JPH0829355A - Appearance inspection device and appearance inspection method - Google Patents

Abstract

(57) [Abstract] [Purpose] Regarding the improvement of the appearance inspection device, it is judged whether the inspection element pattern detected as a defect candidate is a true defect, and the defect which can be repaired is judged as "normal", and the inspection reliability is improved. To improve. A detection unit 11 for detecting a defect of an object 13 to be inspected from image data DIN, and a determination unit 12 for determining whether the defect detected by the detection unit 11 is an allowable defect or a true defect. The determination means 12 uses the image data D
A first labeling circuit 22A for enlarging the inspection element pattern of the inspection target 13 extracted from IN and labeling for each inspection element pattern, and for each inspection element pattern of the inspection target 13 excluding defects from the image data DIN The label output signal S1 from the first labeling circuit 22A and the label output signal S2 from the second labeling circuit 22B are compared with each other by the second labeling circuit 22B for labeling Area determination circuit for determining whether or not there is
22C and.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a visual inspection apparatus and a visual inspection method, and more particularly to an improvement in an apparatus and method for inspecting a wiring pattern on a printed circuit board. 2. Description of the Related Art In recent years, as electronic devices have become higher in density, the printed circuit boards in the electronic devices have been made into a multilayer thin film pattern for the purpose of high-density wiring. In this manufacturing process, pattern inspection such as disconnection due to lack of wiring pattern and short circuit due to copper residue is essential. Since the wiring patterns are becoming finer and finer, this inspection is no longer visible to the operator. Therefore, an automatic visual inspection device is used.

There are roughly two techniques for automatically inspecting a wiring pattern for defects. (1) It is a detection technology that optically reads a wiring pattern and converts it into an electric signal. Generally, a binarized image is used here. (2) It is an inspection theory that detects defects from the obtained image. The present invention relates to (2), for example, using a radial length measuring sensor, measuring the length of the wiring pattern in each direction, checking the presence or absence of the center, and Then, the length values of the patterns in multiple directions are converted into a 16-bit code called a "radial code", and the detected radial code is associated with the dictionary that stores the quality of the radial code of each element in advance. Then, a method of judging pass / fail (normal or defective) is adopted. This method
Since the dictionary data can be exchanged, it has versatility to be applied to various patterns.

However, according to the inspection logic of the conventional example, in the wiring patterns detected as defect candidates, even repairable defects which should be originally judged as "normal" are erroneously determined as defects. However, re-inspection may take a lot of time. Therefore, there is a demand for an apparatus and method capable of determining whether or not the inspection element pattern detected as a defect candidate is a true defect, determining that the defect that can be repaired is “normal”, and improving the inspection reliability. There is.

[0004]

11 to 13 are explanatory views according to a conventional example. FIG. 11 is a configuration diagram of a printed circuit board inspecting apparatus according to a conventional example, and FIG. 12 is a wiring pattern inspection flowchart according to the conventional example. 13 (A) and 13 (B) respectively show a printed circuit board and a copper residual pattern diagram for explaining the problem.

For example, according to the appearance inspection apparatus previously filed by the applicant of the present invention (Japanese Patent Laid-Open No. 5-264240), as shown in FIG. 11, a light source 1, an image acquisition system 2, and a defect inspection system 3 are provided. And a pass / fail judgment rule dictionary 4. The image acquisition system 2 includes a CCD camera 2A, a binarization circuit 2B and a storage circuit 2.
C. The defect inspection system 3 has a length measurement circuit 3A, a center detection circuit 3B, a radial coding circuit 3C, a category conversion circuit 3D, a category map creation circuit 3E and a comparison circuit 3F.

The function of the inspection apparatus will be described. For example,
Illumination light is emitted from the light source 1 to the printed circuit board 23 on which the wiring pattern, which is an example of the inspection target 13, is formed. An image of the inspection target 13 in this state is acquired by the CCD camera 2A. An analog image signal of this image is converted into a binary image signal by the binarization circuit 2B, and this image signal is temporarily stored in the storage circuit 2C. The memory circuit 2C outputs the image data to the length measuring circuit 3A.

The inspection element pattern (wiring pattern) is measured by the length measuring circuit 3A based on the image data. For this purpose, a radial length measuring sensor is used, and the length of the pattern in each direction is measured to check whether or not it is the center. With respect to the central one, the length values of the patterns in a plurality of directions are converted into a 16-bit code called a "radial code".

