JPH0413953A - Detect inspection preprocessor for electronic component molded form - Google Patents

Abstract

PURPOSE:To automatically discriminate propriety of an electronic component molded from with high reliability inexpensively by receiving imaging information of the form having a rectangular surface, obtaining extracted image information, and then further binarizing binary image information at each pixel to obtain binary correction image information. CONSTITUTION:Intensities of pixels are detected by a defect separation processor 16 to obtain variation in the luminous intensity of the adjacent pixels as extracted image information. The extracted image information is compared with a predetermined threshold value at the respective pixels to binary image information, 9 pixels including a defect pixel is then recognized as a defect region to obtain a corrected image information. The binary information is transferred to a defect detection processor 17, which obtains correlation between the directional component and linear pattern of the binary corrected image information in lateral and longitudinal directions to obtain binary correlation image information.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to an inspection device for inspecting the quality of molded products for electronic components, and particularly for testing the quality of molded products for electronic components molded by injection molding. The present invention relates to a preprocessing device used in this process.

[Prior Art] Generally, molded products molded by injection molding, for example,
The quality of plastic molded products for electronic components, especially those that have multiple rectangular squares on one side, such as connectors, is such that the rectangular squares of the molded product have dents due to short shots, Particular problems with defects occur when cracks, sink marks, warpage, etc. occur.

Conventionally, the quality inspection of such molded products for electronic components is usually performed by visual inspection by an operator.

[Problems to be Solved by the Invention] As described above, conventional quality inspections of molded products for electronic components are carried out by visual inspection by operators, so it is inevitable that defective products will be overlooked, and inspections tend to take a lot of time. There is a problem that the inspection cost increases as a result.

SUMMARY OF THE INVENTION An object of the present invention is to provide a pre-processing device for defect inspection of molded products for electronic components, which can determine the quality of molded products for electronic components with high reliability.

Another object of the present invention is to provide a defect inspection pretreatment device for electronic component molded products that can determine whether the electronic component molded product is good or defective at low cost.

Still another object of the present invention is to provide a defect inspection preprocessing device for electronic component molded products that can automatically determine whether the electronic component molded product is good or defective.

[Means for Solving the Problems] According to the present invention, there is provided a plurality of square groups in which a plurality of rectangular squares are consecutively arranged in a predetermined first direction, and the plurality of square groups are arranged in a first direction. It is used for defect inspection of molded products for electronic components that have rectangular surfaces that are continuously arranged in a second direction that is coplanar with the direction of , and perpendicular to the first direction. an imaging means for obtaining the information, and receiving this imaging information;
a first extracting means for extracting a change in brightness to obtain image information; a second extracting means for obtaining extracted image information by removing the edge portion of the square from this image information; The extracted image information is first extracted for each pixel based on a threshold of 1.
and a first 2@ converting means to obtain binary image information by binarizing either one of the first value and the second value; a changing means for obtaining changed image information with a region included as a first value; and a changing means that uses the changed image information to recognize the above region as a first level region and recognize a portion other than the above region as a second level region. Modified image information can be obtained from the imaging information, the first level range in this modified image information is binarized using a preset second threshold, and the second level range is binarized using a preset third threshold. There is obtained a defect inspection preprocessing device for a molded product for electronic components, which is characterized by having a correction means for digitizing and obtaining binarized corrected image information.

In this case, the imaging means includes a camera section disposed above the rectangular surface and a light source section that irradiates light onto the molded product for electronic components, and this light source section is set in advance in the first direction. (e.g., an extremely shallow angle (almost parallel, e.g. 0 to 10 degrees) with respect to the first direction (rectangular surface))
It is desirable to place the

[Operation] In the present invention, a rectangular surface is imaged by an imaging means to obtain imaging information. Next, image information is generated by extracting the change in brightness of this imaging information (that is, the brightness of each pixel is detected and the change value between the brightness of adjacent pixels is generated as extracted image information). Next, the edge portions of the squares described above are removed from this image information to obtain extracted image information. This extracted image information is binarized based on a first threshold value set in advance to obtain binarized image information. Next, regarding the binarized image information, pixels adjacent to the first pixel are expanded (referred to as the first pixel) having a first value (for example, logical "1" level) to be substantially enlarged (referred to as the first pixel). The logic "1" level is set and the first pixel is enlarged and emphasized. Then, the first pixel is enlarged and emphasized as changed image information. Regarding this modified image information, the enlarged and emphasized pixel portion is recognized as a first level region, and the portion (region) other than this pixel portion is recognized as a second level region, and modified image information is obtained from the imaging information. Then, in this modified image information, the first level range is binarized using a preset second threshold value, and the second level range is binarized using a preset third threshold value to obtain binarized corrected image information. .

