JP3367449B2 - Electronic component appearance inspection method, appearance inspection apparatus, and recording medium storing a program for causing a computer to execute the appearance inspection process - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a visual inspection method for an electronic component, a visual inspection apparatus, and a recording medium having a program for causing a computer to implement the visual inspection processing. Appearance inspection method of electronic parts for inspecting the appearance defects
The present invention relates to a visual inspection device and a recording medium in which a program for causing a computer to perform visual inspection processing is recorded.

[0002]

2. Description of the Related Art Conventionally, as a visual inspection method for electronic components, an inspection method or inspection for automatically detecting defects in the manufacturing process such as small holes (hereinafter referred to as voids) formed on the package surface of electronic components. There is a device.

As an inspection method of the first conventional example, Japanese Patent Laid-Open No.
In Japanese Patent No. 280958, an image obtained by photographing a surface of an inspection target is divided into a plurality of unit areas, and binarization processing is performed with an average value of a gray level in each of the unit areas. A method for determining the presence / absence of a defect from the shape is disclosed, and the technique described in the publication will be described as a first conventional example.

FIG. 11 is a block diagram of the defect inspection apparatus of the first conventional example.

In FIG. 11, the scanning beam emitted from the laser light source 20 is applied to the surface of the inspection object 10.
This reflected light is guided to the light receiver 28, and the image signals a1, a2
Is output as. The image signals a1 and a2 are added by the addition means 3
2, the A / D conversion means 34, the density conversion means 36, and the differential filter 38 to output the signal a5 in which the contour of the defect is emphasized. This signal a5 is obtained by the defective address detecting means 40 as the defective address signal Ad, while the signal a5 is obtained.
5 is binarized by the binarization processing means 42, and the defect area extraction means 48 extracts the defect area with reference to the defect address Ad and outputs it as the defect image signal A2.

As a method of determining this defect, the inspection target 1
The address of the defect is obtained based on the differential image of 0, and in the binarization processing means 42, from the first threshold value TH1 obtained from the gray level distribution of the entire differential image and the gray level average value of the neighboring region of the pixel of interest. From the obtained second threshold value TH2, TH3 = TH1-k (TH1-TH2) is set to obtain a third threshold value TH3, and in the defect area extraction means 48, the third threshold value TH3 corresponds to the position of the pixel whose gray level is smaller than the threshold value TH3. It was determined that the surface of the inspection target 10 had a defect.

FIG. 12 is a conceptual diagram of each area in the inspection apparatus of FIG.

FIG. 12 shows the inspection object 10 in FIG.
This is an example in which a more specific electronic component 2 is replaced. The captured image of the light receiver 28 includes a partial area of the package 2a of the electronic component 2 and the terminal 2b. Of these, the region including only the package 2a is the inspection target region Rt. The image of the package 2a captured as the captured image includes the void B which is a defect formed in the package 2a and the unevenness P caused by the surface of the package 2a or reflected light. FIG. 13 shows the defect inspection of FIG. It is a light / dark level L and a binarization level Ls-coordinate C characteristic diagram in an apparatus. FIG. 13 is a view of the level taken along the line XII-XII in FIG. 12, in which a gentle concave portion on the left side of the light and shade level L of the photographed image shows unevenness P, and a sharp concave portion on the right side shows a void B, respectively. .

The void B has a narrow peak width, and the gray level L
, The peak width of the unevenness P is wider than that of the void B, and the gray level L gradually changes.

In FIG. 13, the gray level L is the binarization level L.
A region larger than s is a region without void B and is determined to be normal, and a region with a gray level L smaller than the binarization level Ls is a region with "void B".

As is clear from FIG. 13, in the unevenness P, the light and shade level L slightly decreases over a wide range, but the binarization level Ls obtained as the average value of the light and shade level L also gently decreases. ing.

As the defect inspection method of the second conventional example, as a simpler processing method, an image of the surface of the package 2a is binarized at a constant binarization level over the entire image. There is a method of determining that "the void B exists" when the area of the region where the reflectance of the captured image is small and the gray level is smaller than the binarization level is a preset value or more.

FIG. 14 is a characteristic diagram of the lightness level L and the binarization level Ls-coordinate C in the defect inspection apparatus of the second conventional example. Similar to FIG. 13, FIG. 14 corresponds to the captured image of the package 2a shown in FIG. 12, and as is clear from FIG. 14, the binarization level Ls is constant over the entire area.

In this way, the conventional defect inspection method and apparatus have determined the presence or absence of defects such as voids on the package surface of electronic components.

[0015]

However, the conventional defect inspection method and apparatus have the following problems.

When the electronic component 2 is used as an inspection target,
The image signals a1 and a2 of the package 2a are as shown in FIG.
3, as shown in FIG. 14, the light intensity of the irradiation light source 1 that irradiates the surface of the package 2a is different depending on the direction and place, the resin component of the package 2a and the surface state are not uniform, and the mold of the package 2a is Often, the reflectance of the surface is partially different due to stains etc. adhering from etc., in such a case, in a part of the shooting area,
The unevenness P is likely to occur, in which the gray level L of the image signals a1 and a2 of the package 2a is totally different from other portions.

