JP2843389B2 - Bonding ball inspection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Summary] The present invention relates to a bonding ball inspection apparatus for measuring and inspecting a dimension of a bonding ball diameter which is an important element in managing a bonding state in wire bonding of an integrated circuit or the like. An imaging unit for imaging a bonding ball at a tip of a wire bonded to a pad for the purpose of measuring the size of a bonding ball diameter with high accuracy without being affected by a concentration gradient as much as possible, and an output image of the imaging unit A pixel generating means for generating a pixel by digitally processing a signal; and a sub-pixel for generating a sub-pixel by performing gradation interpolation between pixels of each pixel in an inspection area of a predetermined size among the pixels generated by the pixel generating means. Generating means; each pixel generated by the pixel generating means in the inspection area; A gray-scale image memory for storing the respective sub-pixels generated by the pixel generator, and a second differentiation for the pixels and sub-pixels read from the gray-scale image memory, and zero-crossing the result of the second differentiation. And a zero-cross detecting means for detecting the diameter of the bonding ball based on the detection result of the zero-cross detecting means.

[Industrial applications]

The present invention relates to a bonding ball inspection apparatus, and more particularly to a bonding ball inspection apparatus that measures and inspects a dimension of a bonding ball diameter, which is an important element in managing a bonding state in wire bonding of an integrated circuit or the like.

In recent years, integrated circuits (ICs) and large-scale integrated circuits (LSIs) have been used widely and in large quantities, so that visual inspection of the wire bonding state of ICs and LSIs has been automated. It has become. In the appearance inspection of the wire bonding state, it is necessary to accurately measure the diameter of the bonding ball.

[Conventional technology]

FIG. 7 shows a configuration diagram of an example of a conventional bonding ball inspection apparatus. In the figure, reference numeral 1 denotes a lead frame on which an IC chip 2 is mounted. Pads 3 are formed on the IC chip 2. 4 is a wire, one end of which is a pad 3
A bonding ball 5 is formed thereon and fixed.

The lead frame 1 is transported by a frame feeder 6 which is a transport device, directly below the illumination device 7 and the imaging device 8. Thus, the bonding ball 5 is illuminated by the illumination device 7 and is imaged by the imaging device 8. The video signal extracted from the imaging device 8 is supplied to an information processing device 9, where the shading is corrected by analog processing or digital processing to make the image a uniform shaded state, and then a binary value is automatically obtained. Then, the wire bonding state of the captured image is inspected.

Here, in the conventional wire bonding inspection apparatus, a binarization threshold is obtained as a process in the information processing apparatus 9, and the size of the bonding ball diameter is measured from a grayscale image binarized by the binarization threshold. Is inspected for the shape of the bonding ball 5.

[Problems to be solved by the invention]

However, in the conventional bonding inspection apparatus described above,
Since the video signal is digitized, a quantization error occurs. Although this quantization error can be reduced by increasing the resolution of the imaging optical system, the field of view becomes narrower as the resolution increases, so that there is a limit to the increase in resolution.
For this reason, conventionally, there is a relatively large quantization error, which causes a dimensional error in the diameter of the bonding ball 5.

Further, conventionally, the grayscale image binarized by the binarization threshold is affected by the density gradient of the image, and the shape of the bonding ball 5 is changed by the binarization threshold.

The present invention has been made in view of the above points, and it is an object of the present invention to provide a bonding ball inspection apparatus capable of measuring a dimension of a bonding ball diameter with high accuracy without being affected by a quantization error or an image density gradient as much as possible. With the goal.

[Means for solving the problem]

FIG. 1 is a block diagram showing the principle of the present invention. In the figure, 10
Is an imaging means for imaging the bonding ball at the tip of the wire bonded to the pad. Numeral 11 denotes a pixel generation means for digitally processing a video signal to generate pixels. Reference numeral 12 denotes a sub-pixel generation unit that generates a sub-pixel by performing gradation conversion between pixels of each pixel in an inspection area of a predetermined size.

