JP2003141508A - Method, apparatus and program for visual inspection of substrate, and computer readable storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To increase the efficiency of inspection process by equally dividing a plurality of pads for inspection. SOLUTION: Two processing areas Q11 and Q21 for inspection image data are set so that, in inspection image data provided by picking up an image of a substrate 3 to be inspected with a plurality of pads 1 thereon, the inspected areas (number or total area) of the pads 1 are equally divided. For each of the processing areas Q11 and Q21 , an inspection image processing thread is started via the parallel processing of a multiprocessor 22 to visual inspect the substrate 3.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a board appearance inspection method and apparatus for picking up an image of a printed board and detecting defects such as scratches, spots, and chips on the printed board based on the image data, and an appearance inspection for the board. And a computer-readable storage medium storing this program.

[0002]

2. Description of the Related Art In recent years, with the miniaturization of electronic equipment, the density of electronic components such as printed circuit boards has rapidly increased. With respect to such a printed circuit board, defects such as scratches, stains, and chips are detected by the board appearance inspection device.
However, with the increase in the density of electronic components, the board appearance inspection apparatus has a heavy load on the processing of the inspection image data, resulting in an increase in the inspection processing time. Under such circumstances, there is a demand for a board appearance inspection device capable of inspecting with high accuracy and high speed, and for that purpose, it is necessary to efficiently perform image processing on inspection image data imaged at high resolution.

In order to speed up the processing of the inspection image data, a method of distributing the image processing can be considered. For example, as a technology aimed at speeding up product inspection,
For example, there is a visual inspection device described in Japanese Patent Laid-Open No. 6-222012. This appearance inspection apparatus realizes high-speed inspection by executing inspections of multiple items in parallel, and does not speed up inspection of a single item.

As an image processing method for distributing processing by using a multiprocessor, for example, Japanese Patent Laid-Open No. 6-52 is available.
There is a 296 publication. This image processing method is an image division parallel processing in which processing modules are divided, and is not effective when the degree of dependence on the immediately preceding processing result in the image processing process is high.

On the other hand, when the image processing area for the inspection image data is divided and processed in parallel by the multiprocessors, it is desirable that the load on the multiprocessors be evenly distributed. Therefore, the image size of the inspection image data is equalized. It is generally done by dividing into.

[0006]

However, in the method of simply dividing the image size of the inspection image data evenly, the amount of inspection image data is not always evenly distributed. For example, a plurality of pads 1 as shown in FIG.
When a visual inspection is performed on a plurality of pieces 2 of squares arranged in a square, for example, a substrate 3 to be inspected arranged in two rows, inspection image data is acquired in an inspection range 4 including these two rows of pieces. There is a need to.

When acquiring the inspection image data,
As shown in 2, when the width of the image sensor 5 is shorter than the width of the inspection range 4, the image sensor 5 scans the substrate 3 to be inspected twice, that is, the first-row scanning and the second-row scanning. Become.

The inspection image data obtained by dividing the scan into two pieces as described above is obtained, for example, by scanning the first row by the first row piece and the second row piece as shown in FIG. 13 (a). Becomes the inspection image data Da including a part of the above, and becomes the inspection image data Db including the remaining portion of the second row piece as shown in FIG.

To perform inspection by equally dividing the inspection image data Da and Db into respective sizes, the inspection image data Da is divided into image areas W ₁ and W _2, and the inspection image data Da is obtained. Each image area W ₃ and W for Db
_4, and the inspection processing is executed in parallel by the multiprocessor.

However, in the method of simply dividing the image size of the inspection image data into equal parts, the number of pads 1 (inspection image data amount) included in each divided image area becomes extremely unbalanced, and especially in the case of FIG. In the inspection image data Db of the second column scanning shown in 13 (b), the image area W ₄ does not include the pad 1 at all, and even if the parallel processing is executed by the multiprocessor, each image area W 4
_The effect of making the processing load on ₃ and W ₄ equal and shortening the processing time is not exhibited at all.

On the other hand, in the inspection process for the substrate 3 to be inspected, clustering, which is a general image processing method, is used to extract the pad 1 from the inspection image data. Since this clustering method calculates binarized image data obtained by binarizing inspection image data by concatenating from adjacent pixels, the processing amount is extracted as the number of pixels for which the concatenated search is performed is increased. The larger the number of pads, the longer the processing time.