Accordingly, when the unevenness P and the void B have the same gradation level in the gradation image, it is judged that the gradation level L is smaller than the binarization level Ls after the binarization process. Therefore, although there is no void in the package 2a, there is a problem that it is easy to erroneously determine that "void B exists".

Further, in the defect inspection apparatus of the first conventional example, in order to obtain the average value of the gray level L for each of a plurality of unit areas,
Spatial filtering processing was performed based on the information of each pixel and a large number of pixels surrounding the pixel, resulting in an enormous amount of data to be processed, a long inspection time, and a large shade value of the stamp character near the stamp character. In addition, the average value of the light and shade levels in the unit area becomes high, and the binarization level is set to be high, so that there is a problem that normal inspection cannot be performed.

Further, in the defect inspection apparatus of the second conventional example, FIG.
As shown in FIG. 4, the binarization processing is simply performed at the constant binarization level Ls over the entire inspection target region Rt. However, in this method, the gray level L is entirely in the vicinity of the unevenness P in FIG. Therefore, there is a problem that it is erroneously determined that “void B exists” near the unevenness P.

In the present invention, the appearance inspection method, the appearance inspection apparatus, and the appearance inspection of the electronic component, which can surely discriminate only the true defect without being affected by the unevenness included in the photographed image of the package of the electronic component, etc. A recording medium for recording a program for causing a computer to realize a process is provided.

[0021]

In order to solve the above problems, the present invention employs the following novel characteristic methods and means.

The feature of the appearance inspection method for electronic parts of the present invention is that the surface of the inspection object (102 in FIG. 2) is photographed to obtain the gray level (L in FIG. 4) for each coordinate (C in FIG. 2). , A continuous area (R11 to Rl in FIG. 5) of coordinates (C) in which the gray level (L) is smaller than a preset value (Ls in FIG. 4).
Areas (Rl2 and Rl3 in FIG. 5) whose area is larger than a preset value among 3) are obtained as void candidate areas (Rc2 in FIG. 6A and Rc3 in FIG. 6B), and the void candidate area ( Rc2, Rc3) are spread around to obtain void candidate expansion regions (Re2 in FIG. 7A, Re3 in FIG. 7B), and each gray level (L) for each void candidate expansion region (Re2, Re3). ) Frequency distribution characteristic value (V in FIG. 9)
c2, Vc3), and if the feature value (Vc2, Vc3) is outside a certain range, the void candidate region (Rc2, Rc)
It is to determine that there is a defect in 3).

The feature of the appearance inspection apparatus for electronic parts of the present invention is that the surface of the inspection object (102 in FIG. 2) is photographed and the gray level (L in FIG. 4) is obtained for each coordinate (C in FIG. 2). Coordinates (C) where the image pickup means (103 to 105 in FIG. 1) and the gray level (L) are smaller than a preset value (Ls in FIG. 4).
Of the continuous areas (Rl1 to Rl3 in FIG. 5) whose area is larger than a preset value (Rl2 and Rl in FIG. 5).
3) is a void candidate region (Rc2 in FIG. 6A, FIG.
The void candidate area extraction means (106 to 108 in FIG. 1) to be obtained as Rc3 in (b) and the void candidate area (Rc).
2, Rc3) is expanded to the surroundings to obtain a void candidate expansion region (Re2 in FIG. 7A, Re3 in FIG. 7B), and a void candidate expansion region (109 in FIG. 1). A characteristic value calculation means (110 in FIG. 1) for obtaining a characteristic value (Vc2, Vc3 in FIG. 9) of the frequency distribution of the light and shade level (L) for each (Re2, Re3), and a characteristic value (Vc2, Vc).
3) is outside a certain range, the void candidate region (Rc
2, Rc3) is provided with a determination unit (111 to 112 in FIG. 1) for determining that a defect exists.

The characteristic of the recording medium in which the program for causing the computer to realize the appearance inspection processing of the electronic component of the present invention is recorded is that the surface of the inspection object (102 in FIG. 2) is photographed and the coordinates (in FIG. The lightness level (L in FIG. 4) for each C) is obtained, and the lightness level (L) is a preset value (L in FIG. 4).
s) a smaller continuous region (R11 to R13 in FIG. 5)
Regions (Rl2, Rl3 in FIG. 5) whose area is larger than a preset value are void candidate regions (Rc in FIG. 6A).
2, Rc3 in FIG. 6B), and the void candidate regions (Rc2, Rc3) are spread to the periphery to obtain void candidate expansion regions (Re2 in FIG. 7A, Re3 in FIG. 7B).
Characteristic value of frequency distribution of light and shade level (L) for each void candidate extended region (Re2, Re3) (Vc2, Vc3 in FIG. 9)
When the feature values (Vc2, Vc3) are out of a certain range, it is determined that a defect exists in the void candidate region (Rc2, Rc3).