That is, the measured value of the pattern is output to the center detection circuit 3B. The center of the inspection element pattern is detected by the center detection circuit 3B, and the center signal is output to the radial encoding circuit 3C. The center signal and the length measurement value are converted into a 16-bit radial code by the encoding circuit 3C, which is output to the category conversion circuit 3D. The category conversion circuit 3D converts a 16-bit radial code into a category code, and based on this. The category map creation circuit 3E creates a category map and detects a defect signal indicating a defect candidate such as copper residue. The comparison circuit 3F compares the defect signal with the dictionary. With this, a dictionary storing the quality of each element radial code,
Correspond the detected radial code to the normal one,
It is classified as a defect candidate. The pass / fail result is the result output means 14
Output from

Next, a specific method for inspecting a wiring pattern will be described. For example, in the case of inspecting the printed circuit board 23 in which the copper residue-as shown in FIG.
As shown in FIG. 12, in step P1, the copper residue is extracted as an input category from the created category map. Then, in step P2, it is compared whether or not there is a category in the dictionary that matches the input category. At this time, if it is compared with the dictionary and there is it (YES), the process proceeds to step P3. If it does not exist (NO), the quality is output.

In step P3, the "two-sided search method" is read from the search method dictionary, and then in step P4, both sides are searched centering on the copper residue. Here, it is compared whether or not there is a hit category. Then, in step P5, it is determined whether or not there is a gap defect on both sides. At this time, if the search category can be detected (YES), a defect is detected. Specifically, the copper residue generated between the wiring pattern having the defective spacing and the land as shown in FIG. 13B and the copper residue generated next to the land and having the defective spacing are detected as defects. If it is not detected (NO), the wiring pattern is determined to be “normal”. Specific normal patterns are the copper residue isolated on the printed circuit board as shown in FIG. 13B and the copper residue generated between the wiring pattern and the land.

[0011]

By the way, according to the inspection logic of the conventional example, the copper residue as shown in FIG.
, Since there is a space defect category on both sides of the copper residual category, it is determined to be a copper residual defect, and when there is no defect, it is determined to be a normal copper residual. For this reason,
A copper residue having a property different from that of the copper residue defect shown in (B) will be determined as a copper residue defect. This is because the copper residue is judged to be a copper residue defect because there are gap defect categories on both sides. However, there is no fear that the copper residue will contact (short) with other normal patterns,
Since the conductor spacing is sufficiently wide, the pattern may be determined to be normal.

As a result, the inspection logic of the conventional example is effective for the copper residuals, but is not always effective for the copper residuals. As a result, the residual copper that can be repaired is erroneously determined to be a defect, which requires a lot of time for re-inspecting the printed circuit board and lowers the production yield. The present invention was created in view of the problems of the conventional example, determines whether the inspection element pattern detected as a defect candidate is a true defect, the defect that can be repaired is determined to be "normal", An object of the present invention is to provide an appearance inspection device and an appearance inspection method capable of improving inspection reliability.

[0013]

FIG. 1 shows a principle diagram of an appearance inspection apparatus according to the present invention. As shown in FIG. 1, the appearance inspection apparatus of the present invention includes a detection unit 11 for detecting a defect of an inspection target 13 from image data DIN, and the detection unit 1 described above.
It is characterized by further comprising: a determination unit 12 that determines whether or not the defect detected by 1 is an allowable defect.

In the appearance inspection apparatus of the present invention, the judging means 12 is the inspection object 1 extracted from the image data DIN.
A first labeling circuit 22A for enlarging the inspection element pattern No. 3 and labeling for each inspection element pattern, and a second labeling circuit 22A for each inspection element pattern of the inspection object 13 in which defects are removed from the image data DIN Of the labeling circuit 22B, the label output signal S1 from the first labeling circuit 22A and the label output signal S2 from the second labeling circuit 22B are compared to determine whether or not there is a defect generated in the defective area. And a region determination circuit 22C for determining.

The appearance inspection method according to the present invention uses the image data DIN.
The inspection element pattern of the inspection target 13 is extracted from the above, the extracted inspection element pattern is enlarged to set the inspection area, and it is determined whether or not the defect generated in the inspection area is a true defect. Characterize. In the appearance inspection method of the present invention, the determination as to whether the defect is a true defect is made by adding a serial number to each inspection element pattern of the enlarged inspection target 13.
And a second labeling process for assigning a serial number to each inspection element pattern of the inspection target 13 excluding the defect, and the defect is determined based on the correspondence between the first and second labeling processes. It is characterized in that it is determined whether it is a defect occurring in a region or a defect occurring in a region other than the defect region.