Further, in the inspection processing for scanning the periphery of the extracted pad 1 to detect a defect such as a chip, the time required for the processing is affected by the total amount of contour points of the pad 1.

Therefore, in such an inspection process, when the image size of the inspection image data is equally divided and processed in parallel, the inspection image data D shown in FIG.
In the state where the number of pads 1 is disproportionately divided as shown in b, the processing time on one side of the image region W ₃ where the pad 1 is included is long, and the processing on the other side of the image region W 4 ends quickly. Extra latency will occur.

Therefore, it is an object of the present invention to provide a board appearance inspection method and apparatus capable of efficiently dividing inspection image data of a plurality of inspection objects to improve the efficiency of inspection processing.

Further, according to the present invention, a board appearance inspection program for evenly dividing the inspection image data amount of a plurality of inspection objects to improve the efficiency of the inspection process, and a computer that stores this program can read the program. The purpose of the present invention is to provide a flexible storage medium.

[0016]

The present invention provides a board appearance inspection method for picking up an image of a board to be inspected on which a plurality of objects to be inspected are arranged and performing an appearance inspection of the board to be inspected based on the image data. The processing area for the image data is divided into a plurality of sections so that the inspection areas of the inspection object in the data are evenly divided, and image processing is performed for each of these processing areas to perform a visual inspection of the substrate to be inspected. This is a method for inspecting the appearance of a substrate.

According to the present invention, in a board appearance inspecting apparatus for picking up an image of a board to be inspected on which a plurality of objects to be inspected are arranged and inspecting the appearance of the board to be inspected based on the image data, the object to be inspected in the image data. Appearance of the substrate to be inspected by performing image processing on each of the plurality of processing areas divided by the area dividing means so as to divide the processing area for the image data into a plurality of areas so that each inspection area is equally divided. An inspection apparatus for inspecting a substrate, comprising: inspection processing means for inspecting.

According to the present invention, the image data obtained by picking up an image of the substrate to be inspected on which a plurality of inspection objects are arranged is taken in, and the inspection area of the inspection object in this image data is equally divided. A board appearance inspection program is characterized in that a processing area for image data is divided into a plurality of portions, and image processing is performed on each of the divided plurality of processing areas to perform an appearance inspection of a substrate to be inspected.

[0019]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of a substrate visual inspection apparatus.
On the transfer stage 10, as shown in FIG. 11, a plurality of pieces each having a plurality of pads 1 arranged in a square shape are arranged, for example, a substrate 3 to be inspected is arranged in two rows. The transfer stage 10 constitutes a transfer device 11 as shown in FIG. This transport device 11 is
The cassette 12 that stores the uninspected substrate 3 to be inspected, and the uninspected substrate 3 that is stored in the cassette 12 are held and transported above the transport stage 10 to transfer the uninspected substrate 3 to the uninspected substrate 3. Loader 13 mounted on the transfer stage 10
Then, the non-defective product cassette 14 and the defective product cassette 15 and the inspected substrate 3 that has been inspected are ejected from the transport stage 10,
An unloader 16 that stores the inspected substrate 3 that has been inspected in the good product cassette 14 or the defective product cassette 15 according to the determination result of the substrate inspection is provided.

The image pickup device 17 is provided above the transfer stage 10. This image pickup device 17 includes
An illumination light generating means (light source) 18 for illuminating
The illumination light generated by the illumination light generating means 18 is applied to the entire surface of the substrate 3 to be inspected through an illumination optical system 19 such as an optical fiber. Further, in the image pickup device 17, the image sensor 5 is provided via the image forming optical system 20 in a direction almost directly above the substrate 3 to be inspected. The image sensor 5 includes, for example, a plurality of CCD elements. The image signal output from the image pickup device 17 is sent to the image processing device 21.

The image processing device 21 takes in an image signal output from the image pickup device 17 and stores it as inspection image data. The image processing device 21 sends an image input device 24 and a plurality of image data to the multiprocessor 22 via the system bus 23. The image memories 25-1 to 25-N are connected. The image input device 24 takes in the image signal output from the image pickup device 17 and temporarily stores it as inspection image data. The image memories 25-1 to 25-N are
An arbitrary image size of inspection image data, the number of inspection image data to be stored, and the like can be set.