Reference numeral 13 denotes a gray scale image memory for storing each pixel and sub-pixel in the inspection area. Further, reference numeral 14 denotes a zero-cross detection method, which performs second-order differentiation of the pixels and sub-pixels from the gray-scale image memory 13, detects zero-crossings of the second-order differentiation results, and measures the dimensions of the diameter of the bonding ball.

[Action]

In the present invention, the sub-pixel generating means 12 generates a sub-pixel by performing gradation interpolation between each pixel (pixel) in the inspection area, and stores the sub-pixel in the gray-scale image memory 13. Stored together with the pixels. Therefore, the number of pixels extracted from the grayscale image memory 13 is increased by the number of sub-pixels as compared with the related art, so that the digital quantization error of the image is reduced.

By the way, in human vision, lateral inhibiti
On) has a function to emphasize and detect the object outline. On the other hand, a zero crossing method is known as a method for accurately detecting the contour of an object in a digital image (Ma
rr and Hidreth, 1980). This method uses a Gaussian filter to detect changes in image intensity.
r), the image is blurred, the Laplacian (∇ ² ) is applied to the blurred image, and a zero cross (zero cross) of the image is extracted to be an intensity change position (contour) of the image.

The work of blurring the image intensity function I (x, y) with the Gaussian function G is mathematically written as G * I, and this Laplacian is written as ∇ ² (G * I). Through a Gaussian filter and Laplacian applied.
^{In the} 2G * I image, a zero cross is detected by extracting an address of a pixel where a pixel having a positive value and a pixel having a negative value are adjacent to each other.

The present invention detects the contour of an image in view of the zero crossing method. However, in the present invention, since the gradation interpolation is performed by linear approximation from the original pixels by the sub-pixel generation means 12, the image after the gradation interpolation has high image density and blur at the same time. . Therefore, in the present invention, a pixel in which a zero-crossing occurs is extracted by the zero-crossing detecting means 14 from the pixels and sub-pixels subjected to Laplacian (∇ ² ), that is, subjected to the second differentiation, without applying a Gaussian filter to the image. Therefore, the contour of the bonding ball can be accurately detected. This zero-cross detection is less susceptible to a density gradient than a binary image generated by binarization.

As described above, by increasing the number of pixels by the number of sub-pixels in the present invention, the digital quantization error can be reduced, and the same effect as that obtained by performing the blurring process on the image can be obtained. The outline of the bonding ball can be detected more accurately by zero-cross detection. it can.

〔Example〕

FIG. 2 shows a block diagram of one embodiment of the present invention. In the figure,
The same components as those in FIG. 7 are denoted by the same reference numerals, and description thereof will be omitted. In FIG. 2, the lighting device 7 and the imaging device 8
Constitutes the imaging means 10 shown in FIG. In FIG. 2, reference numeral 21 denotes an image input circuit which constitutes the pixel generation means 11 and digitally processes an input video signal. Reference numeral 22 denotes an image memory, 23 denotes an image processing memory, and these constitute the gray-scale image memory 13.

Reference numeral 24 denotes an image processor, which constitutes the sub-pixel generation unit 12 and the zero-cross detection unit 14.
An operation is performed according to a flowchart shown in FIG. An image display circuit 25 causes the display 26 to display an image. The pixel data transfer between the image input circuit 21, the image memory 22, the image processing memory 23, the image processor 24, and the image display circuit 25 is performed via the image bus 27.

A program memory 28 stores a program for controlling the entire bonding ball inspection apparatus. Reference numeral 29 denotes a system processor, which controls the entire circuit of the bonding ball inspection apparatus in accordance with a program from a program memory 28.

