JP4676315B2 - Bonding equipment - Google Patents

Description

  The present invention relates to a bonding apparatus that recognizes a chip.

  Conventionally, when bonding a chip such as a semiconductor to a lead frame or the like, a bonding apparatus has been used.

  This bonding apparatus is configured to bond a chip picked up from a wafer to a lead frame or the like. The bonding apparatus is configured to set a wafer ring, and a wafer sheet is stretched on the wafer ring. A wafer is stuck on the wafer sheet, and the wafer is diced. Thus, dicing streets are formed vertically and horizontally on the wafer, and a chip is formed by each street.

  In the bonding apparatus in which the wafer ring is set, an image of the wafer is acquired and image processing is performed on the wafer image to determine the presence / absence of a chip and pass / fail.

  FIG. 8 is a diagram showing a wafer image 801 and a histogram 802 acquired by the camera. In the histogram 802, the gray value at each pixel in the wafer image 801 is classified into 256 levels and shown on the X axis 803. The number of pixels in each classified gray value is shown on the Y axis 804.

  When the presence / absence of the chips 811,... Is recognized using the wafer image 801, the wafer image 801 is binarized to form a binarized image, and the chips 811,. Clarified the division with ...

  However, in such a bonding apparatus, a threshold value is used when a binarized image is formed from a wafer image 801 acquired by a camera. This threshold value has been selected by an operator.

  For this reason, erroneous recognition may occur depending on the selection of the threshold value.

  That is, (a) of FIG. 9 shows a binarized image 821 binarized by setting the threshold value to “93” indicating a gray value in a valley with relatively low brightness, and the chip portion. .., And street portions 823... Are clearly separated.

  On the other hand, FIG. 9B is a diagram showing a binarized image 831 that is binarized by setting the threshold value to “148” that indicates a light and shade value in a valley with relatively high brightness. Only the pattern portions 832,... Provided inside are displayed in white (bright portions). In this case, there is a possibility that the presence / absence and quality of the chip 811 cannot be accurately grasped.

  The present invention has been made in view of such a conventional problem, and an object of the present invention is to provide a bonding apparatus capable of selecting an appropriate threshold value.

In order to solve the above-mentioned problem, in the bonding apparatus according to claim 1 of the present invention, in the bonding apparatus in which a threshold value is used when binarizing each pixel of a wafer image based on shading, each of the wafer image Histogram forming means for classifying pixels for each gray value and forming a histogram showing the distribution of the number of pixels for each gray value, and detecting a minimum with a small number of pixels as a valley in the distribution curve formed by the histogram, And a threshold value selection means for updating the histogram by setting the gray value of the valley showing the darkest gray value among the detected valleys as the threshold value, and discarding the gray value data larger than the threshold value.

  That is, when obtaining the threshold value for binarizing the wafer image, first, each pixel of the wafer image is classified for each gray value, and a histogram indicating the distribution of the number of pixels for each gray value is formed. Then, a minimum having a small number of pixels is detected as a valley in the distribution curve formed by the histogram, and the valley indicating the darkest gray value among the detected valleys, that is, the gray value of the valley with the lowest lightness is the threshold value. It is said.

  As a result, the minimum valley, which is the valley with the lowest brightness, is detected from the wafer image, and the gray value at the minimum valley is set as the threshold value.

In the bonding apparatus according to claim 2, the threshold selection unit includes: a temporary threshold calculation unit that calculates a temporary threshold from the histogram formed by the histogram formation unit using an Otsu method; and the temporary threshold calculation unit. Histogram updating means for updating the histogram by discarding data whose gray value is larger than the calculated temporary threshold, and a comparison threshold for calculating the comparison threshold from the updated histogram updated by the histogram updating means using the Otsu method a threshold calculating means, and the approximate curve calculation means for the obtaining an approximate curve between comparative threshold calculated by the temporary threshold and the comparator threshold calculation means calculated in the temporary threshold value calculating means calculates at the approximate curve calculation means judgment means for approximating a curve to determine whether in Yamagata having a maximum was, the approximate curve is the determining means determines that there Yamagata When, the comparison threshold as the temporary threshold, and repeating means to move to the histogram updating unit, when the determining unit determines that no said approximation curve by Yamagata, threshold for the temporary threshold and the threshold value Determining means.

  That is, when obtaining the threshold value for binarizing the wafer image, first, each pixel of the wafer image is classified for each gray value, and a histogram indicating the distribution of the number of pixels for each gray value is formed.

  Then, a temporary threshold is calculated from this histogram using the Otsu method, and data having a gray value larger than the temporary threshold is rounded down to update the histogram.

  Next, a comparison threshold value is calculated from the updated histogram using the Otsu method, an approximate curve between the temporary threshold value and the comparison threshold value is obtained, and whether the approximate curve is a mountain shape having a maximum. Judge whether or not.

  At this time, when the approximate curve has a mountain shape, it can be seen that there is a peak showing a maximum between the temporary threshold value and the comparison threshold value, so that the comparison threshold value is a valley. Therefore, in this case, using the comparison threshold value as the temporary threshold value, the updating of the histogram, the calculation of the comparison threshold value, the calculation of the approximate curve, and the determination are repeated in order until the approximate curve is no longer chevron.