In the appearance inspection method of the present invention, the labeling processing is characterized by using a filter having a radius of a minimum interval between inspection element patterns, thereby achieving the above object.

[0017]

[Operation] The operation of the visual inspection apparatus of the present invention will be described. For example, when the detection means 11 detects a defect of the inspection object 13 from the image data DIN, the determination means 12 determines whether or not the defect is an allowable defect. Specifically, the inspection element pattern of the inspection object 13 extracted from the image data DIN is enlarged by the first labeling circuit 22A, and the first labeling process of adding a serial number to each inspection element pattern is executed.

For example, the circuit 22A performs a labeling process using a filter having a radius which is the minimum interval between inspection element patterns. On the other hand, in the second labeling circuit 22B, the inspection element pattern of the inspection object 13 in which the defect is removed from the image data DIN is created, and the second labeling process of giving a serial number to each inspection element pattern is executed. . As a result, the label output signal S from the first labeling circuit 22A
Label output signal S from the first and second labeling circuits 22B
2 is compared by the area determination circuit 22C to determine whether it is a defect generated in the defective area or a defect other than the defective area. The pass / fail result is the result output means 1
It is output from 4.

Therefore, among the defect candidates of the inspection element pattern detected by the detecting means 11, the defect which can be repaired can be judged as "normal" by the area judging circuit 22C, and only the true defect should be detected. You can This eliminates the need to judge a repairable defect as a “defect” as in the conventional example, and contributes to the provision of a highly reliable printed circuit board inspection device and the like.

Further, according to the appearance inspection method of the present invention, it is a defect that has occurred in the defective region or a defect that has occurred in a region other than the defective region based on the correspondence relationship between the first and second labeling processes. Is determined. For this reason, the defects generated in the defective area that fails the inspection,
That is, it is possible to separate defects that have occurred in a normal area that passes the inspection. The latter defect is a case where there is no possibility of contact (short) with other inspection element patterns and a sufficiently wide conductor interval is satisfied.

This contributes to the provision of a highly reliable visual inspection device for inspecting the wiring pattern of the printed circuit board. It also contributes to the substantial improvement of the production yield of the inspection object.

[0022]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, each embodiment of the present invention will be described with reference to the drawings. 2 to 10 are diagrams illustrating an appearance inspection device and an appearance inspection method according to each embodiment of the present invention.
FIG. 2 is a block diagram of a printed circuit board inspection apparatus according to an embodiment of the present invention, and FIG. 3 is a block diagram of its labeling circuit. 4A is an explanatory diagram of the pattern search method, and FIG. 4B is an explanatory diagram of the inspection area setting method.

5 (A) and 5 (B) are explanatory views of the labeling filter and its function, and FIGS. 6 (A) and 6 (B) are
It is explanatory drawing of relabeling of an inspection element pattern. FIG. 7 is a functional explanatory diagram of the area determination circuit. For example, a printed circuit board inspection device, which is an example of an appearance inspection device, includes a light source 1, an image acquisition system 2, and an image acquisition system 2 as shown in FIG.
A pass / fail judgment rule dictionary 4, a defect detection system 21 and a defect re-judgment system 22 are provided. Since the light source 1 and the image acquisition system 2 have the same functions as those of the conventional example, the description thereof will be omitted.

That is, the defect detection system 21 is an example of the detection means 11 and detects a defect of the printed circuit board (an example of the inspection target 13) 23 from the image data DIN. The defect detection system 21 includes a length measurement circuit 21A, a center detection circuit 21B, a radial coding circuit 21C, a category conversion circuit 21D, a category map creation circuit 21E and a comparison circuit 21.
F. The length measuring circuit 21A measures the wiring pattern (inspection element pattern) from the image data DIN. For example, the length of the pattern in each direction as shown in FIG. 4A is measured using a radial length measuring sensor.

The center detection circuit 21B detects the center of the wiring pattern from the length measurement value and outputs the center signal and the length measurement value to the radial coding circuit 21C. Radial coding circuit
21C converts the center signal and the measured value into a 16-bit radial code, and outputs it to the category conversion circuit 21D.
The category conversion circuit 21D converts the 16-bit radial code into a category code and outputs it to the category map creation circuit 21E. The category map creation circuit 21E is
Create a category map from the category code. The category map shows defect candidates such as copper residue, and these candidates are output to the comparison circuit 21F as the defect signal S0. Comparison circuit 21
F is compared with the defect signal S0 and the dictionary, and these comparison result signals SOUT are output to the selector 22D of the system 22.

The defect re-determination system 22 is an example of the determination means 12 and determines whether or not the defect candidate detected by the defect detection system 21 is an allowable defect. For example, the system 22 includes the first and second labeling circuits 22A and 22B, the area determination circuit 22C, and the selector 22.
Have D. The labeling circuit 22A, as shown in FIG.
It has an area designation unit 201, a primary labeling unit 202, a temporary label image frame memory 203, a label table 204, a connection information input / output unit 205, and a secondary labeling unit 206. The area designation unit 201 is the printed circuit board 2 extracted from the image data DIN.
Fig. 4 is an enlarged view of the inspection pattern for land 3 and straight leads.
Areas 1 and 2 as shown in (B) are set.

The primary labeling unit 202 uses a filter whose radius is the minimum interval between the wiring patterns as shown in FIG. 5A, and the center of the mask corresponds to the pixel of interest. As a result, when the label filter is applied to the inspection element pattern as shown in FIG. 5B, the same label (hereinafter, also referred to as a temporary label) is attached to the black portion of the in-circle drawing. As a specific example, as shown in FIG. 6 (A), a portion where the linear lead and the land are attached with “1” and a portion where the linear lead is attached with “2” are mixed in the first labeling.

The temporary label image frame memory 203 is a memory for temporarily storing the temporary label value. Label table 204
Registers, for example, a pair of temporary labels and a true label as shown in FIG. 6B as the combined information. The connection information input / output unit 205 inputs / outputs connection information. Secondary labeling section 20
6 re-labels the combined information (temporary label value) as shown in FIG. 6 (B) and outputs a true label image (multi-valued) to the area determination circuit 22C. This true label image (multi-valued)
Becomes the label output signal S1.

As a result, the wiring pattern or the like of the printed board 23 extracted from the image data DIN is enlarged, labeling processing is performed for each wiring pattern, and the label output signal S1 is output from the labeling circuit 22A to the area determination circuit 22C. be able to. As a specific example, all the linear leads and lands as shown in FIG. 6A are marked with "1" by the second labeling.

The inside of the labeling circuit 22B is the circuit 22.
The same as A. The function is to remove defects from the image data DIN of the printed circuit board 23 based on the comparison result signal SOUT from the selector 22D, add a serial number to each inspection element pattern, perform labeling processing, and output the label output signal S2 to the area determination circuit. Output to 22C. The area determination circuit 22C compares the label output signal S1 with the label output signal S2 and determines whether or not the defect is generated in the defective area. As a result, the area determination circuit 22C outputs the defect re-determination signal S3 to the pass / fail result portion. The output of the comparison circuit 21F is "defective" in the selector 22D.
Or select "Normal", for example, if "Defective",
The comparison result signal SOUT is output to the labeling circuit 22B.

The result output means 14 outputs the comparison result signal SOUT based on the defect re-determination signal S3. This allows
The pass / fail result of the printed circuit board 23 is output. In the pass / fail judgment rule dictionary 4, a defect search method, a search category, pass / fail when the search category is detected, and pass / fail when the search category is not detected are described as category 1. As a specific example, regarding copper residue on the printed circuit board 23, double-sided search, gap defect, copper residue defect, normal copper residue, etc. are described. The search method dictionary is referred to as the dictionary 4.

Next, with respect to the wiring pattern inspection method according to the embodiment of the present invention, the operation of the inspection apparatus will be described.
8 and 9 are flowcharts (No. 1 and 2) of inspecting a wiring pattern according to the embodiment of the present invention.
(B) is a supplementary explanatory diagram of the labeling process. For example, in the case of inspecting the printed circuit board 23 in which the copper residue is produced as shown in FIG.
First, in step P1, a category map is created. For example, each radial code corresponds to the category conversion circuit 21
Converted to a category (for example, copper remaining, protrusion) in D.

