JP2720678B2 - Lead flatness measurement method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for measuring lead flatness of a surface mount type integrated circuit package (hereinafter referred to as SMD), and more particularly to a binary search algorithm having an image processing technique and a backtrack function. The present invention relates to a method for measuring the flatness of a lead using the above calculation means.

[0002]

2. Description of the Related Art Conventionally, when measuring the flatness of a lead of an S / M / D4, a side surface of the S / M / D4 is scanned with a two-dimensional CCD camera 5.
As shown in FIG. 7, the system CPU 11b composed of the horizontal scanning unit 13, the vertical profile counter 21, and the flatness calculating unit 22 processes and outputs the image obtained by the imaging in accordance with the algorithm shown in FIG.

First, the horizontal scanning unit 13 assumes the vicinity of the bottom of the S, M and D 4 in the binarization memory 9 and moves from left to right in accordance with a read scan line 23 provided in advance by the CPU 14 from left to right. Horizontal scanning while distinguishing between light and dark. From the pixel that has changed from a bright point to a dark point, a continuous section until the dark point changes to a bright point again is recognized as a lead. Here, in a section that the horizontal scanning unit 13 has recognized as a read, the vertical profile counter unit 21
All the pixels in the same row below the lead scan line 23 are vertically scanned, and the number of all bright points in the scanning process (dotted line) is counted as shown in FIG. 9, and the counting result and the lead which is the starting point of the vertical scanning are read. The pixels on the scan line 23 are converted into coordinates in the binarization memory 9 and
The flatness calculating unit 22 determines the minimum value of the number of consecutive bright spots, that is, the floating amount of the lead, based on the numerical data stored in the data cell. Flatness calculator 2
2 retrieves the maximum value from all the data stored in the memory unit 12 and sets it as the read flatness of the SMD4. The result is transferred to the CPU 14, and the CPU 14 multiplies all the data transferred from the flatness calculating unit 22 by a pre-stored resolution of the two-dimensional CCD camera 5 and converts the converted real value to a predetermined value. A pass / fail judgment (passed within the standard) was performed for the stored standard of the lead flatness, and the result was displayed on the CRT 15 via the I / O 10.

[0004]

According to the conventional image processing means, when obtaining the lead flatness from a binarized image, FIGS.
If the lead is extremely deformed as shown in FIG. C (FIGS. 9A and 9C or a short FIG. 9B), the presence of the lead is recognized by the horizontal lead scan line 23 for the lead A and C that are inclined. And the vertical profile counter 21
Is larger than the actual lead floating amount because the presence or absence of a lead is searched for in the section where the lead is recognized by the horizontal lead scan line 23, and the number of all bright spots from the lead side surface to the reference surface 26 is counted. There is a problem. (See part 25 in FIG. 9) The short lead B cannot be recognized by the horizontal lead scan line 23.

[0005]

According to the present invention, as an image processing means, a reference plane 26 of a screen obtained by imaging the side surfaces of S, M and D 4 by a CCD camera 5 or 20 is arranged in units of rows in the horizontal direction and units of one pixel in the vertical direction. And means for recognizing the read by dividing the data (horizontal parallel lines in FIG. 216), means for converting the recognized read portion into numerical data, and a binary search algorithm having a backtrack function in the data operation processing means. ing.

[0006]

Next, the present invention will be described with reference to the drawings.

FIG. 1 is a system configuration diagram of an SM lead flatness measuring apparatus embodying the lead flatness measuring method of the present invention, and FIGS. 2 and 3 show a side view of the SM lead D4 in one dimension. CC
An image is taken by the D camera 20, the binarized memory 9 and the system C
CRT15 of the binarized image binarized by the PU 11a
FIG. 4 is a tree display diagram of a measurement data processing process processed by the backtrack function, and FIG.
5 is a flowchart illustrating a processing procedure of FIG. 4.

The S, M and D 4 to be measured are positioned on the milky white resin projection 3 provided on the measurement stage 2 along the side surface of the mold section and the inner surface of the lead, and emit light emitted from the light source 1 from below the measurement stage 2. By irradiating the light after diffusing it at the milky white resin projection 3, the lead shadow image of the SMD 4 is projected on the one-dimensional CCD camera 20. The one-dimensional CCD camera 20 takes in the projected image and converts it into a video signal. The video signal is amplified by the AMP 6, digitized by the A / D 7, temporarily stored in the frame memory 8,
The binarized data is binarized to [0, 1] (0; light point, 1; dark point) in comparison with the binarization level stored in the CPU 14 and stored in the binarization memory 9. One-dimensional CCD camera 20a,
20b, 20c, 20d, 20e, 20f, 20g, 2
0h, eight pulse motor controllers 24a, 24
4b, 24c, 24d, 24e, 24f, 24g, 24
The read image is successively accumulated in the binarization memory 9 while stepping up by an amount corresponding to the existing range of the read by h. The horizontal scanning unit 13 steps up the accumulated binarized data of the lead image by one pixel unit perpendicularly to the reference plane 26, and the horizontal scanning unit 13 determines the presence or absence of a read by [0, 1] determination. Then, the total number of bright spots, that is, the floating amount from the 0 judgment, that is, the bright spot portion, to the 1 judgment, that is, the presence of the lead (lead end) is counted. The CPU 14 calculates the value of the floating lead by multiplying the count result of the horizontal scanning unit 13 by the resolution of the one-dimensional CCD camera 20 stored in advance.

FIG. 3 is a diagram for explaining the backtrack function, and FIG. 4 is a tree diagram showing a result of the processing of the backtrack function. FIG. 5 is a flowchart of the backtrack function when obtaining the result of FIG.

According to the image processing method of the lead image described above,
First, the minimum lead floating amount 15 in the binarized screen is detected. The CPU 14 stores the number of horizontal steps from the reference surface 26 when the 15 is detected at the center address of the memory unit 12 and searches the left side of the same screen where the lead has been processed with reference to the lead 15 which is the minimum floating amount. . The smallest floating amount lead 11 in the search is detected, and the number of horizontal steps at this time is stored in the memory unit 12 at the left address of the data of 15. Next, the left side of the lead 11 detected here is searched. Since there is no lead on the left side of the screen of the minimum floating amount lead 11 as a reference, the lead on the right side of the screen is searched. Then the smallest 13 in the search
Is detected. The floating amount of the 13 lead is stored in the 11 right address. Next, when the left side is searched with reference to 13, 12 is detected and stored at the left address of 13. Even if the left and right sides of the 12 leads are searched, there are no leads, and there are no unmeasured leads on the left side with reference to 13. Then, when the right side is searched, 14 is detected and stored in the 13-lead right address. Since there is no lead on the right or left from 14, the process returns to 15 which was initially the reference lead. Here, the right side is searched based on 15. Thereafter, similarly, based on the lead determined to be the minimum floating amount during the search, S, M, D
Search on the left side of the binarization screen of 4. When the lead with the smallest floating amount is detected during the search, the left side of the screen is searched based on the lead, and the right side is searched when there are no unmeasured leads. Hereinafter, FIG. 4 is a diagram showing the result of the processing by repeating the processing. This image processing is performed by the above-described processing means regardless of which lead has what floating amount.
The lead No. of M · D4 matches the order counted from the left side of the captured lead image. The pass / fail judgment of the device with respect to the standard (passing within the standard) is determined by the lead flatness. The read flatness is arranged in ascending order of the read No. by the above processing, and the maximum value is searched by the CPU 14 from the stored data in the memory unit 12, and the maximum value is found by the CPU 14. The result of the pass / fail is compared with the pass / fail judgment standard stored in advance, and the result of the pass / fail
To be displayed.

In FIG. 6, a one-dimensional CCD camera, which is the optical system of the above-described embodiment, is replaced by a two-dimensional CCD camera 5a, 5b, 5c, 5d, 5e, 5f, a system embodying the present invention.
This is an example using a total of 8 units of 5g and 5h. The binarized image of the lead of S, M and D 4 captured by the two-dimensional CCD camera 5 is processed at the binarization level stored in the CPU 14,
It is displayed on the CRT 15 together with the pass / fail judgment.

[0012] All the steps of the apparatus system such as the measurement positioning step and the binarized image processing method are the same as those in the above-described embodiment, and thus are omitted. As a result,
By using the two-dimensional camera in the imaging process, the pulse motor controller is not required, the step error of the pulse motor is not included, and the output time of the determination result is further reduced.

[0013]

As described above, the present invention uses the image processing means for scanning in units of rows in the horizontal direction and in units of one pixel in the vertical direction and a binary search algorithm having a backtrack function. In the processing means, the extreme lead deformation that was erroneously recognized (FIG. 2A,
B, C) can be used to solve the erroneous recognition in the floating amount, and the data obtained thereby can be used to calculate the floating amount of the lead of the surface mount type integrated circuit package while performing backtracking, and to increase the lead numbers in ascending order. Can be arranged.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a lead flatness measuring device.

FIG. 2 is a binarized image and a pixel division diagram obtained by a two-dimensional CCD camera.

FIG. 3 is a binarized image diagram used for explaining a binary search algorithm as a data processing means.

FIG. 4 is a processing result diagram obtained by a backtrack function as a binary search algorithm.

FIG. 5 is a flowchart of a backtrack function.

FIG. 6 is a configuration diagram of a lead flatness measuring apparatus according to a second embodiment.

FIG. 7 is a configuration diagram of a conventional lead flatness measuring device.

FIG. 8 is a flowchart of a conventional lead floating amount.

FIG. 9 is a diagram showing a state that has been a conventional problem.

[Explanation of symbols]

 Reference Signs List 1 light source 2 measurement stage 3 measurement stage protrusion 4 SMD 5a, b, c, d two-dimensional CCD camera 6 AMP 7 analog / digital converter 8 frame memory 9 binarization memory 10 I / O 11a, b system CPU 12 memory 13 horizontal Read scan line 14 CPU 15 CRT 16 Horizontal parallel line of pixel processing 17 Reflector 18 Lens 19 Positioning block 20a, b, c, d One-dimensional CCD camera 21 Vertical profile counter 22 Maximum value calculation unit 23 Horizontal line 24a, b, c , D Pulse motor controller 25 Misrecognition part in previous image processing means 26 Reference plane

Claims (1)

(57) [Claims] 1. A means for taking an image of a side surface of a surface-mounted integrated circuit package and binarizing the obtained image data, and observing the binary data in a lead flatness measurement of the surface-mounted integrated circuit package. Means for recognizing the lead portion, means for replacing the recognized lead portion with numerical data, and processing means for the numerical data, wherein the floating amount of each lead of the surface mount type integrated circuit package while performing backtracking. And a binary search algorithm for calculating the read flatness and arranging the leads in ascending order of the number.