By adopting such a method and means, the recording medium on which the program for causing the computer to implement the appearance inspection method, appearance inspection device and appearance inspection process of the electronic component of the present invention is first void. The candidate area is extracted, then the feature value is obtained from the distribution of the frequency of the gray level at each coordinate in these void candidate areas, and the presence or absence of voids is determined based on this feature value, so the effect of unevenness is eliminated. The presence or absence of defects such as voids can be reliably determined.

[0026]

FIG. 1 is a block diagram of a visual inspection apparatus for electronic parts according to a first embodiment of the present invention.

The irradiation light source 101 emits irradiation light.

The electronic component 102 reflects the irradiation light from the irradiation light source 101 according to the reflectance of the surface.

The CCD camera 103 images the reflected light from the electronic component 102 over a certain area and outputs it as an analog image signal Sa.

The analog image signal Sa from the CCD camera 103 is A / D converted by the A / D conversion means 104 and output as digital grayscale image data Db.

The grayscale image data storage means 105 temporarily stores the digital grayscale image data Db therein.

The binarization means 106 reads out the digital grayscale image data Db from the grayscale image data storage means 105, and the grayscale level L (FIG. 4) is preset inside for each coordinate C (FIG. 2). Binarized image data Dc in which the coordinate C is set to the "1" level if the value is equal to or lower than the level Ls (FIG. 4) and "0" in the opposite case is generated and output.

The labeling means 107 is supplied with the binarized image data Dc, and labels the respective regions in which the "1" level of the binarized image data Dc is continuous with the labeling regions R11 to R11.
Rl3 (FIG. 5), and the label data Dd as a set of the label names and the areas of the labeling areas Rl1 to Rl3.
Output as.

The label candidate data Rd is supplied to the void candidate area extracting means 108, and the labeling areas Rl1 to Rl are supplied.
3 (FIG. 5), only the labeling areas Rl2, Rl3 (FIG. 5) having an area larger than the area preset inside are void candidate areas Rc2, Rc3 (FIG. 6 (a),
(B)) and output as void candidate label data De.

The void candidate area expanding means 109 is provided for each void candidate area Rc2 represented by the void candidate label data De.
Rc3 (FIGS. 6 (a) and 6 (b)) is expanded to the surroundings by a predetermined range, and void candidate expansion regions Re2 and R2 are generated.
e3 (FIGS. 7A and 7B), and output as void candidate area expansion data Df.

The histogram calculating means 110 is used for the digital grayscale image data Db and the void candidate area expansion data Df.
Is supplied to each of the void candidate expansion regions Re2 and Re3 (FIG. 7 (a),
For each (b)), the distribution of the frequency of the gray level L (FIG. 4) represented by the digital gray image data Db is obtained by, for example, a histogram notation, and is output as histogram data Dg for each void candidate extended region Re2, Re3.

The feature value calculating means 111 is used for each of the void candidate expansion regions Re2 and Re3 represented by the histogram data Dg.
Each histogram (see FIGS. 8A and 8B) (see FIGS. 8A and 8B)
Based on (a) and (b), a feature value Vc described later is obtained.
2, Vc3 (FIG. 9) is calculated and output as feature value data Dh.

The judging means 112 compares each of the characteristic values Vc2 and Vc3 (FIG. 9) represented by the characteristic value data Dh with an internally preset value, and at least one of the characteristic values Vc2 and Vc3 is an internal value. If it is not within the predetermined range set in advance, it is determined that “void B exists” in the void candidate expansion regions Re2 and Re3, and the determination signal Si is output.

FIG. 2 is a conceptual diagram of each area in the appearance inspection apparatus for electronic parts of FIG.

Referring to FIG. 2, the CCD camera 103 shown in FIG.
The photographed image includes a partial area of the package 102a of the electronic component 102 and the terminal 102b. Of these, the region including only the package 102a is the inspection target region Rt. The image of the package 102a captured as a captured image includes a void B which is a defect formed in the package 102a and the unevenness P caused by the surface of the package 102a or reflected light. FIG. 3 illustrates the electronic component of FIG. 3 is an inspection process flowchart in the appearance inspection apparatus of FIG.

Irradiation light emitted from the irradiation light source 101 of FIG. 1 is reflected on the surface of the package 102a (FIG. 2), and CC
The light incident on the D camera 103 is converted into A / D conversion means 104.
The grayscale image data storage means 105 is converted to digital grayscale image data Db which is analog-to-digital converted by the digital signal and indicates the grayscale level L (FIG. 4) at each coordinate C (FIG. 2).
Is temporarily stored in (step S301).

The first binarizing means 106 of FIG. 1 reads the digital grayscale image data Db stored in the grayscale image data storage means 105 (step S302).

The first binarizing means 106 shown in FIG. 1 sets the gray level L (FIG. 4) indicated by the digital gray-scale image data Db for each coordinate C (FIG. 2) in advance as a binary value. First binarized image data in which the coordinate C whose gray level L is lower than the binarization level Ls as compared with the binarization level Ls (FIG. 4) is set to the “1” level and the opposite coordinate C is set to the “0” level. It is output as Dc (step S303).

The labeling means 107 of FIG. 1 sets the continuous areas of the "1" level represented by the first binarized image data Dc as the labeling areas Rl1 to Rl3 (FIG. 5), respectively. The label names and areas of the regions R11 to R13 are output as label data Dd as a set (step S304).