Next, in step P2, it is compared whether or not there is a category in the dictionary that matches the input category. If the input category and the dictionary are compared with each other (YES), the process proceeds to step P3. If not (NO),
Output pass / fail. In step P3, the "specific search method" is read from the search method dictionary, and then in step P3.
In step 4, a desired category is searched for on the category map. Here, it is compared whether or not there is a hit category. The search method in this case is, for example, as shown in FIG.
And a method for checking whether or not the designated category Y exists in both directions A and 180 that differ from the direction A by 180 degrees.

Then, in step P5, it is determined whether or not the search category can be detected. At this time, if the search category can be detected (YES), a defect is detected. At this time, the surrounding category distribution is searched around the defect candidate category, and the quality and defect type are globally determined based on the quality determination rule. Specifically, FIG. 13 (B)
The copper residue generated between the wiring pattern and the land having the defective spacing and the copper residue generated next to the land and having the defective spacing are detected as defects.

Up to this point, the process is similar to the conventional example, but in the embodiment of the present invention, the residual copper and the residual copper defect re-determination process is added in steps P6 to P9. That is, in step P6, it is determined whether the copper residue and the copper residue are defects. If it is a copper residual defect (YES), the process proceeds to step P7. When it is not a copper residual defect (NO), the quality is judged by the conventional method.

In step P7, a labeling process is performed based on the image data DIN and the comparison result signal SOUT. Specifically, the defect detection system 21 outputs the comparison result signal SOUT.
However, the selector 22D determines whether or not there is a copper residual defect. At this time, the input image p (i,
If j) is not a copper residual defect, the comparison result signal SOUT
Is output as a pass / fail result as it is, and if it is a copper residual defect, it is output to the labeling circuit 22B as a defect signal.

Also during this time, the image data DIN is transferred to the labeling circuits 22A and 22B. Labeling circuit 22
In A, the following operation is performed. For example, an inspection pattern such as a land or a straight lead of the printed circuit board 23 extracted from the image data (original image) DIN is enlarged by the area designation unit 201, and areas 1 and 2 shown in FIG. 4B are set. It In FIG. 4B, region 1 is a defect (defective region) because it includes two normal patterns of leads (lands with leads), and region 2 has only one normal pattern (land). , Normal (normal area).

Further, in the primary labeling unit 202, as shown in FIG.
From the target pixel as shown in (A), a circular one whose radius is the minimum conductor interval defined by the inspection standard is used, and the center of the mask corresponds to the target pixel. As a result, FIG.
The same label (hereinafter also referred to as a temporary label) is attached to the black portion of the in-circle drawing of (B). As a specific example, as shown in FIG. 6 (A), a portion where the linear lead and the land are attached with “1” and a portion where the linear lead is attached with “2” are mixed in the first labeling. This temporary label value is temporarily stored in the frame memory 203.

On the other hand, in the label table 204, for example,
A pair of temporary labels and a true label as shown in FIG. 6B are registered as the combined information. The connection information input / output unit
By controlling the input or output of the combined information by 205, the secondary labeling unit 206 re-labels the combined information (temporary label value) as shown in FIG. (Multi-value) is output to the area determination circuit 22C as the label output signal S1 (first labeling process).

That is, in the first labeling process, as shown in FIG.
As shown in (A), the filter center is made to correspond to the pixel of interest, the label of the pixel of interest is read, and if the label is attached to the inspection element pattern, the labels are attached in order from 1. Then, the conductor portion in the filter is labeled with the same label as the pixel of interest. This scanning is performed on the entire screen while sequentially checking according to raster scanning.

At this time, if the pixel of interest has no label and two or more different labels are attached at adjacent points, the oldest label attached to the pixel of interest is given to the pixel of interest. Furthermore, since each pixel adjacent to the pixel of interest should originally have the same label, it is recorded in the memory 203 that these labels are the same so that the data can be processed later.

When the first labeling is finished,
As shown in (A) and (B), the pixels to which the same label should originally be attached are re-labeled. With these processes, labeling of the entire screen is completed, and the screen shown in FIG.
Sequential numbers 1 to 14 as shown in are added to each inspection element pattern. In addition, a second labeling process is performed in which a serial number is added to each wiring pattern of the printed circuit board 23 excluding defects. At this time, the labeling circuit 22B removes the copper residue from the original image. Specifically, a 3 × 3 filter is used for the original image and labeling is performed in advance.

When the selector 22D transfers the comparison result signal SOUT indicating "there is a copper residual defect" to the circuit 22B, the labeling circuit 22B refers to the label corresponding to the pixel output as the copper residual defect. , Delete all pixels with the same label. As a result, the copper residual defect is deleted. Once all the copper residue in the image has been removed,
A code different from that of the first labeling process, for example, Roman letters, is used to add the serial numbers A to O to the respective inspection element patterns from which defects are removed as shown in FIG. 10B (second labeling process). .

