JPH04158206A - Component inspecting apparatus - Google Patents

Abstract

PURPOSE:To reduce the amount of storage memory of image data and to shorten a time for an image data processing by detecting the attitude of a held component to be inspected and by detecting the amount of bend of an input-output pin through a window set on the basis of detection of the attitude. CONSTITUTION:An image of the whole of a component 1 to be inspected whereon input-output pins 2 are arranged is picked up bya TV camera 4, and a first video signal V1 thereof is subjected to A/D conversion. From first image data D1 obtained from the signal and stored 5, an image processing element 6 detects the attitude of the component 1, such as an inclination thereof. Meanwhile, a window dividing the whole of the component 1 into prescribed areas is set on the basis of the detected attitude, an image of each area is picked up by the camera 4 and a second video signal V2 thereof is subjected to the A/D conversion 9. From second image data obtained from this signal and stored 5, the processing element 6 detects the position of the pin 2 in the window. This position and reference data D4 are compared by a discriminating element 8 and the result is delivered as discrimination data D5. Accordingly, the detection of the presence or absence of a bend of the pin 2 due to a posisional slippage, as well as the display of the position of the pin 2 in arrangement in the case when the amount of the bend exceeds an allowance value, can be executed and thus detection of a faulty part is enabled.

Description

[Detailed Description of the Invention] [Summary] A part to be inspected in which input/output pins are arranged is photographed by a television camera, and the positional deviation of the input/output pins is detected based on the photographed image data, thereby inspecting the part to be inspected. With regard to component inspection equipment that performs a photographing section that outputs a first video signal that photographs the entirety of the part to be inspected in which the output pins are arranged and a second video signal that photographs a window obtained by dividing the whole part into predetermined regions; A/
an image storage unit that stores first and second image data A/D converted by a D converter; an image processing unit that processes the first and second image data stored in the image storage unit; A main storage unit that stores processed data processed by the image processing unit, and a determination unit that compares the processed data stored in the main storage unit with reference data and outputs determination data for determining quality. The image processing unit detects a posture of the inspected part based on the first image data, and corrects the second image data based on the posture, thereby detecting a position of the input/output pin in the window. , the discrimination data indicating the positional deviation of the input/output pin is output from the discrimination section, and the inspected component is inspected.

[Industrial application field]

The present invention relates to a component inspection device that inspects a component to be inspected by photographing a component to be inspected in which input/output pins are arranged using a television camera, and detecting a positional shift of the input/output pins based on the photographed image data. .

Electronic components such as semiconductor packages used in electronic devices are formed by arranging a large number of input/output pins.

In addition, these electronic components are becoming more compact and highly functional, leading to higher densities.Especially, when electronic components are mounted on a board, the margin for mutual alignment becomes extremely small, and input/output pins are If the input/output pins are not arranged at a predetermined pitch due to bending or the like, mounting cannot be performed.

Therefore, an inspection device is required to detect whether the input/output pins are arranged at a predetermined pitch.

[Conventional technology]

Conventionally, the configuration was as shown in the conventional configuration diagram shown in FIG.

As shown in FIG. 9, the part to be inspected 1 is
By irradiating the fluorescent lamp illumination 31 through the light diffusing plate 32 while the part is held by the slit 0, a silhouette image of the inspected part l is created by the slit illumination formed by the illumination 33.35 and the electromagnetic shutter 34.36. The input/output pins 2 are arranged on the part to be inspected by irradiating the lights 33A and 35A.
images were respectively taken by a camera 51 having an optical system 50.

The part to be inspected l in this case is formed as shown in the part to be inspected in Fig. 1θ. FIG. 10(a) is a side view, (
b) is a plan view.

As shown in (a) and (b), the input/output pins 2 of the base IA are arranged at a predetermined pitch P. For example, when L is about 10 mm, the pitch P is set to 0. By making it .45 mm, the density is increased so that 400 input/output pins 2 are arranged.

Further, the camera 51 is fixed to a fixed base 47, and an X-Y table 48 driven by a moving base movable part 46 moves the
, the camera 5I is formed to be moved in the Y direction, and the camera 5I has a line sensor 52 installed therein, scans the image 53,
The image data D is configured to be sent to the image storage section 54A of the image processing device 54.

Furthermore, the image processing device 54 is provided with a data processing section 54B, a reference data storage section 54C, and a discrimination section 54D, and the data processing section 54B processes the image data D stored in the image storage section 54A, and outputs the input/output pins. Detect the sequence of 2,
The arrangement detected by the discrimination unit 54D is stored in the reference data storage unit 54C.
It compares it with the reference value stored in , and sends out the discrimination data D5 which is determined to be good or bad.

Therefore, when performing an inspection, the system control device 40 controls the drive control device 41 and the image processing device 54, and first the drive control device 41 drives the suction drive section 44 and the head movable section 43. While holding the component 1 to be inspected at a predetermined position, the fluorescent lamp 31 and the slit illumination light 33.35A are blinked by driving the illumination 0N10FF drive unit 42 and the shutter opening/closing drive unit 45.

Next, the camera 51 is positioned directly below the part to be inspected 1, and the line sensor 52 is scanned by a command from the drive control device 41, and image data D based on the image 53 of the part to be inspected 1 is scanned.
is captured by the image processing device 54.

Therefore, by outputting the discrimination data D5 sent from the discrimination unit 54D of the image processing device 54 by the printer, it is possible to check whether the arrangement of the input/output pins 2 is all within the reference value, and if it deviates from the reference value. The position of the input/output pins 2 is detected, and the inspection of the part to be inspected l is automatically performed.

[Problem to be solved by the invention]

Input/output pin 2 of such a high-density inspected component l
In order to inspect the array, a resolution performance of about 2.5 μm/pixel is required, and inspection is possible by using, for example, a line sensor 52 with 4096 pixels, which is currently in practical use. High quality image data D can be obtained.

However, in order to obtain image data D by scanning the line sensor 52, the scanning speed of -step must be increased by 2.
．． Then, when the line sensor 52 having 4096 pixels is used, approximately 33.6 seconds are required to input one screen as shown in equation (1).

4096×4096×2μs #33.6s・・・・・・
(1) Also, by using an 8-bit configuration, 16 MB of image memory is required as shown in equation (2).

4096x 4096x 8bit= 16MB ・
...(2) Therefore, the image data D to be inspected
It takes a long time to obtain the desired results, requires a large number of inspection man-hours, and also requires a large capacity memory.

Therefore, it is an object of the present invention to reduce the number of inspection steps by increasing the image processing speed, and to downsize the apparatus by minimizing the required memory capacity.

[Means to solve the problem]

FIG. 1 is a diagram explaining the principle of the present invention.

As shown in FIG. 1, a first video signal Vt captures the entirety of the inspected part 1 on which the input/output pins 2 are arranged by a television camera 4, and a first video signal Vt captures the entirety of the inspected part 1 in which the input/output pins 2 are arranged. a photographing unit 3 that outputs a video signal (2) of 2; first and second image data D1 obtained by A/D converting the first and second video signals V1 and V2 by an A/D converter 9; an image storage unit 5 that stores image data D2; an image processing unit 6 that processes the first and second image data D1 and D2 stored in the image storage unit 5; and processed data processed by the image processing unit 6. A main storage unit 7 that stores D3, and a main storage unit 7 that stores the processed data D3 stored in the main storage unit 7 as reference data D4.
and a determination unit 8 that outputs determination data D5 for determining pass/fail by comparison with detecting the position of the input/output pin 2 in the window by correcting the second image data D2;
The discrimination data D5 indicating the positional deviation of the input/output pin 2 is outputted from the discrimination section 8, and the inspected part 1 is inspected.

With this configuration, the above-mentioned problem is solved.

[Effect]

That is, the input/output pin 2 is
A first video signal Vt capturing the entire array of inspected parts 1 is A/D converted by an A/D converter 9,
The first image data DI is stored in the image storage section 5, and the image processing section 6 detects the orientation of the inspected part l from the first image data DI, such as the inclination.

On the other hand, a window is set for dividing the entire part to be inspected l into predetermined areas according to the detected posture of the part to be inspected l, and a second video signal v2 captured by the television camera 4 for each window is similarly set. , A/D converter 9
The stored second image data D2 is converted into A/D by the image processing unit 6, and is stored in the image storage unit 5 as second image data D2. The detected position is compared with reference data D4 by the discrimination section 8, and the comparison result is sent from the discrimination section 8 as discrimination □ data D5.

Therefore, it is possible to detect whether or not there is a bend due to the positional deviation of the output pin 2 using the discrimination data D5, and to display the arrangement position of the input/output pin 2 if the amount of bend exceeds a predetermined tolerance. , it is possible to detect defective parts.

〔Example〕

The present invention will be explained in detail below with reference to FIGS. 2 to 8.

FIG. 2 is a configuration diagram of an embodiment according to the present invention.

FIG. 3 is a flowchart of the present invention, FIG. 4 is a density histogram of the first video signal, and (a1) (a) in FIG.
2) (b1) (b2) are projection distribution maps in the vertical and horizontal directions, (a) and (b) in Figure 6 are diagrams explaining the processing of the first image data, Figure 7 is a diagram explaining the window, and Figure 8 (aXb
)(c) is an explanatory diagram of processing of the second image data. The same reference numerals indicate the same objects throughout the figures.

As shown in FIG. 2, the photographing section 3 for photographing the inspected part l disposed at the fixed station 47 is configured to photograph the image reflected by the half mirror 14 in the direction of arrow A using a low magnification lens 16. A television camera 4A supported by a post 13 and a television camera 4B which captures a transmitted image as indicated by an arrow B using a high magnification lens 15 are provided.

Further, the first video signal (1) captured by the television camera 4A and the second video signal (2) captured by the television camera 4B are each sent to the camera selector 12, and the camera selector 12 is connected to the drive control device 41. According to the command, either the first or second video signal V1 or V2 is sent to the A/D converter 9, the image storage section 5, the image processing section 6, the main storage section 7, the discrimination section 8, and the reference data. Storage part 1
The configuration is such that the data is sent to an image processing apparatus IO equipped with 1 and 1, and the other configurations are the same as those described above.

Therefore, the part to be inspected l is held by the vacuum suction head 30 and kept stationary directly above the photographing section 3 , and the fluorescent lamp 31 is turned on by the illumination 0N10FF drive part 42 to capture the silhouette image of the part to be inspected 1 on the half mirror 14 . The first video signal captured by the television camera 4A due to reflection.
1 is sent to the A/D converter 9.
Control 2.

The first video signal Vl sent to the A/D converter 9 is A
After the /D conversion, the image storage unit 5 stores the first image data Di as the first image data Di.
is stored in

In this case, A/D conversion is performed by the image processing unit 6 from 0 to 2
Count the pixels of density divided into 55 stages, create a graph shown in the density histogram diagram of the first video signal in FIG. Find the area S shown by from the dark part of the density,
A/D conversion is performed by binarizing the density t at that time as a slice level to form first image data D,
The image is stored in the image storage unit 5.

Next, the vertical and horizontal projection distribution map shown in FIG. 5 is formed using the first image data DI.

For example, if the first image data DI is in the state shown in (a1) in FIG. 5, the graph shown in (a2) can be created by adding the number of pixels that are I- for each column on the By forming the stored coordinates X1 and X2 and the distribution length m can be obtained.

Similarly, if the state shown in (b1) in FIG.
By adding the number of pixels equal to 1' for each column on the Y axis, the stored coordinate Y is created by forming the graph shown in (b2).
1, Y2 can be found.

In this case, the center points GX, GY of the first image data Di
can be determined using equations (3) and (4).

GX = 1/2 (Xl+X2) ・・・・・・(3
) GY = 1/2 (Yl + Y2) ...... (4
) Furthermore, as shown in (a) in the diagram for explaining the processing of the first image data in FIG.
The length of AI is calculated by performing a search in the axial direction and first detecting the coordinate that is l'.

In this case, the relationship A1=m-Bl holds, so (5) f
The inclination angle θ can be determined using equation 6).

If AI < Bl, θ=TAN-'(AI/B1) (5)
If AI < Bl, θ=-TAN-'(At/B1) (6)
Note that, since the expected maximum inclination angle θ is within ±10 degrees, the case where AI=81 (θ=45 degrees) is not considered.

The first image data old is divided according to the center points GX and GY calculated in this way, their inclination angle θ, and the pitch P of the input/output pin 2 according to the design value, and the data is divided as shown in the explanatory diagram of the window in Fig. 7. The windows W1 to WN are formed as follows, and the center WOI of each window W1, W2 to WN.

Determine WO2 to WON.

Therefore, the TV camera 4B is placed at the center WO of the window Wl.
The X-Y table 48 is moved by the movable table movable part 46 so as to move to the position I, and the second video signal v2 of the enlarged image of the high magnification lens 15 by the television camera 4B is inputted under the control of the camera selector 12.

In this case, it is necessary to physically match the center positions of the TV cameras 4A and 4B, but even if they do not match, if the offset value is calculated in advance, it can be easily corrected using that value during calculation. I can do things.

When inputting the second video signal v2, the fluorescent lamp 31 is turned off and the slit illumination lights 33A and 35A are applied to input the second video signal v2 to the A/D converter 9.
The A/D conversion is performed by creating a density histogram diagram equivalent to that shown in FIG. Image data D2 is formed and stored in the image storage section 5.

However, when forming the second image data D2, the input/output pin 2
Since the image is obtained by reflected light, the bright side is set to 1-, and the dark side is set to 0'.

Next, as shown in FIGS. 8A, 8B, and 8C, which are explanatory diagrams for explaining the processing of the second image data, the second image data D2 is processed.

The number of pixels of l in the window Wl shown in (a) is integrated on the X-axis and Y-axis to form the vertical and horizontal projection distribution maps shown in (b) and (c), and the areas shown by diagonal lines in each are SP
Find the locations Xg and Yg that are 1/2 of .

Therefore, X at the center WOI of window Wl,
When the center lines in the Y direction are Xc and Yc, respectively, the positional deviation δX of the input/output pin 2 in the X-axis direction and the positional deviation GY in the Y-axis direction can be calculated using equations (7) and (8). Furthermore, the amount of bending δ in this case can be calculated using equation (9).

δX =Xc-Xg ・・・・・・Fist・(7)
GY=Yc-Yg・・・・・・・・・・・・(8)
δ = 6T11] 7... (9) When the amount of curvature δ in the window Wl is detected in this way,
Next, the window W2 is photographed, and the video signal v2 is similarly processed to detect the amount of bending δ in the window W2.
By detecting the bending amount δ in N, the bending amount δ in all windows W1, W2 to WN is detected, and the detection of the bending amount δ in all windows Wt, W2 to WN is completed. The main storage unit 7 stores the processed data D3 as
is stored in

Further, the determination unit 8 retrieves reference data D4 indicating the position of the input/output pin calculated based on the design value from the previously stored reference data storage unit 11, and compares it with the processed data D3 stored in the main storage unit 7. It is determined whether the total amount of bending δ of the input/output pins 2 is within an allowable range, and the result is output as determination data D5.

Therefore, in the inspection of the part to be inspected, as shown in FIG.
An image of the entire part to be inspected l is taken by , the posture of the part to be inspected 1 such as inclination is detected, and then windows W1, W2 to WN are set by dividing the part to be inspected l, and each window W1, TV camera 4 captures images from W2 to WN.
B, and the amount of bending δ of the input/output pin 2 is detected.