The void candidate area extraction means 108 of FIG.
Each labeling area R11 to R represented by the label data Dd
Among the l3 (FIG. 5), only the labeling regions Rl2 and Rl3 which are larger than the area set in advance as the void candidate regions Rc2 and Rc3 (FIGS. 6 (a) and (b)),
The data is output as void candidate area data De (step S
305).

The void candidate area expanding means 109 shown in FIG.
Each void candidate area R represented by the void candidate area data De
c2 and Rc3 (FIGS. 6 (a) and 6 (b)) are each expanded to the surroundings by a certain range, and each void candidate expansion region Re2
The void candidate area expansion data Df representing Re3 (FIGS. 7A and 7B) is output (step S306).

The histogram calculating means 110 of FIG. 1 is arranged for each coordinate C (FIG. 2) for each void candidate expansion area Re2, Re3 (FIGS. 7A, 7B) represented by the void candidate area expansion data Df. The distribution of the frequency of the light and shade level L (FIG. 4) is represented as a histogram (FIGS. 8A and 8B) and output as histogram data Dg (step S30).
7).

The histogram value Dg is supplied to the feature value calculation means 111 of FIG. 1, and each void candidate expansion region Re is supplied.
2, Re3 (FIGS. 7A and 7B), the feature values Vc2 and V, which will be described later, of each histogram (FIGS. 8A and 8B).
c3 (FIG. 9) is obtained and output as feature value data Dh (step S308).

The determination means 112 of FIG. 1 determines the void candidate extension areas Re2, R indicated by the label data Dg for each of the void candidate extension areas Re2, Re3 (FIGS. 7A and 7B).
The characteristic values Vc2 and Vc3 (FIG. 9) of e3 are compared with the values set in advance, and the characteristic values Vc2 and Vc are compared.
If c3 does not exist within the preset value range,
It is determined that there is a defect in the void candidate extension regions Re2 and Re3, and the determination signal Sj is output (step S3).
09).

FIG. 4 is a characteristic diagram of the lightness level L and the binarization level Ls-coordinate C in the appearance inspection apparatus for electronic parts of FIG.

FIG. 4 is a view of the level seen along the line II-II in FIG. 2, in which the gentle concave portion on the left side is uneven P.
The sharp recess on the right side is due to the void B.

In FIG. 4, the gray level L is the binarization level Ls.
The smaller areas are labeled areas R11 and R12.

FIG. 5 is a conceptual diagram of the labeling areas R11 to R13 in the appearance inspection apparatus for electronic parts of FIG. As shown in FIG. 5, in the binarized image data Dc binarized by the binarization means 106, each labeling area Rl
1 to Rl3 are labeled respectively to form labeling areas Rl1 to Rl3.
The areas of 11 to Rl3 are 4, 14, and 13, respectively.

FIG. 6 is a conceptual diagram of the labeling candidate regions Rc2, Rc3 in the electronic component appearance inspection apparatus of FIG.

When the area (previously set area) = 6, the void candidate area extracting means 108 has only the labeling areas R12 and R13 in FIG. 5 which are larger than 6 as the void candidate areas Rc2 and Rc3, respectively. It is output as void candidate area data De.

FIG. 7A is a conceptual diagram of the void candidate expansion area Re2 in the appearance inspection apparatus for electronic parts of FIG.
(B) is a conceptual diagram of a void candidate expansion region Re3 in the electronic component appearance inspection apparatus of FIG. 1.

The void candidate area expanding means 109 expands the surrounding two pixels around each coordinate C in each of the void candidate areas Rc2, Rc3, and expands the void candidate expanding area R respectively.
e2 and Re3 are generated.

As the extension method, a maximizing filter of size (3 × 3) centered on each coordinate C is processed by the number of pixels to be extended, or each coordinate C is centered (5 ×).
5) A method in which the size maximizing filter is processed once and expanded by two pixels can be considered, but the present invention does not limit this expansion method.

FIG. 8A is a light and shade level L in the void candidate expansion region Re2 in the electronic component appearance inspection apparatus of FIG.
8B is a histogram of the light and shade level L in the void candidate expansion region Re3 in the electronic component appearance inspection apparatus of FIG.

When the void B exists, the gray level L changes abruptly near the outer periphery of the void candidate regions Rc2, Rc3 (FIGS. 6A and 6B), but when the void B does not exist, the gray level L changes. Level L does not change rapidly. Therefore, the void candidate regions Rc2, Rc3 are expanded to the surroundings by a certain range, and the void candidate expansion regions Re2, Re3 (see FIG.
(A), (b)), and these void candidate expansion regions Re
If the distribution of the change in the lightness level L at each coordinate C of 2 and Re3 is expressed as a histogram, the distribution of the lightness level L on the histogram is clearly different between the void B portion and the unevenness P portion. Vc2, Vc
If 3 (FIG. 9) is obtained, both can be distinguished.