  On the other hand, when the approximate curve is not a mountain shape, there is no peak showing a maximum between the temporary threshold value and the comparison threshold value, and it can be seen that the comparison threshold value is not a valley. For this reason, the temporary threshold at this time can be determined as the minimum valley that is the valley with the lowest brightness, and the gray value at the minimum valley is the threshold.

  As described above, in the bonding apparatus according to the first aspect of the present invention, the minimum valley that is the valley having the lowest brightness can be detected from the wafer image, and the gray value at the minimum valley can be used as a threshold value.

  For this reason, it is possible to automatically select an appropriate threshold value as compared with the conventional case where the operator has selected the threshold value. And in the binarized image binarized using this threshold value, it is possible to accurately grasp the presence / absence and quality of the chip.

  In the bonding apparatus according to the second aspect, the valley position can be detected from the histogram by using a generally used Otsu method.

  Then, the data with a gray value larger than the threshold value obtained by the Otsu method is rounded down and the histogram is updated, and further, the gray value at the minimum valley bottom is detected by repeating the detection operation for detecting the valley position by the Otsu method. And can be used as a threshold value.

  This makes it possible to reliably obtain the gray value at the minimum valley bottom while utilizing the method of Otsu, which tends to obtain the valley position at the intermediate position as the threshold value.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a bonding apparatus 1 according to the present embodiment. The bonding apparatus 1 picks up a chip 4 on a wafer sheet 3 stretched on a wafer ring 2 and leads a lead frame (not shown). It is a device that moves up and bonds.

  A wafer 5 is stuck on the wafer sheet 3 and the wafer 5 is diced. As a result, streets 6,... Formed by dicing are formed vertically and horizontally on the wafer 5, and the chips 4,.

  The bonding apparatus 1 includes a camera 12 that acquires an image of the wafer 5 on the wafer sheet 3 set on the wafer table 11. The camera 12 is configured to acquire a wafer image 13 (see FIG. 2) obtained by photographing the wafer 5 on the wafer sheet 3. The wafer image 13 acquired by the camera 12 is an image processing unit. 14 is configured so as to be output to 14.

  The image processing unit 14 is configured around a microcomputer, and is configured to process the wafer image 13 from the camera 12 to form a histogram 21 (see FIG. 2). The image processing unit 14 is configured so as to be able to inspect the presence / absence of the chip 4 and the state of inclination, for example, by performing image processing on the wafer image 13.

  The image processing unit 14 is configured to output a signal to the wafer table control unit 31, and the wafer table control unit 31 is configured to output the wafer table 11 based on the signal from the image processing unit 14. For example, a pickup at a chip position where the chip is missing is prevented, or the tilt of the chip 4 is corrected.

  The operation of the bonding apparatus 1 according to the present embodiment having the above configuration will be described with reference to the flowchart shown in FIG.

  That is, the image processing unit 14 that has acquired the wafer image 13 from the camera 12 creates the histogram 21 from the wafer image 13 as shown in FIG. 3 (S1).

  More specifically, the brightness of each pixel constituting the wafer image 13 is set as a shade value from “0” to “255”, and the number of pixels indicating each shade value is calculated. That is, the density value of the dark pixel with the lowest brightness is set to “0”, the brightness value of the lightest pixel with the highest brightness is set to “255”, and the number of pixels with the brightness value set to “0” is obtained. In addition, the number of pixels whose gray value is set to “1” to the number of pixels whose gray value is set to “255” are obtained in order. Next, each gray value is set on the horizontal axis (X axis), and the number of pixels set for each gray value is set on the vertical axis (Y axis), indicating the distribution of the number of pixels at each gray value. The histogram 21 is formed.

  Then, as shown in FIG. 4, a temporary threshold value is calculated from the distribution curve 41 formed by the histogram 21 using the Otsu equation 51 defined by the Otsu method (S2). Then, as shown in FIG. 5A, the gray value of the second valley 43 indicating the second minimum from the lower gray value is calculated as the temporary threshold.

  Next, data larger than the gray value is rounded down by setting the number of pixels at each gray value having a gray value larger than the provisional threshold determined by the gray value of the second valley 43 to “0” (S3). ), A new updated histogram 61 is formed as shown in FIG. 5B (S4).

  A threshold for comparison is calculated from the distribution curve 41 formed by the updated update histogram 61 using the Otsu's equation 51 shown in FIG. 4 (S5). Then, as shown in FIG. 6A, the gray value of the first valley 42 indicating the first minimum from the lower gray value is calculated as the comparison threshold value.

  After calculating the approximate curve 71 as shown in FIG. 6B by calculating the least-squares method two-dimensional polynomial 52 shown in FIG. 4 between the comparison threshold and the provisional threshold. (S6) It is determined whether or not this approximate curve 71 is a mountain having a maximum (S7).

  At this time, when the approximate curve 71 has a mountain shape, there is a mountain having a maximum 72 between the temporary threshold value and the comparison threshold value, so that the comparison threshold value is a valley. For this reason, in this case, the comparison threshold value is set as the temporary threshold value (S8), and the steps S3 to S7 are repeated in order until the approximate curve 71 is not a mountain shape.

  That is, data larger than the gray value is rounded down by setting the number of pixels at each gray value having a gray value larger than the provisional threshold determined by the gray value of the first valley 42 to “0” (S3), As shown in FIG. 7A, a new update histogram 65 is formed (S4), and from the distribution curve 41 formed by the updated update histogram 65, the Otsu equation 51 shown in FIG. 4 is used. The threshold for comparison is calculated (S5). