim 2, wherein a feature amount having a large degree of separation obtained by using the following expression 1 is used. [Formula 1] 4. The discoloration inspection method according to claim 1, wherein a separation value between non-defective products and defective products is set by using two feature quantities selected from the feature quantities. 5. The discoloration inspection according to claim 1, wherein the separation value is half the distance between the good product average coordinate and the defective product coordinate closest to the good product average coordinate. Method. 6. The discoloration inspection method according to claim 5, wherein the feature amount obtained from the image of the inspection object product is arranged on the coordinates having the good product average coordinates as the origin. 7. A white light illuminating means for illuminating a work, a color image pick-up means arranged almost directly above the work, and a plurality of feature amounts of a combination of red, green and blue components of the obtained image. A discoloration inspection apparatus comprising a determination processing unit that determines a color difference using one or two or more feature amounts in which a separation value between a good product and a defective product is a certain value or more. 8. The discoloration inspection apparatus according to claim 7, wherein red / green values, green / blue values, and red / blue values, which are ratios of red, green, and blue components, are used as the feature quantities. 9. The feature quantity extraction means for selectively extracting one or more feature quantities having a large degree of separation obtained by using the following expression (1) is provided. Discoloration inspection device described in. 10. The non-defective item coordinate distance closest to the non-defective item average coordinate and the non-defective item average coordinate is calculated using the feature amount selected and extracted by the feature amount extracting means, and half of the inter-coordinate distance is calculated. The discoloration inspection apparatus according to claim 9, further comprising a threshold value setting unit that sets the color separation value as a separation value. 11. The determination processing unit arranges and determines the feature amount obtained from the image of the inspection target product on the coordinates having the good product average coordinates as the origin.
Discoloration inspection device described.