The image processing device 21 is an image input device 2
A process of transferring the inspection image data temporarily stored in 4 to any one of the plurality of image memories 25-1 to 25-N by DMA (direct memory access) transfer without interposing the multiprocessor 22; Image processing such as board inspection by 22 is asynchronously executed in parallel.

Further, the image processing device 21 has the area dividing means 2 for performing the temporary inspection mode prior to the actual inspection operation.
It has 6 functions. The area dividing means 26 stores the inspection image data of one inspected substrate 3 in the image memory 25-1.
25-N, the processing areas for the inspection image data are divided into a plurality of processing areas, for example, two processing areas so that the inspection areas (quantity or total area) of the pads 1 in the inspection image data are equally divided. It has a function of creating an area table 27 including information of the division result.

Specifically, the area dividing means 26 is caused by the area dividing algorithm executing means 28 to detect the pad 1 from the inspection image data to the pixel position (X pixel) as shown in FIG.
In order to create a histogram showing the frequency of the pixel value “1” corresponding to the above, and to calculate two processing areas in which the numbers of the pads 1 are equally divided from this histogram, the image data start pixel number P1 and the pixel value “1” are calculated. The pixel number P2 and the image data end pixel number P3 for which the cumulative total of the frequencies of "1" is 1/2 of the total number are calculated, and these pixel numbers P1 and P
2, P3 has a function of creating an area table 27 as shown in FIG.

The area table 27 has image memory numbers (25-1 to 25-N) assigned to the plurality of image memories 25-1 to 25-N, an offset "1", and a width "1". , Offset “2” and width “2”.

[0027] Among the offset "1", which corresponds to the image data start pixel number P1 shown in FIG. 3, the image memory number 25-1 For example, the image data start pixel number P1 is recorded as S1 ₁ .

The width "1" is obtained by subtracting the image data start pixel number P1 from the pixel number P2, and in the image memory number 25-1, P2-P1 = W1.
Recorded as ₁ .

The offset "2" corresponds to the pixel number P2, and in the image memory number 25-1,
The pixel number P1 is recorded as S2 ₁ .

The width "2" is obtained by subtracting the pixel number P2 from the image data end pixel number P3. In the image memory number 25-1, P3-P2 = W2 ₁
Is recorded as.

Further, the image processing device 21 is
Inspection processing means 2 for automatically inspecting the substrate 3 to be inspected
It has 9 functions. The inspection processing means 29 reads the information of the offset “1”, the width “1”, the offset “2” and the width “2” from the area table 27 created by the area dividing means 26, and based on these information, the image memory. The inspection image data stored in 25-1 to 25-N is divided into, for example, two processing use areas, and the image processing is performed for each of these processing areas to perform a visual inspection of the substrate 3 to be inspected. There is.

The inspection processing means 29 performs the DMA transfer without the interposition of the multiprocessor 22 as described above to capture the inspection image data obtained by imaging the substrate 2 to be inspected and the inspection image thus captured. The process of performing a visual inspection of the inspected substrate 2 from the data is performed in parallel.

The image processing device 21 captures the inspection image data obtained by imaging the substrate 3 to be inspected on which the plurality of pads 1 are arranged, and stores the image data in the image memories 25-1 to 25-.
In N, the area table 27 for dividing the processing area for the inspection image data into a plurality of areas is created so that the number of pads 1 in the inspection image data (inspection image data quantity) is equally divided. A program that can be read by the multiprocessor 22 that stores a board appearance inspection program for performing an image inspection of each of the plurality of processing areas based on the information recorded in the area table 27 and performing an appearance inspection of the board 3 to be inspected. The memory 30 is connected to the system bus 23.

This board appearance inspection program is made to generate a histogram showing the frequency of the pixel value "1" corresponding to the pad 1 with respect to the pixel position (X pixel) from the inspection image data, and the inspection area of the pad 1 is respectively generated from this histogram. In order to calculate two processing areas that are evenly divided,
The image data start pixel number P1, the pixel number P2 at which the cumulative frequency of pixel values “1” is ½ of the total number, and the image data end pixel number P3 are determined, and these pixel numbers P
The program of the area division algorithm for creating the area table 27 from 1, P2 and P3 is included.

Next, the operation of the apparatus configured as described above will be described.

First, the inspection operation of the entire visual inspection of the substrate 3 to be inspected will be described with reference to the flowchart of the inspection procedure shown in FIG.

In step # 1, the image processing apparatus 21 is input with board shape information such as the length and width of the board 3 to be inspected. As a result, in the image processing device 21, the scanning length and the number of times of scanning by the transport stage 10 necessary to image the entire substrate 3 to be inspected are determined.

Next, the image processing apparatus 21 proceeds to step # 2.
In FIG. 6, the image signal output from the image pickup device 17 is captured and the inspection image data of the entire substrate 3 to be inspected is stored in any of the image memories 25-1 to 25-N, and the inspection image data is shown in FIG. The inspection range 4 including such a piece to be inspected is designated.

Next, in the image processing device 21, step #
In 3, the binarization threshold value for extracting the pad 1 from the inspection image data, the area range recognized as the pad 1, and other inspection parameters are set.

Next, the image processing apparatus 21 proceeds to step # 4.
In order to adjust the detection result obtained by the above-mentioned set inspection parameters, the temporary inspection is executed using one inspected substrate 3 prior to the actual inspection operation.

This temporary inspection will be described with reference to the temporary inspection flowchart shown in FIG.

The area dividing means 26 of the image processing device 21 is
In step # 10, the image memory 25 for N sheets necessary for inputting the image data of the inspection range 4 shown in FIG.
-1 to 25-N are secured, and in the next step # 11, n = 1 is set to input the inspection image data of the entire first column of the substrate 3 to be inspected, and in the next step # 12. , The imaging device 1 that has imaged the first row of the board 3 to be inspected
Image memories 25-1 and 25-2 that capture the image signal from
Store in any of 5-N.

Here, the image pickup device 17 has the above-mentioned inspection range 4
In order to acquire the image data of, the imaging of the substrate 3 to be inspected is performed as follows. If the image pickup resolution of the image sensor 5 in the image pickup device 17 is R and the number of effective pixels of the image sensor 5 is M, the image pickup range Wa of the image sensor 5 is Wa = R × M (1) as shown in FIG. Becomes

Here, when the width of the inspection range is Wp, and there is a relationship of Wa <Wp ≦ Wa × 2 (2) between the imaging range Wa and the inspection range width Wp, the imaging device 17 In order to obtain the inspection range 4, the substrate 3 to be inspected is scanned twice (first-row scanning and second-column scanning) to image the substrate 3 to be inspected. The ratio between the imaging range Wa and the inspection range width Wp differs depending on the type of substrate size of the inspected substrate 3.

FIGS. 9 (a) and 9 (b) are schematic diagrams of inspection image data captured by the first-row scanning and the second-row scanning.
(a) is the inspection image data Da acquired by scanning the first column
FIG. 3B shows the inspection image data Db captured by scanning the second column. The inspection image data Da is image data including a part of the pieces in the first row and the pieces in the second row,
The inspection image data Db is image data including the rest of the pieces in the second row.

Next, the area dividing means 26 performs step # 1.
3, the inspection image data of the first-row scan stored in any of the image memories 25-1 to 25-N is read out, and binarization processing is performed on the inspection image data using a preset threshold value. The pad 1 and its background information are separated.
At this time, the pad 1 in the binarized inspection image data
Has a pixel value of "1", and the background information has a pixel value of "0".

Next, the area dividing algorithm executing means 28 of the area dividing means 26, in step # 14, executes the processing shown in FIG.
As shown in (1), a histogram showing the frequency of the pixel value “1” corresponding to the pad 1 with respect to the pixel position (X-direction pixel column) is created from the inspection image data of the first-row scanning.

Next, the area division algorithm execution means 28
In order to calculate two processing areas in which the inspection area (quantity or total area) of the pad 1 is equally divided from the histogram, first, in step # 15, the image data start pixel number P1 of the histogram is calculated. Then, in the next step # 16, the pixel number P2 at which the cumulative frequency of the pixel value “1” in the histogram is ½ of the total number is calculated, and in the next step # 17, the image data end pixel number P3 in the histogram is calculated. calculate.

Next, the area division algorithm executing means 28
Is the image data start pixel number P in the calculated inspection image data for the first column scan in step # 18.
1, the pixel number P2 and the image data end pixel number P3 are recorded in the area table 27 as shown in FIG.

In the area table 27, the image data start pixel number P1 is set to the start point S1 in the column of offset "1" for easy understanding in comparison with FIG. 3 and FIG. 9 (a).
Is recorded as _1, P2-P1 = W obtained by subtracting the image data start pixel number P1 from the pixel number P2 column width "1"
1 ₁ is recorded, and the pixel number P1 is recorded in the column of offset “2”.
P that but recorded as the starting point S2 _1, minus the pixel number P2 from the image data end pixel number P3 in the column width "2"
3-P2 = W2 ₁ are recorded.

Therefore, as shown in FIG. 9A, the inspection image data Da of the first-row scan is divided into two processing regions of width W1 ₁ and width W2 _{1 in} which the numbers of the pads 1 are evenly distributed. Will be done.

The offset "1", the width "1", the offset "2" and the width "2" recorded in the area table 27 are for the inspection image data Da of the first row scanning and are for the two rows. At the time of scanning, the offset "1", the width "1", the offset "2", and the width "2" are recorded in the same manner as above. FIG. 9B shows the inspection image data Db of the second row scan.
It is a schematic diagram of each starting point F1 ₁ , F2 ₁ and each width E1
_2, E2 ₂ is shown.

In the above-described temporary inspection operation, since the input of the inspection image data of the inspection range 4 shown in FIG. 6 is completed, that is, the inspection image data is stored in the N image memories 25-1 to 25-N. It is repeated until the storage is completed (N <n).

Next, the image processing apparatus 21 proceeds to step # 5 shown in FIG. 5 to automatically inspect the substrate 3 to be inspected according to the automatic inspection flowchart shown in FIG. In step # 30, the inspection processing unit 29 of the image processing apparatus 21 includes the image memories 25-1 to 25-N of the number of sheets required to capture the length of the inspection range 4 by the first column scanning shown in FIG.
Secure. For example, when the image sensor 5 is a line sensor type, the size of one image memory is such that the horizontal width is the number of effective pixels of the line sensor 5 and the vertical width is the number of lines serving as one reading unit.

In actual reading, double buffering processing is further subdivided by a memory in the image input device 24 for temporarily storing image data.
In this double buffering process, there are two buffers, and while the data is input to one buffer, the other buffer passes through the system bus 23 to the image memory 2
5-1 to 25-N are DMA-transferred. When the data transfer to the image memories 25-1 to 25-N for the number of lines of the reading unit is completed, the next image memories 25-1 to 25-1 to 25-N are completed.
The transfer destination changes to 25-N.

Next, the inspection processing means 29 determines in step # 3.
1, n = 1 is set in order to confirm the first substrate 3 to be inspected which is sequentially transported by the loader 13 and placed on the transport stage 10, and in the next step # 32,
From the area table 27 shown in FIG. 4, the starting point S1 ₁ of the column of the offset “1” and the width W1 ₁ of the column of the width “1” for the inspection image data of the first column scanning are read out, and the next step #
In 33, the inspection image data in the first column scan sets a processing area Q1 ₁ to the image memory 25-1 to 25-N that are stored on the basis of the these start point S1 ₁ and width W1 _1.

Next, the inspection processing means 29 performs step # 3.
4, the starting point S2 ₁ of the column of offset “2” and the width W2 ₁ of the column of width “2” from the area table 27 shown in FIG.
Is read out, and in the next step # 35, the processing area is stored in the image memories 25-1 to 25-N in which the inspection image data of the first-row scanning is stored based on the starting point S2 ₁ and the width W2 _1. setting the Q1 _2.

Next, the inspection processing means 29 performs step # 3.
6 and # 37, the setting of the processing areas Q1 _N for the number of image memories 25-1 to 25-N necessary for the first column scanning is repeated.

By the above operation, the image memories 25-1 to 25-2
As shown in FIG. 9, two processing areas Q1 ₁ and Q2 ₁ are set for the entire inspection image data of the first column scanning for 5-N.