In the binarized corrected image information obtained in this way, the first level range is essentially emphasized. can be easily understood.

By the way, by arranging the light source section of the imaging means at a predetermined angle with respect to the first direction, for example, in a direction substantially horizontal to the molded product for electronic components, the squares of the molded product for electronic components can be If there is a defect in the part, this defective part becomes a shadow, and the defective part is emphasized in the image information taken by the camera part.

[Example] Next, the present invention will be explained with reference to examples.

Referring to FIG. 1, the defect inspection device according to the present invention inspects the quality of a molded product for electronic parts, for example, a connector 11. The connector 11 has one surface (rectangular surface) as shown in the figure.
11a is divided into a plurality of rectangular squares by a partition wall, and these rectangular squares constitute a plurality of square groups (square space group) that are consecutively arranged in a predetermined first direction. A plurality of square groups are continuously arranged in a second direction that is on the same plane as the first direction and perpendicular to the first direction, thereby forming a rectangular surface. During inspection, the connector 11 is placed on the inspection table 12 with the rectangular surface 11a facing upward.

The defect inspection device has an irradiator 13 that irradiates a light beam onto the rectangular surface 11a of the connector 11.

This irradiator 13 is connected to, for example, a semiconductor laser (not shown) as a light source (wavelength near 80 nm).
). The irradiator 13 is equipped with a light emission window section 13a through which light is emitted as shown in the figure, and the length of this light emission window section 13a (
In FIG. 1, the length in the left-right direction) is longer than the length in the left-right direction of the connector 11. As a result, the light emission window 1
3a, the entire rectangular surface 11a can be irradiated.

The irradiator 13 is placed at an extremely shallow angle (
It is placed at a position where it can be irradiated with light almost horizontally (for example, 0 degrees to 10 degrees).

Above the rectangular surface 11a is a TV left camera, for example, a CCD.
The TV left camera 4 converts the reflected light from the rectangular surface 11a into an electrical imaging information signal (brightness level signal (reflected light pattern signal)).

As shown in FIG. 2(a), there are good connectors 11 with no defective parts at all, and inferior connectors 11 with defective parts as shown in FIG. 2(b).

When the connector 11 is imaged by the TV left camera 4 to obtain image information, the brightness of this image information changes depending on the presence of the defective portion.

Referring to FIGS. 3(a) to 3(c), assume that a defective product is imaged and imaging information is obtained. When observing the brightness distribution along the partition wall where there is no defective part, it is shown in Fig. 3(a).
As shown in the figure, the luminance distribution of the partition wall varies in proportion to the amount of scattered light depending on the surface roughness of the partition wall, but as shown in the figure, this variation is extremely small. On the other hand, when observing the brightness distribution along a partition wall where a large defect exists, the brightness distribution varies greatly in this defect as shown in FIG. 3(b). Furthermore, when observing the brightness distribution along the partition wall 11d where a small defect exists, the brightness distribution changes at this defect as shown in FIG. 3(c).

Here, with reference to FIGS. 1 and 4, the TV left camera 4
The captured image information (258×25 [i pixels)] is input to the defect detection device 15 (step sl). The defect detection device 15 includes a defect separation processing section 16 and a defect detection processing section 17. First, the defect separation processing section 16 separates a defective portion from the imaging information.

When the defect separation processing unit 16 receives the imaging information, it first performs high frequency component extraction processing (step s2). In this high frequency component extraction process, image information is obtained by extracting changes in brightness. That is, the brightness of each pixel is detected and the amount of change in brightness of adjacent pixels is used as an extracted image.

By the way, the connector 11 is installed in the defect separation processing section 16 in advance.
Position information indicating the position where the is placed is given. The defect separation processing unit 16 knows the boundary between the background and the rectangular surface 11a (square group) based on the above position information, and removes the rate of change in brightness (high frequency component) at the boundary. (The rate of brightness change is extremely large). In other words,
Step s to perform edge removal processing from the extracted image
3) Generate extracted image information.

Next, the defect separation processing unit 16 compares the brightness of each pixel of this extracted image information with a predetermined threshold, and sets the brightness to a logic "1" level when the brightness is equal to or higher than the first threshold.
When the brightness is less than the threshold value, it is set to a logical "0" level, and the image information is converted into binary image information. Then, pixels at the logic "1" level are determined to be defective pixels (step s4).

In this binary image information, all pixels adjacent to the defective pixel are rewritten to logic "1" level. In other words, all 9 pixels including the defective pixel are logically "
1'' level, this 9-pixel portion is recognized as a defective area (step s5), and modified image information is obtained.

Next, this defective area is recognized as a high-level area, the area other than this defective area is recognized as a low-level area, the high-level area and the low-level area are identified in the captured image, and corrected image information is obtained (step s6). In other words,
This corrected image information is then output as a defect isolated image.

As mentioned above, in the preprocessing according to the present invention, the defective part is first identified using the imaging information, the defective part is extracted and emphasized, and then the extracted defective part is represented on the imaging information. The corrected image information is as shown in Figure 5(a).
Only defective parts can be highlighted and identified, and as a result, defective products can be inspected extremely easily. On the other hand, when a captured image is simply binarized to obtain binarized image information (hereinafter referred to as simple binarized image information), it is possible to highlight and identify only the defective part, as shown in FIG. 5(b). Can not.

Here, referring to FIGS. 6 and 7, as shown in FIG. 6, the TV left camera 4 is placed above the rectangular surface of the connector 11, and the light from the irradiator 13 is directed to 11a
When irradiated perpendicularly to the rectangular surface 11a, the light is reflected upward as reflected light from the rectangular surface 11a, and the light is scattered at the defective portion. Then, imaging information is obtained from the change in reflected light intensity (change in brightness), and the defective portion is identified based on this imaging information.

Generally, a defective portion is extremely small (1 mm or less), so it may be difficult to distinguish between a change in reflected light intensity (change in brightness) due to the defective portion and a change in brightness due to the surface roughness of the rectangular surface 11a. As a result, as shown in Figure 7,
The defective part may not be properly imaged (in FIG. 7, it is shown as a simple binarized image).

On the other hand, when the rectangular surface 11a of the connector 11 is irradiated with light from an oblique direction (specifically, the rectangular surface 11a
When the light is irradiated at a shallow angle (θ), for example, 5 degrees with respect to a), that is, when the light is irradiated almost horizontally on the rectangular surface 11a, the light is not reflected at the defective part. As a result, the defective portion is imaged as a shadow by the imaging device 13. When the image information captured in this manner is subjected to simple binarization processing, it can be seen that the defective portion is properly imaged, as shown in FIG.

The image information preprocessed in this way, that is, the binarized corrected image information, is passed to the defect detection processing unit 17, and the defect detection processing unit 17 detects the quality of the connector 11 based on the binarized corrected image information. Execute.

Here, to explain the outline of the quality inspection process, referring to FIG. 10 and FIG. turn 18 (this vertical line pattern 18 is, for example, 15
The vertical line segments of the binarized corrected image information and the vertical linear pattern 18 are correlated with each other by sequential scanning in the horizontal direction to obtain first cross-correlation information (step s7). In other words, the correlation between the vertical line pattern defining the vertical portion of each square and the vertical portion of each square is calculated and the first correlation value is output. Then, the first cross-correlation information is standardized to 8-bit information (0 to 255) and becomes first standardized information. Similarly, the horizontal linear pattern 19 (this horizontal linear pattern 19 is about 15 pixels, for example) extending in the second direction (hereinafter simply referred to as the horizontal direction) is sequentially scanned in the horizontal direction, The correlation between the horizontal line segment of the digitized corrected image information and the horizontal linear pattern 18 is determined to obtain second cross-correlation information (step s8). In other words, the correlation between the horizontal linear pattern defining the horizontal portion of each square and the horizontal portion of each square is calculated and a second correlation value is output. Then, the second cross-correlation information is standardized to 8-bit information (0 to 255) and becomes second standardized information.

These first and second standardization information are preset third standardization information.
(for example, 255 x (14/15)), and when the first and second normalized information exceed the second threshold, they are each set to "1". That is, the first and second
If the normalized information exceeds the second threshold, the first and second
Each of the normalized information is substantially binarized as "1" (steps s9 and 510).

As described above, steps s7 to slO are performed in the vertical and horizontal directions of the binarized corrected image information. In the vertical direction, the vertical correlation image information shown in FIG. 12(a) is obtained, and in the horizontal direction, the vertical correlation image information shown in FIG. 12(b) is obtained.
) can be obtained.