In the histogram of FIG. 8A, the void B
The gray level L in the void candidate expansion region Re2 including
Since it is clearly divided into the void B and the part other than the void B,
There are two peaks in the histogram, which are distributed in the lightness level L due to the void B and the lightness level L other than the void B.

On the contrary, in the histogram of FIG. 8B, the gray level L in the void candidate expansion region Re3 including the unevenness P
, There is no region where the gray level L changes abruptly, so the gray level L is distributed in a narrow range, and the histogram has only one peak.

Characteristic value Vc obtained by the characteristic value calculation means 111
2 and Vc3, the gray level L of the histogram
There are the number of peaks of the frequency, the variance value S of the frequency of the gray level L of the histogram, the standard deviation of the frequency of the gray level L of the histogram, the maximum value of the differential value of the frequency of the gray level L of the histogram, and the like.

When the characteristic values Vc2 and Vc3 are the number of peaks of the frequency of the light and shade level L of the histogram, from the minimum value to the maximum value of the light and shade level L, or in the opposite direction,
The peak number can be obtained by comparing the frequency of a certain gray level L with the frequency of the adjacent gray level L and counting the number of points where the frequency changes from increasing to decreasing.

Therefore, in the judging means 112, the peak is 2
If there is a void, it can be determined that the void B exists in the void candidate expansion region Re2 (FIG. 7A).

If the lightness level L of the histogram is smoothed before obtaining the number of peaks, small peaks due to noise components will not be counted as the number of peaks in the narrow lightness level L range of the histogram data. The number is required with high accuracy.

Next, when the characteristic values Vc2 and Vc3 are the variance value S of the frequencies of the gray level L of the histogram, the variance value S is obtained from (Equation 1).

[0068] XL is the frequency at the gray level L of the histogram, XA
Is the average value of the frequencies of all the light and shade levels L in the histogram, and N is the maximum value on the horizontal axis of the histogram, that is, the gradation of the light and shade level L.

Void candidate expansion region R in which void B exists
At e2, as shown in FIG. 8A, the frequency of the light and shade level L is distributed over the void B and the two peaks other than the void B, so the variance value S of the histogram becomes large.

Void candidate expansion region Re in which unevenness P exists
In FIG. 3, the frequencies of the light and shade level L are concentrated and distributed in one peak as shown in FIG. 8B, so that the variance value S of the histogram becomes small.

Therefore, if the threshold value is set in advance in the determination means 112 between the dispersion value S due to the void B and the dispersion value S due to the unevenness P, the void B is detected when the dispersion value S is larger than the threshold value. Void candidate extended area Re2 (Fig. 7 (a))
It can be determined to exist inside.

Further, as the characteristic values Vc2 and Vc3, the standard deviation, which is the square root of the variance value S, is used instead, which is the same as the variance value S.

Next, when the characteristic values Vc2 and Vc3 are the maximum values of the differential value of the frequency of the grayscale level L of the histogram, these maximum values are determined by the histogram (FIG. 8A,
It can be obtained by differentiating each frequency of the light and shade level L in (b)) to obtain each differential value which is the change amount of the frequency of the light and shade level L, and then searching the maximum value of these differential values.

When the void B exists, the lightness level L changes abruptly around the void B area, and therefore the differential value of the frequency of the lightness level L in the histogram becomes larger than in the case of the unevenness P.

Therefore, if a threshold value is set in advance in the judging means 112 between the maximum value of the differential value by the void B and the maximum value of the differential value by the unevenness P, the maximum value of the differential value is larger than the threshold value. It can be determined that the void B exists in the candidate region Re2.

This differential gray level Li is obtained from the maximum value of the difference between the frequency of the gray level L to be calculated in the histogram and the frequency of the surrounding gray level L, or the sum of the absolute values of the differences of each frequency. To be However, the present invention does not limit the method of obtaining the differential gray level Li.

As described above, in the present embodiment, (1)
A void candidate region in which a void may exist is extracted, (2) each void candidate region is expanded to obtain each void candidate expansion region, and (3) the density level frequency in each void candidate expansion region is calculated. The distribution was obtained, and (4) the feature value of this frequency distribution was determined to be present when there was a void, so the effect of unevenness could be eliminated, and the presence or absence of voids could be reliably detected. The judgment can be reduced significantly.

Next, a modified example of this embodiment will be described.

As a first modification, in the labeling means 107 of FIG. 1, a part of the inspection target area Rt of FIG.
A mask area excluded from the void candidate areas Rc2 and Rc3 may be set.

As a second modification, the labeling means 107 in FIG. 1 calculates the areas of the labeling areas Rl1 to Rl3, and the void candidate area extracting means 108 calculates the void candidates from the areas of the labeling areas Rl1 to Rl3. Region Rc2, R
Instead of extracting c3, the labeling means 107 causes the most distant 2 on the outer circumference of each of the labeling areas R11 to R13.
It is also possible to calculate the distance between the points and extract only the labeling regions R12 and R13 whose void candidate region extracting means 108 has determined that this distance is equal to or greater than a preset threshold value as the void candidate regions Rc2 and Rc3, respectively. .