Next, the inspection processing means 29 determines in step # 3.
8, the above n = 1 is set, and in the next step # 39, the image input command (first time) is input to the image memories 25-1 to 25-N in which the inspection image data of the first column scanning is initially stored. ) Is issued and the inspection image data of the first column scanning is stored next in step # 40.
An image input command (second time) is issued to 5-1 to 25-N.

At this time, since the inspection image data is taken in asynchronously with the process of inspecting the appearance of the substrate 2 to be inspected from the inspection image data taken in as described above, the multiprocessor 22 issues each image input command. You can immediately proceed to the next step and execute the process.

The inspection image data fetched by the image input device 24 by issuing the second image input command is placed in the execution waiting state of the appearance inspection process, during which the inspection image data fetched first is By the DMA transfer, the data is transferred to the image memories 25-1 to 25-N without interposing the multiprocessor 22.

Next, the image processing apparatus 21 proceeds to step # 4.
In step 1, it is confirmed that the image input of the inspection image data first captured is completed. Actually, since the image input device 24 does not allow a plurality of image pickup execution waits, a synchronization process is performed inside the image input device 24, and the time of the multiprocessor 22 is not consumed for confirmation of completion here.

Next, the inspection processing means 29 causes the step # 4.
In 2, the above step # Start inspection image processing threads for processing region Q1 ₁ shown in FIG. 9, which is set (a) at 33, followed by step # in 43 is set at step # 35 9 (a) Shown processing area Q
The inspection image processing thread for 2 ₁ is started. These inspection image processing threads are the minimum units that can be assigned to the multiprocessor 22, and the activated inspection image processing threads are executed separately from the execution of the automatic inspection flowchart shown in FIG.

By executing this inspection image processing thread, the inspection of the substrate 3 to be inspected is carried out for defects such as scratches, stains and chips.

After starting the inspection image processing threads for these processing areas Q1 ₁ and Q2 ₁ , the inspection processing means 29
Confirms in step # 44 and # 45 whether or not the scanning for the first column is completed. If the scanning for the first column is not completed, the process returns to step # 40 and the third image input command is issued. It is issued to the image memories 25-1 to 25-N.

When the third image input command is issued and the image pickup input corresponding to the image input command is in the waiting state,
The inspection image data for the second image input command is D
The MA is transferred and stored in the image memories 25-1 to 25-N. Then, the inspection processing means 29 activates the inspection image processing thread for the processing area in the inspection image data.

After this, until the board inspection processing for the inspection image data for the first column scanning is completed, the above step #
40 to # 45 are repeated.

When the board inspection processing for the inspection image data for the first column scanning is completed, the above steps # 40 to # 45 are repeatedly executed for the inspection image data for the second row scanning to perform the board inspection processing. Is done. For example, the inspection processing means 29, in step # 42, carries out the process shown in FIG.
Start inspection image processing threads for processing region Q1 ₁ shown in (b), followed by starting the inspection image processing threads for processing region Q2 ₁ in step # 43, the second column scanning operation by the execution of these inspection image processing threads Defect inspections such as scratches, spots, and chips on the substrate 3 to be inspected are executed from the inspection image data of 1.

When a series of automatic inspection processing for one substrate 3 to be inspected is completed as described above, the transport device 11 is notified of the determination result of the inspection. The unloader 16 of the transfer device 11 ejects the inspected substrate 3 that has been inspected from the transfer stage 10 and stores the inspected substrate 3 that has been inspected in the non-defective product cassette 14 or the defective product cassette 15 according to the determination result of the substrate inspection. Sort and store.

At the same time, the loader 13 takes out the uninspected substrate 3 to be inspected next from the cassette 12, holds it, and conveys it to the upper side of the conveying stage 10 to convey the uninspected substrate 3 to the conveying stage 10. Place on top. As a result, the inspection of the substrate 3 to be inspected is executed in the same manner as above.