The detection of the amount of bending δ in all input/output pins 2 is confirmed, and finally, the results are outputted using the discrimination data D5 for inspection.

In such a configuration, the CCD used for the television cameras 4A and 4B has sufficient resolution with 512 x 512 pixels, so the amount of memory required for an 8-bit configuration is 512 x 512 x 8 = 256.
It is clear that the amount of memory is reduced compared to the conventional structure shown in FIG. 9.

Furthermore, since the images were taken using video cameras 4A and 4B, the input time for video signals V1 and V2 was 1760 seconds (standard video rate), and the total image input time was 400 (the number of input/output pins 2). In case, 401 XI/604
=; 6.7 seconds, which is approximately 175 seconds compared to the conventional method.

In the present invention, the entire image and the enlarged image are taken using two cameras, the video cameras 4A and 4B, but the entire image and the enlarged image are taken using one camera using an auto zoom lens. Even with this configuration, the same effect can be obtained.

〔Effect of the invention〕

As explained above, according to the present invention, the holding posture of the part to be inspected is detected by photographing the entire image of the part to be inspected,
Next, set the window according to the detected pose,
By detecting the amount of bending of the input/output pin using an enlarged image using a window, the amount of memory for storing image data can be reduced (and the processing time of image data can be shortened).

Therefore, compared to the conventional method, the device can be made smaller and the inspection work can be made more efficient, which has great practical effects.

[Brief explanation of drawings]

FIG. 1 is a diagram explaining the principle of the present invention. FIG. 2 is a configuration diagram of an embodiment according to the present invention. FIG. 3 is a flow chart diagram of the present invention. FIG. 4 is a density histogram diagram of the first video signal. (at) (a2) (b1) (b2) in FIG. 5 are projection distribution maps in the vertical and horizontal directions. (a) and (b) of FIG. 6 are explanatory diagrams of processing of the first image data. FIG. 7 is an explanatory diagram of the window. (a), (b), and (c) of FIG. 8 are explanatory diagrams of processing of the second image data. FIG. 9 is a conventional configuration diagram. Figure 1O is an explanatory diagram of the part to be inspected, where (a) is a side view, (
b) shows a plan view. In the figure, 1 is the component to be inspected, and 2 is the input/output pin. 3 is the photography department, 4 is the TV camera. 5 is an image storage unit, and 6 is an image processing unit. 7 is a main memory section, and 8 is a discriminator section. 9 is an A/D converter, and 1 is the first video signal. v2 is the second video signal, DI is the first image data. ；Teisuke gen gift misfire 88M explanation mouth haze 1 ro misfire day/) 70-Yatro 1 '3 day pixel count 1r: ゛teoi lucky issue/) density distogram mouth millet 4
Misfortune] (α1) Let's reconsider! Projection distribution record of Eihei senku 75 Kasumi (shi) ★F If)! Statue 7' and evening) 9ri Of Risage F March Melon] No. 6 [Vinh7/) Speech S@E1 Y-axis (a) ×3 (C) Part 2 4*-7'-? /ly, x scissors n mouth 8th mouth

Claims (1)

[Claims] A first video signal (V1) that captures the entire part to be inspected (1) in which input/output pins (2) are arranged by a television camera (4), and a second video signal that captures a window that divides the entire part into predetermined areas. a first video signal (V2) that outputs a second video signal (V2); Second
An image storage unit (5) that stores image data (D1, D2) of
), an image processing unit (6) that processes the first and second image data (D1, D2) stored in the image storage unit (5), and processing performed by the image processing unit (6). data(
D3); and a main storage unit (7) that stores
The processed data (D3) stored in the reference data (D4
), and the image processing section (6) outputs the discrimination data (D5) for determining the quality of the inspected part (1).
), and by correcting the second image data (D2) based on the orientation, the position of the input/output pin (2) in the window is detected, and the position of the input/output pin (2) is detected. The discrimination data (D5) indicating positional deviation is processed by the discrimination unit (
8) and inspects the inspected component (1).