As a third modification, the labeling means 107 of FIG. 1 calculates the areas of the labeling areas Rl1 and Rl2, and the void candidate area extracting means 108 calculates the voids from the areas of the labeling areas Rl1 and Rl2. Instead of determining the presence or absence, the labeling means 107 causes the labeling areas Rl1 to Rl3 to have areas and the labeling areas Rl.
1 to Rl3 and both of the distances between the most distant points on the outer circumference are calculated, and the void candidate region extracting means 108 sets thresholds for both the area and the distance in advance, and either one of them is set. Only the labeling regions R12 and R13 that are larger than the respective thresholds may be extracted as the void candidate regions Rc2 and Rc3, respectively.

Further, instead of performing the binarization processing by using the binarization means 106, the binarization level Ls may be obtained from the histogram obtained by the histogram calculation means 110 and the binarization processing may be performed. Examples of this method include a modal method and a method of subtracting a predetermined offset value from the gray level L at which the frequency of the histogram peaks, but the present invention does not limit the binarization processing method.

Further, as the characteristic values Vc2 and Vc3 obtained by the characteristic value calculating means 111, the presence or absence of the void B is determined by combining the number of peaks of the histogram frequency, the variance value S or the standard deviation, and the maximum value of the differential value. You may.

Further, the void candidate area expanding means 109 may obtain a circumscribed rectangle for each of the void candidate areas Rc2 and Rc3, and the rectangular areas obtained by expanding the circumscribed rectangle by a certain range may be used as the void candidate expanded areas Re2 and Re3, respectively.

FIG. 10 is a block diagram of the appearance inspection apparatus for electronic parts according to the second embodiment of the present invention. The present embodiment describes an embodiment suitable for a case where the electronic component appearance inspection apparatus according to the first embodiment of FIG. 1 is configured especially on a computer.

The appearance inspecting apparatus for electronic parts according to the present embodiment includes an input section 201, a CPU 202, and a memory 203.
, Irradiation light source 204a, CCD camera 204b, A
The A / D converter 204c, the display processing unit 205, the display unit 206, the external storage device 207, the interface unit 208, and the bus 209 are included.

The input section 201 is an operation means such as a keyboard and a remote controller which is composed of various operation keys and the like.
Various instructions are given to the PU 202. This means is realized by an alphanumeric keyboard, a dedicated input device, a mouse, a remote controller, etc., especially in a computer.

The CPU 202 is composed of a microcomputer or the like, and operates by a program or the like which is built in or supplied from the outside and controls each unit.

The memory 203 is composed of a memory such as a RAM and stores various data under the control of the CPU 202. This means, especially in computers, RAM
It is realized by various recording media such as a flash memory and a hard disk.

The irradiation light source 204a emits irradiation light.

The CCD camera 204b is the electronic component 102.
The reflected light from is imaged over a certain area and output as an analog image signal Sa.

The A / D converter 204c performs analog / digital conversion on the analog image signal Sa from the CCD camera 204b and outputs it to the bus 209 as digital grayscale image data Db.

The display processing unit 205 is connected to the bus 209 and converts the image data sent via the bus 209 into a display signal Sdp and outputs it.

The display unit 206 is an image display means such as a display device or a monitor, and displays the display signal Sdp as an image. This means, especially in computers,
It is realized by various display devices.

The external storage device 207 is a storage medium that stores various processing programs of the CPU 202, data similar to that stored in the memory 203, and the like.
Various data are written and read under the control of the PU 202. This means, especially in computers,
It is realized by various recording media such as AM, flash memory, and hard disk.

The interface unit 208 includes the CPU 20.
It is an interface with the outside that is connected to 2.

The bus 209 includes an input unit 201 and a CPU 20.
2, the memory 203, the A / D converter 204c, the display processing unit 205, the external storage device 207, and the interface unit 208 are mutually connected.

The appearance inspection method for electronic parts according to the second embodiment of the present invention is the same as the appearance inspection method for electronic parts according to the first embodiment.

Next, a recording medium recording a program for causing a computer to implement the appearance inspection processing for electronic parts according to the second embodiment of the present invention will be described.

In the appearance inspection method for an electronic component and the appearance inspection device for an electronic component according to the first embodiment, a program for causing a computer to implement the appearance inspection process for an electronic component is used as a control program for a dedicated appearance inspection device. Although built-in, in the present embodiment, similar processing is realized by a software program or the like in a general-purpose computer device or the like.

As a recording medium of this software program, it may be stored in an external recording device 207 such as a memory card, a floppy (registered trademark) disk, a hard disk, a CD-ROM, a DVD-RAM, or in the data storage section 203. A storage area dedicated to the program may be provided to store the program read from the external recording device 207.

Other configurations and processing flowcharts are the same as those of the recording medium in which the program for causing the computer to implement the appearance inspection method, the appearance inspection device, and the appearance inspection processing method for the electronic component according to the first embodiment. .

In each of the above embodiments, the binarization processing is performed for each pixel of the images captured by the CCD cameras 103 and 204b. However, in order to speed up the processing, a plurality of pixels are collectively binarized. I don't mind.

[0104]

By adopting the above-mentioned methods and means, a recording medium recording a program for causing a computer to implement the appearance inspecting method, appearance inspecting apparatus and appearance inspecting apparatus for electronic parts according to the present invention, It has the following effects.

As a first point, first, a void candidate area is found, then each void candidate area is expanded to the surrounding area to obtain a void candidate expanded area, and then the density level frequency in each void candidate expanded area is calculated. Since the void area and the nonuniformity area are discriminated from the characteristic value of the distribution, there is an advantage that the misjudgment of the nonuniformity as the void is significantly reduced and the inspection accuracy can be greatly improved.

As a second point, it is sufficient to extract only the void candidate regions and calculate the feature values of these void candidate regions, which has the advantage of significantly speeding up the inspection process.

[Brief description of drawings]

FIG. 1 is a block configuration diagram of an appearance inspection apparatus for electronic parts according to a first embodiment of the present invention.

FIG. 2 is a conceptual diagram of each area in the electronic component appearance inspection apparatus of FIG.

FIG. 3 is an inspection processing flowchart in the appearance inspection apparatus for electronic parts of FIG.

FIG. 4 is a characteristic diagram of a lightness level L and a binarization level Ls-coordinate C in the appearance inspection apparatus for electronic parts of FIG.

5 is a conceptual diagram of labeling regions Rl1 to Rl3 in the appearance inspection apparatus for electronic components of FIG.

6A is a conceptual diagram of a void candidate region Rc2 in the electronic component appearance inspection device of FIG. 1, and FIG. 6B is a conceptual diagram of a void candidate region Rc3 in the electronic component appearance inspection device of FIG.

7A is a conceptual diagram of a void candidate extension region Re2 in the electronic component appearance inspection apparatus of FIG. 1, and FIG. 7B is a void candidate extension area R in the electronic component appearance inspection apparatus of FIG.
It is a conceptual diagram of e3.

8A is a histogram of a gray level L in a void candidate expansion region Re2 in the electronic component appearance inspection apparatus of FIG. 1, and FIG. 8B is a void candidate extension area Re3 in the electronic component appearance inspection apparatus of FIG. Gray level L
Is a histogram of.

FIG. 9 is a conceptual diagram for determining a void B in the electronic component appearance inspection apparatus of FIG. 1.

FIG. 10 is a block configuration diagram of an appearance inspection device for electronic parts according to a second embodiment of the present invention.

FIG. 11 is a block configuration diagram of a defect inspection apparatus of a first conventional example.

12 is a conceptual diagram of each area in the defect inspection apparatus of FIG.

13 is a gray level L in the defect inspection apparatus of FIG.
3 is a characteristic diagram of a binarization level Ls and a coordinate C. FIG.

FIG. 14 is a characteristic diagram of a lightness level L and a binarization level Ls-coordinate C in a defect inspection apparatus of a second conventional example.

[Simple explanation of symbols]

2 electronic components 2a package 2b terminal 10 inspection target 20 laser light source 28 Light receiver 32 addition means 34 A / D conversion means 36 Density conversion means 38 Differential filter 40 defective address detecting means 42 Binarization processing means 48 defective area extracting means 101 Irradiation light source 102 electronic components 102a package 102b terminal 103 CCD camera 104 A / D conversion means 105 grayscale image data storage means 106 binarization means 107 labeling means 108 Void candidate area extraction means 109 Void candidate area expansion means 110 Histogram calculation means 111 Feature value calculation means 112 determination means 201 Input section 202 CPU 203 memory 204a irradiation light source 204b CCD camera 204c A / D converter 205 display processing unit 206 display 207 External storage device 208 Interface part 209 bus

─────────────────────────────────────────────────── ─── Continuation of the front page (58) Fields investigated (Int.Cl. ⁷ , DB name) G01B 11/00-11/30 102 G01N 21/84-21/958 G06T 7/00

Claims (24)