As described above, in the above embodiment,
2 with respect to the inspection image data so that the number of pads 1 in the inspection image data acquired by imaging the substrate 3 to be inspected on which the plurality of pads 1 are arranged is equally divided.
Two processing areas Q1 ₁ and Q2 ₁ are set, and the inspection image processing threads are activated by the parallel processing of the multiprocessor 22 for each of these processing areas Q1 ₁ and Q2 ₁ , and the appearance inspection of the substrate 3 to be inspected is performed. Therefore, the plurality of pads 1 can be evenly distributed in each of the two processing areas Q1 ₁ and Q2 ₁ , the processing speed can be optimized, the inspection processing can be executed at high speed, and the image processing for the inspection image data can be performed. Can be minimized and the efficiency of the inspection process can be improved.

A plurality of pads 1 are connected to two processing areas Q1 ₁ ,
In order to uniformly disperse each Q2 _1, a histogram showing the frequency of pixel values corresponds to the pad 1 "1" to the pixel position from the inspection image data, so that the quantity of the pad 1 from the histogram is equally divided respectively Since two processing areas Q1 ₁ and Q2 ₁ are determined in
When the inspection target substrate 3 is imaged by performing column scanning and second column scanning on the inspection target substrate 3 based on the relationship between the imaging range Wa of the image sensor 5 and the inspection range width Wp for the inspection target substrate 3, FIG. Even if the images have a plurality of pads 1 disproportionately distributed like the inspection image data Da and Db of the first-row scanning and the second-row scanning shown in 9 (a) and (b), these inspection image data Da , Db, it is possible to reliably obtain the respective processing regions Q1 ₁ , Q2 ₁ , G1 ₁ , G2 _{1 in} which a plurality of pads 1 are balancedly distributed.

In this case, even when an automatic inspection is performed on a plurality of types of substrates 3 to be inspected having different substrate sizes, the processing areas Q1 ₁ and Q2 corresponding to the substrate sizes are obtained.
₁ , G1 ₁ , G2 ₁ can be set.

The inspection image is provided with the area dividing means 28 for dividing the two processing areas Q1 ₁ and Q2 ₁ so that the inspection area (quantity or total area) of the pad 1 to be inspected is equally divided. If the processing time for data can be made efficient, the image sensor 5 having a high operating frequency can be used, and the inspection time of the entire board appearance inspection device when performing automatic inspection operation can be shortened.

Further, since the process of taking in the inspection image data obtained by picking up the image of the substrate 3 to be inspected and the inspection process of performing the appearance inspection of the inspected substrate 3 from the taken in inspection image data are processed in parallel, The inspection processing for the inspection substrate 3 can be speeded up.

As a result, the increase in the density of electronic parts such as a printed circuit board rapidly progresses with the downsizing of electronic equipment, and the increase in the inspection time is minimized even when the processing amount per unit of the printed circuit board or the like increases. You can keep it to the limit.

The present invention is not limited to the above-mentioned one embodiment, but can be variously modified at the stage of implementation without departing from the spirit of the invention.

Furthermore, the above-described embodiments include inventions at various stages, and various inventions can be extracted by appropriately combining a plurality of disclosed constituent elements. For example, even if some constituent elements are deleted from all the constituent elements shown in the embodiment, the problem described in the section of the problem to be solved by the invention can be solved, and it is described in the section of the effect of the invention. When the effect of being obtained is obtained, a configuration in which this constituent element is deleted can be extracted as an invention.

[0080]

As described above in detail, according to the present invention, a method and apparatus for inspecting a substrate, which can divide the inspection image data amount of a plurality of inspection objects evenly to improve the efficiency of the inspection process. Can be provided.

Further, according to the present invention, a board appearance inspection program for evenly dividing the inspection image data amount of a plurality of inspection objects to improve the efficiency of the inspection process, and a computer storing the program. A readable storage medium can be provided.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing an embodiment of a substrate visual inspection apparatus according to the present invention.

FIG. 2 is a configuration diagram of a transfer device and an image pickup device in an embodiment of a substrate appearance inspection device according to the present invention.

FIG. 3 is a diagram showing a histogram created by an area dividing unit in one embodiment of the substrate visual inspection apparatus according to the present invention.

FIG. 4 is a schematic diagram showing an area table in one embodiment of the substrate visual inspection apparatus according to the present invention.

FIG. 5 is a flowchart of the inspection procedure in the embodiment of the board appearance inspection apparatus according to the present invention.