By the way, the defect detection processing section 17 stores in advance a standard pattern of a rectangular surface of a non-defective product. This standard pattern has the number and length of vertical structures (vertical line portions) and the number and lengths of horizontal structures (horizontal line portions) on the rectangular surface. The defect detection processing unit 17 determines the outer circumference of a predetermined area of the vertical correlation image information, for example, the portion corresponding to each square (the length of the line segment that separates the portion corresponding to the square is less than or equal to 2, which is referred to as the vertical actual circumference length). ) (step 511)
. Similarly, the defect detection processing unit 17 determines the outer circumference length of the portion corresponding to each square of the horizontal correlation image information (the length of the line segment that separates the portion corresponding to the square is less than or equal to 2, referred to as the horizontal actual measured circumference length). (Step 512). Next, the defect detection processing unit 17 determines the vertical outer circumference length (hereinafter referred to as the vertical standard outer circumference length) and the horizontal direction outer circumference length (hereinafter referred to as the horizontal standard outer circumference length) of each square of the standard pattern, the vertical measured outer circumference length, and the horizontal measured outer circumference length. By comparing the measured outer circumference length with each square of the correlation image information (image information that specifies the rectangular surface 11a) obtained from the vertical correlation image information and the horizontal correlation image information, the outer circumference of each square portion of the standard pattern is calculated. A square with a short length is identified, and a position having 10'' is investigated for this identified square to identify a defective portion (step 513).

Thereafter, the defect detection processing unit 17 indicates the defective portion (for example, marks it with a circle) on the binarized corrected image, and displays the defective part on the display unit (not shown), for example, as shown in FIG. 12(c). Display as follows. Furthermore, the defect detection processing section 17 outputs the quality determination result to the outside.

[Effects of the Invention] As explained above, in the present invention, a molded product for an electronic component is imaged to obtain imaging information, and from this imaging information, preprocessing is performed to obtain binarized corrected image information in which defective parts are identified and emphasized. This has the advantage that using this binarized corrected image information, it is possible to easily and automatically determine whether a molded product for electronic parts is good or bad with high reliability.

Furthermore, a shallow angle (
For example, if the light is irradiated horizontally, the defective part will not be hit by the light, and the defective part will be captured as a shadow, that is, it will be emphasized. Defects can be determined.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the defect inspection device according to the present invention, FIGS. 2(a) and (b) are perspective views showing good and defective connectors, respectively, and FIGS. 3(a) to (
C) is a diagram for explaining a change in brightness due to a defective part in the connector. FIG. 4 is a flowchart for explaining the preprocessing in the defect inspection apparatus shown in FIG. 1, FIG. Figure (b) is a diagram showing simple binary image information, and Figure 6 is a diagram showing a state in which a TV camera is placed above the rectangular surface of the connector and light is irradiated perpendicularly to this rectangular surface. Fig. 7 is a diagram showing the image information captured in the state shown in Fig. 6 in a simple binarized image state, and Fig. 8 is a diagram showing a TV camera placed above the rectangular surface of the connector and Figure 9 shows a state in which light is irradiated at a shallow angle.
The figure shows the image information captured in the state shown in Fig. 8 in a simple binarized image state, and Fig. 10 shows the binarized corrected image information and the predetermined line in the defect inspection device shown in Fig. 1. Diagram for explaining calculation of correlation value with minute image, 11th
The figure is a flowchart for explaining defect detection in the defect inspection apparatus shown in FIG. 1, and FIGS. 12(a) to 12(c) are diagrams showing images obtained in the defect detection process, respectively. 11... Connector 12... Inspection table, 13
... Light projector, 14 ... TV camera, 15 ... Defect detection device, 16 ... Defect separation processing section, 17 ... Defect detection processing section. Figure 4 Figure 5 (Q) Figure 6 Figure 7 Figure 9 Figure 10 Figure 11 @ Mouth E Figure 12 (C1) Figure 12 (b) Figure 12 (c)

Claims (1)

[Claims] 1. It has a plurality of square groups in which rectangular squares are consecutively arranged in a predetermined first direction, and the plurality of square groups are on the same plane as the first direction. is used when inspecting molded products for electronic components having rectangular surfaces continuously arranged in a second direction orthogonal to the first direction for defects, and obtains imaging information by imaging the rectangular surfaces. a first extraction means that receives the imaging information and extracts a change in brightness to obtain image information; and a second extraction means that removes an edge portion of the square from the image information to obtain extracted image information. means, and first binarization for obtaining binarized image information by binarizing the extracted image information into either a first value or a second value for each pixel based on a first predetermined threshold value. and a changing means for obtaining changed image information by setting an area including pixels adjacent to pixels having the first value in the binarized image information as the first value;
The area is recognized as a first level area in the changed image information, the part other than the area is recognized as a second level area, corrected image information is obtained from the captured image information, and in the corrected image information, the area is recognized as a first level area. The second level range is set in advance.
A molded article for electronic parts, characterized in that it has a correction means for obtaining binarized corrected image information by binarizing the second level range using a preset third threshold. defect inspection pre-processing equipment. 2. In the defect inspection device for molded products for electronic components as set forth in claim 1, the imaging means includes a camera section disposed above the rectangular surface, and a camera section disposed above the rectangular surface; 1. A pretreatment device for defect inspection of molded products for electronic components, characterized by having a light projecting section disposed in a direction of a predetermined angle.