(57) [Claims] 1. A surface area of an object to be inspected is photographed to obtain a gray level for each coordinate, and an area having a larger area than a preset value in a continuous area of coordinates in which the gray level is smaller than a preset value. Is obtained as a void candidate region, the void candidate region is expanded to obtain a void candidate extended region, the feature value of the distribution of the frequency of the gray level for each of the void candidate extended regions is obtained, and the feature value is outside a certain range. In this case, it is determined that there is a defect in the void candidate area in this case. 2. The appearance inspection method for an electronic component according to claim 1, wherein the characteristic value is the number of peaks in the distribution of the frequency of the gray level for each of the void candidate expansion regions. 3. The appearance inspection method for an electronic component according to claim 1, wherein the feature value is a variance value in a frequency distribution of the gray level for each of the void candidate expansion regions. 4. The appearance inspection method for an electronic component according to claim 1, wherein the feature value is a standard deviation in a frequency distribution of the gray level for each of the void candidate expansion regions. 5. The appearance inspection method for an electronic component according to claim 1, wherein the feature value is a maximum value of a differential value of a frequency of the gray level for each of the void candidate expansion regions. 6. The appearance inspection method for an electronic component according to claim 1, wherein a region excluded from the void candidate region is set in a region of the photographed image. 7. The gray level is preset instead of obtaining an area having a larger area than a preset value among consecutive areas having coordinates whose gray level is smaller than a preset value, as a void candidate area. 7. A void candidate area is defined as an area in which a distance between two farthest points on the outer circumference is larger than a preset value among continuous areas having coordinates smaller than the specified value. A method for inspecting the appearance of electronic parts as described in (1). 8. A void candidate expansion region is obtained by expanding a rectangular region circumscribing the void candidate region to the periphery instead of expanding the void candidate region to the periphery to obtain the void candidate expansion region. The appearance inspection method for an electronic component according to any one of claims 1 to 6. 9. An image pickup means for photographing a surface of an inspection object to obtain a gray level for each coordinate, and an area having a preset value in a continuous region of coordinates in which the gray level is smaller than a preset value. A void candidate area extracting unit that obtains a larger area as a void candidate area, a void candidate area expanding unit that spreads the void candidate area around to obtain a void candidate expanded area, and a frequency of the gray level for each void candidate expanded area. The appearance of the electronic component, which comprises: a feature value calculation unit that obtains a feature value of the distribution of the distribution; and a determination unit that determines that a defect exists in the void candidate region when the feature value is outside a certain range. Inspection device. 10. The appearance inspection apparatus for an electronic component according to claim 9, wherein the characteristic value is the number of peaks in the distribution of the frequency of the gray level for each of the void candidate expansion regions. 11. The appearance inspection apparatus for an electronic component according to claim 9, wherein the feature value is a variance value in a frequency distribution of the gray level for each of the void candidate expansion regions. 12. The appearance inspection apparatus for an electronic component according to claim 9, wherein the characteristic value is a standard deviation in a frequency distribution of the gray level for each of the void candidate expansion regions. 13. The appearance inspection apparatus for an electronic component according to claim 9, wherein the characteristic value is a maximum value of a differential value of the frequency of the gray level for each of the void candidate expansion regions. 14. The appearance inspection apparatus for an electronic component according to claim 9, wherein a region excluded from the void candidate region is set in a region of the photographed image. 15. The void candidate area extracting means may be configured to obtain, as a void candidate area, an area having an area larger than a preset value among continuous areas having coordinates whose gray level is smaller than a preset value. In the continuous region of the coordinates where the gray level is smaller than a preset value, a region where the distance between the two most distant points on the outer circumference is larger than a preset value is obtained as a void candidate region. The visual inspection device for an electronic component according to any one of claims 9 to 14. 16. The void candidate area extracting means expands the void candidate area to the periphery to obtain a void candidate expansion area, and instead expands a rectangular area circumscribing the void candidate area to the void. The appearance inspection apparatus for an electronic component according to any one of claims 9 to 14, wherein a candidate extended area is obtained. 17. A surface of an inspection object is photographed to obtain a gray level for each coordinate, and the area is larger than a preset value in a continuous region of coordinates where the gray level is smaller than a preset value. Obtain a region as a void candidate region, obtain a void candidate extended region by expanding the void candidate region to the surroundings, obtain the characteristic value of the distribution of the frequency of the gray level for each void candidate extended region, and the characteristic value is a constant value. A recording medium recording a program for causing a computer to implement a visual inspection process for an electronic component, which is characterized by determining that a defect exists in the void candidate area in the above case. 18. The computer implements the appearance inspection process for an electronic component according to claim 17, wherein the characteristic value is the number of peaks in the distribution of the frequency of the gray level for each of the void candidate expansion regions. A recording medium on which a program for recording is recorded. 19. The computer is made to implement the appearance inspection process of an electronic component according to claim 17, wherein the characteristic value is a dispersion value in a distribution of frequencies of the gray level for each of the void candidate expansion regions. A recording medium on which a program for recording is recorded. 20. The computer is made to implement the appearance inspection process for an electronic component according to claim 17, wherein the characteristic value is a standard deviation in a frequency distribution of the gray level for each of the void candidate expansion regions. A recording medium on which a program for recording is recorded. 21. The appearance inspection process of an electronic component according to claim 17, wherein the characteristic value is a maximum value of a differential value of the frequency of the gray level for each of the void candidate expansion regions. A recording medium on which a program for recording is recorded. 22. The appearance inspection process for an electronic component according to claim 17, wherein a region excluded from the void candidate region is set in a region of the photographed image. A recording medium recording a program to be realized by a computer. 23. Instead of obtaining, as a void candidate region, a region whose area is larger than a preset value among continuous regions whose coordinates are smaller than a preset value, the gray level is preset. 23. Among the continuous regions having coordinates smaller than the specified value, a region in which the distance between the two most distant points on the outer circumference is larger than a preset value is obtained as a void candidate region. A recording medium having recorded therein a program for causing a computer to implement the appearance inspection process of the electronic component described in 1. 24. Instead of expanding the void candidate area to obtain a void candidate expansion area, a rectangular area circumscribing the void candidate area is expanded to obtain a void candidate expansion area. A recording medium recording a program for causing a computer to implement the appearance inspection process for an electronic component according to any one of claims 17 to 22.