FIG. 6 is a diagram showing a piece frame in one embodiment of the substrate visual inspection apparatus according to the present invention.

FIG. 7 is a temporary inspection flowchart in one embodiment of the board appearance inspection apparatus according to the present invention.

FIG. 8 is a schematic diagram showing the relationship between the imaging range of the imaging device and the width of the board to be inspected in one embodiment of the board visual inspection apparatus according to the present invention.

FIG. 9 is a schematic diagram of each inspection image data captured by the first-row scanning and the second-row scanning in the embodiment of the board visual inspection apparatus according to the present invention.

FIG. 10 is an automatic inspection flowchart in one embodiment of the board appearance inspection apparatus according to the present invention.

FIG. 11 is an external view of a substrate to be inspected in which a plurality of pieces each including a plurality of pads are arranged.

FIG. 12 is a diagram showing the relationship between the width of the image sensor and the width of the inspection range.

FIG. 13 is a schematic diagram of inspection image data acquired by an image sensor.

[Explanation of symbols]

1: Pad 2: Peace 3: Inspected board 4: Inspection range 5: Image sensor 10: Transport stage 11: Transport device 12: cassette 13: loader 14: Good quality cassette 15: Defective cassette 16: Unloader 17: Imaging device 18: Illumination light generating means (light source) 19: Illumination optical system 20: Imaging optical system 21: Image processing device 22: Multiprocessor 23: System bus 24: Image input device 25-1 to 25-N: Image memory 26: Area dividing means 27: Area table 28: Area division algorithm execution means 29: Inspection processing means 30: Program memory

Continued front page    F term (reference) 2F065 AA49 BB02 CC01 DD06 FF04                       HH12 JJ03 JJ26 LL02 PP11                       QQ31 QQ43                 2G051 AA65 AB02 CA04 CB01 CC09                       DA05 EA14 EA19 EB01 EC02                       ED22                 5B057 AA03 BA11 CA08 CA12 CA16                       CB08 CB12 CB16 CE09 CE12                       CH04 DB02 DB08 DC04 DC19

Claims (6)

[Claims] 1. A board appearance inspection method for picking up an image of a board to be inspected on which a plurality of objects to be inspected are arranged, and inspecting the appearance of the board to be inspected based on the image data. The board appearance, characterized in that the processing area for the image data is divided into a plurality of parts so that each of the inspection areas is evenly divided, and image processing is performed for each of these processing areas to perform a visual inspection of the inspected board. Inspection method. 2. A board appearance inspecting apparatus for picking up an image of a board to be inspected, on which a plurality of objects to be inspected are arranged, and inspecting the appearance of the board to be inspected based on the image data. Area dividing means for dividing the processing area for the image data into a plurality of areas so that each inspection area is equally divided, and image processing is performed for each of the plurality of processing areas divided by the area dividing means. A board appearance inspection apparatus comprising: an inspection processing unit that performs an appearance inspection of an inspection board. 3. The area dividing means creates a histogram showing the frequency of pixel values corresponding to the inspection object with respect to pixel positions from the image data, and the inspection areas of the inspection object are evenly divided from the histogram. The substrate appearance inspection apparatus according to claim 2, wherein a plurality of the processing regions are indexed. 4. The inspection processing means performs parallel processing of a process of capturing image data obtained by capturing an image of a substrate to be inspected and a process of performing a visual inspection of the substrate to be inspected from the captured image data. The board appearance inspection apparatus according to claim 2, having a function. 5. The image data obtained by picking up an image of a substrate to be inspected on which a plurality of inspection objects are arranged is captured, and the inspection area of the inspection object in the image data is divided evenly. A board appearance inspection program, wherein a processing area for image data is divided into a plurality of portions, and image processing is performed on each of the divided plurality of processing areas to perform an appearance inspection of the substrate to be inspected. 6. The image data obtained by picking up an image of a substrate to be inspected on which a plurality of inspection objects are arranged is taken in, and the inspection area of the inspection object in the image data is divided evenly. A computer characterized in that a processing area for image data is divided into a plurality of areas, and a board appearance inspection program for performing an appearance inspection of the board to be inspected by performing image processing on each of the plurality of divided processing areas is stored. Readable storage medium.