Then, in step P8, the correspondence is checked based on the first and second labeling processes to determine whether it is a defective area or a non-defective area. Specifically, FIG. 10 (A),
Raster scanning is performed from the upper left of the acquired image on the printed circuit board 23 as shown in (B), and the label of the corresponding pixel is referred to. When the comparison of the entire image is completed, the labeling circuit
From 22A to the area determination circuit 22C, serial numbers 1 to 6, 2, 2 (copper remaining), 2, 7, 8 (copper remaining), 9, 10, 11,
12, 13, and 14 are output as the label output signal S1. In addition, the labeling circuit 22B to the area determination circuit 22C
Further, serial numbers A to F, B, no corresponding label, G, H, I, no corresponding label, and J to O are output as the label output signal S2.

As a result, the area determination circuit 22C grasps the correspondence as shown in FIG. 7, and in this correspondence, when one label corresponds to a plurality of labels,
The defective area is determined, and the other areas are determined to be normal areas. Next, whether the defect has occurred in the defective area in Step P9,
Alternatively, it is determined whether the defect is a defect other than the defective region. Specifically, the area determination circuit 22C compares the label output signal S1 with the label output signal S2 to determine whether the defect is in a defective area or a defect other than the defective area. To be done.

In this case, the criterion for judgment is that a normal wiring pattern having only one normal wiring pattern in one area is normal (normal area). Those having two or more exist as defective areas. For example, in FIG. 7, since the label 2 corresponds to the two labels B and G, the copper residue generated in the defective area is determined as a defect. Further, regarding the copper residue, since there is no corresponding pixel, it is determined that there is no label, and this label absence is the copper residue that has occurred in the normal area and is determined as a defect to be repaired.

That is, at the end of the determination of this area,
It is referred to whether the pixel determined as "copper residual defect" is a defective region, and if the defective region is a "defect", if the pixel is a normal region, it is "normal" and the region determination circuit 22C re-determines the defect in the quality result part. The signal S3 is output. As a result, the copper residual defect, which has been treated as a defect in the conventional example, is inspected as a normal pattern, and the copper residual, ... Is correctly determined.

As described above, the printed circuit board inspection apparatus according to the embodiment of the present invention includes the defect detection system 21 and the defect re-determination system 22, as shown in FIG.
The defect re-determination system 22 has two labeling circuits 22A,
22B, a region determination circuit 22C, and a selector 22D. Therefore, even among the defect candidates of the wiring pattern detected by the defect detection system 21, the labeling circuit 22
Two sets of serial numbers 1 and 2 labeled by A and 22B
6,2,2 (copper remaining), 2,7,8 (copper remaining),
9, 10, 11, 12, 13, and 14, and serial numbers A to F,
Based on the correspondence relationship between B, no corresponding label, G, H, I, no corresponding label, and J to O, the copper residual defect that can be repaired can be judged as "normal" by the area judgment circuit 22C, and Only the defects can be detected.

As a result, it becomes unnecessary to erroneously determine a repairable defect as a “defect” as in the conventional example, and it is possible to accurately determine the remaining copper. This contributes to the provision of a highly reliable printed circuit board inspection device and the like. According to the appearance inspection method of the present invention,
Then, it is determined whether the defect has occurred in the defective region or the defect has occurred in a region other than the defective region based on the correspondence between the second labeling process.

Therefore, it is possible to separate a defect generated in the defective area which fails the inspection from a defect other than the defective area, that is, a defect generated in the normal area. That is,
A copper residue having a property different from that of the copper residue defect as shown in FIGS. 10 (A) and 13 (B) can be determined as “normal” although it is a copper residue. This is because even if there is a space defect category on both sides, the copper residue will not be regarded as a defect because it is a defect that has occurred in a region other than the defect region. As a result, there is no possibility of contact (short-circuit) with other normal patterns, and the copper residue satisfying a sufficiently wide conductor interval is determined to be "normal".

As a result, it is possible to inspect whether or not the straight leads or lands that are normal patterns satisfy the sufficient conductor spacing defined by the inspection standard that does not cause a short circuit. As a result, unlike the conventional example, the copper residue that can be repaired is not erroneously determined as a defect, only the true copper residue defect can be detected, and the burden of reinspecting the printed circuit board is reduced as much as possible. In addition, it contributes to substantially improve the production yield of the printed circuit board.