A camera controller 30 controls the operation of the imaging device 8. An illumination controller 31 adjusts the amount of illumination of the illumination device 7 with respect to the IC chip 2 to optimize the S / N of the video signal. Reference numeral 32 denotes a frame feeder controller which controls the transfer of the frame feeder 6 which is a transfer device. 33 is I / O
Controller and keyboard and joystick 34,
The interface between the printer 35 and the system processor 29 is provided. The keyboard and the joystick 34 are used for setting inspection conditions and the like, and the printer 35 is used for printing out inspection results and the like.

A floppy disk controller (FDC) and a hard disk controller (HDC) are connected to the floppy disk 37 and the hard disk. The floppy disk 37 and the hard disk 38 store inspection results, inspection standards, and the like. 39 is a system bus,
This is a data transfer bus between the above circuits.

Next, the operation of this embodiment will be described. The imaging device 8
IC chip 2 of wire 4 bonded to IC chip 2
An image of the bonding ball 5 formed at the end on the side is taken, and a video signal obtained thereby is supplied to the image input circuit 21. The image input circuit 21 digitally processes the input video signal to generate image data in which the pixel groups are synthesized in time series,
The image data is transferred to the image memory 22 via the image path 27.
To be stored. The system processor 29 cuts out pixels (image data) in a region D of a predetermined position and size from the image memory 22 and stores them in the image processing memory 23.

By using all the pixels in the area D stored in the image processing memory 23 in this manner, the image processor 24
The sub-pixel generation means 12 and the zero-cross detection means 14 are realized according to the flowchart shown in the figure.

First, as shown in FIG. 3, the image processor 24 reads out (cuts out) all the pixels in the area D (step 4).
1). Next, the image processor 24 performs gradation interpolation (sub-pixel generation) between the read pixels (in FIG. 3,
Step 42). The inter-pixel gray scale interpolation as shown in FIG. 4, among the pixels in the region D read out from the image processing memory 23, shows four pixels Tonariru any phase by a black circle P ₁ to P _4,
The pixel gradation values are represented by P ₁ = f ′ (i, j) and P ₂ = f ′
(I + 1, j), P 3 = f '(i, j + 1), P 4 = f' (i + 1,
If j + 1) to, white circle following equation to the position indicated by the SP f (x, y) = P 1 · (1-α) · (1-β) + + P 2 · α · (1-β) + P 3 · (1 a method for generating a sub-pixel having the tone value of _{-α) · β + P 4 ·} α · β represented by f (x, y). However, the horizontal distance from the above formula alpha P _1, beta denotes the vertical distance from P _1.

By this inter-pixel gradation interpolation, sub-pixels are generated between the four pixels P _{1 to} P _{4 at the} positions indicated by white circles in FIG. 4 in the same manner as described above. As a result, the image of the bonding ball 5 before the above-described inter-pixel tone interpolation has a large jagged portion in the oblique image portion as shown in FIG. 5 (A). As a result, the jagged portion of the image of the bonding ball 5 in the oblique image portion is greatly reduced as shown in FIG. 7B, and an image in which the quantization error associated with digitization is greatly reduced can be obtained. The sub-pixels generated in this manner are stored in a different area from the storage area of the original pixels in the inspection area D of the image processing memory 23.

Next, the image processor 24 reads out each pixel and sub-pixel in the inspection area D from the image processing memory 23 and performs a second-order differentiation process on them (step 43 in FIG. 3). The zero crossing of the secondary differentiation result is detected (same as above, step 44). That is, this step 43
And 44 are processes for realizing the zero-cross detecting means 14.

Here, the image read from the image processing memory 23 in step 43 and the image of the inspection area D by the sub-pixels are images representing the shapes of the pads 3, the wires 4, and the bonding balls 5, as shown in FIG. Assuming that there is, the relationship between the pixel position of the portion indicated by the solid line a in FIG. 4A and the pixel gradation value is as shown in FIG. 4B, and the contour of the bonding ball 5 indicates the density gradient. .

Assuming that the image tone value of this image is represented by f (x, y), when the image is subjected to the second differentiation in step 43, the second differentiation result D (x, y) becomes D (x, y). x, y) = ∇ ² * f (x, y), as shown in FIG. 6 (C).

The zero-cross detection in step 44 is performed in each of the rightward and leftward directions from the center 0 shown in FIG.
The first zero cross position C ₁ and C ₂ respectively detected. Here, zero-cross detection from the center 0 to the right is D (x + 1,
y)> 0, D (x−1, y) <0, which satisfies both conditions D (x,
This is performed by extracting a pixel having y), and zero-cross detection from the center 0 to the left is D (x−1, y)> 0, D (x + 1,
y) <0, by extracting pixels having a value D (x, y) satisfying both conditions.

In addition, the secondary differentiation processing and step
The zero-crossing detection process at 44 is performed not only in one direction with respect to the solid line a in FIG. 6 (A) but also in each of the directions indicated by broken lines in FIG. 6 (A). However, the detection direction does not extend over the wire 4. Wire 4
Is not the bonding ball 5.

Next, the image processor 24 measures the diameter of the bonding ball 5 from the difference between the two zero-cross detection values in each of the detection directions obtained in this way (step in FIG. 3).
45). In this case, the value of each pixel of the image obtained by converting the center of the bonding ball 5 into, for example, a gradation-converted image is obtained from the position of the first detected maximum value by searching from the direction opposite to the image of the wire 4, and the center of the bonding ball 5 is obtained. By performing the processing of the above steps 43 to 45 on a straight line passing through, the diameter of the bonding ball 5 can be accurately measured.

In addition, by measuring the diameter of the bonding ball 5 in each of a plurality of directions, the deformation of the shape of the bonding ball 5 can be inspected.

〔The invention's effect〕

As described above, according to the present invention, the number of pixels in the inspection area can be increased by the number of sub-pixels, so that the quantization error can be reduced as compared with the related art. For Laplacian (∇ ² )
, The outline of the bonding ball can be accurately detected by zero-cross detection.
Compared to the case of using a binarized gray image, the size measurement accuracy of the bonding ball can be improved, and therefore, the inspection accuracy of the bonding ball can be improved as compared with the related art.

[Brief description of the drawings]

FIG. 1 is a block diagram of the principle of the present invention, FIG. 2 is a block diagram of one embodiment of the present invention, FIG. 3 is a flowchart for explaining the operation of one embodiment of the present invention, and FIG. 5, FIG. 5 is a diagram showing an example of an image before and after gray-scale interpolation between pixels, FIG. 6 is an explanatory diagram of ball diameter measurement by zero-cross detection, and FIG. 7 is a configuration diagram of an example of a conventional device. In the figure, 2 is an IC chip, 3 is a pad, 4 is a wire, 5 is a bonding ball, 10 is an imaging means, 11 is an image generation means, 12 is a sub-pixel generation means, 13 is a gradation image memory, and 14 is a zero crossing. The detection means is shown.

Continuation of the front page (58) Field surveyed (Int. Cl. ⁶ , DB name) H01L 21/60 321 H01L 21/66 G01N 21/88 H01L 21/60 301

Claims (1)

(57) [Claims] An image pickup means (10) for picking up an image of a bonding ball at the tip of a wire bonded to a pad, and a pixel generation means (11) for digitally processing an output video signal of the image pickup means (10) to generate a pixel. ), And a sub-pixel generating means (12) for generating a sub-pixel by performing gradation interpolation between pixels of each pixel in a test area of a predetermined size among the pixels generated by the pixel generating means (11).
A grayscale image memory (13) that stores each pixel generated by the pixel generation means (11) in the inspection area and each subpixel generated by the subpixel generation means (12); Zero cross detection means (14) for performing second differentiation on the pixels and sub-pixels read from the gradation image memory (13) and detecting zero cross of the second differentiation result.
A bonding ball inspection apparatus, comprising: measuring a diameter of the bonding ball based on a detection result by the zero cross detection means (14).