In the embodiment of the present invention, the inspection target 13
The case of the printed circuit board 23 has been described above, but the present invention is not limited to this.

[0053]

As described above, according to the appearance inspection apparatus of the present invention, it has the detection means for detecting the defect of the inspection object from the image data, and the first and second labeling circuits and the area determination circuit. And a determining means. Therefore, among the defect candidates of the inspection element pattern detected by the detection means, the repairable defect can be determined as “normal” by the area determination circuit, and only the true defect can be detected.

Further, according to the appearance inspection method of the present invention, it is a defect that has occurred in the defect area or other than the defect area based on the correspondence relationship between the first and second labeling processes taking the conductor spacing into consideration. It is determined whether or not the defect has occurred in 1. For this reason, it is possible to separate defects that have occurred in a defective area that fails inspection and defects that have occurred in areas other than the defective area, that is, in normal areas that should pass inspection.
In this way, a defect that can be relieved like the conventional example is a "defect".
There is no need to judge.

This greatly contributes to the provision of a highly reliable appearance inspection device for inspecting the wiring pattern of the printed circuit board and also to the improvement of the production yield of the object to be inspected.

[Brief description of drawings]

FIG. 1 is a principle diagram of an appearance inspection apparatus according to the present invention.

FIG. 2 is a configuration diagram of a printed circuit board inspection device according to an embodiment of the present invention.

FIG. 3 is an internal configuration diagram of a labeling circuit according to an embodiment of the present invention.

FIG. 4 is an explanatory diagram of a pattern search and area setting method according to an embodiment of the present invention.

FIG. 5 is a diagram illustrating a labeling filter according to an embodiment of the present invention and a functional explanatory diagram thereof.

FIG. 6 is an explanatory diagram of relabeling of the inspection element pattern according to the embodiment of the present invention.

FIG. 7 is a functional explanatory diagram of the area determination circuit according to the embodiment of the present invention.

FIG. 8 is a wiring pattern inspection flowchart (part 1) according to the embodiment of the present invention.

FIG. 9 is a wiring pattern inspection flowchart (part 2) according to the embodiment of the present invention.

FIG. 10 is a supplementary explanatory diagram of the labeling process according to the embodiment of the present invention.

FIG. 11 is a configuration diagram of a printed circuit board inspection device according to a conventional example.

FIG. 12 is an inspection flowchart of a wiring pattern according to a conventional example.

FIG. 13 is a printed board and copper residual pattern diagram for explaining the problems in the conventional example.

[Explanation of symbols]

 11 ... Detecting means, 12 ... Judging means, 22A ... First labeling circuit, 22B ... Second labeling circuit, 22C ... Region judging circuit, 22D ... Selector, DIN ... Image data, SOUT ... Comparison result signal, S1, S2 ... label output signal, S3 ... defect re-determination signal.

Claims (5)

[Claims] 1. A detection means for detecting a defect to be inspected from image data, and a determination means for determining whether the defect detected by the detection means is an allowable defect or a true defect. Appearance inspection device. 2. The determining means enlarges the inspection element pattern of the inspection object extracted from the image data and labels the inspection element pattern for each inspection element pattern, and removes defects from the image data. Within a defective area by comparing a label output signal from the first labeling circuit and a label output signal from the second labeling circuit with a second labeling circuit that labels each inspection element pattern to be inspected. 2. An area determination circuit for determining whether or not the defect has occurred in 1.
Appearance inspection device described. 3. An inspection element pattern to be inspected is extracted from image data, the extracted inspection element pattern is enlarged to set an inspection area, and whether a defect generated in the inspection area is a true defect or not. A visual inspection method characterized by determining whether or not. 4. The determination as to whether or not the defect is a true defect is performed by performing a first labeling process of adding a serial number to each of the enlarged inspection element patterns of the inspection object, and excluding the defect. A second labeling process is performed by adding a serial number to each target inspection element pattern, and the defect is a defect that has occurred in the defect region based on the correspondence relationship between the first and second labeling processes, or a defect other than the defect region is detected. The appearance inspection method according to claim 3, wherein it is determined whether or not the defect has occurred. 5. The appearance inspection method according to claim 4, wherein the labeling process uses a filter having a radius of a minimum interval between inspection element patterns.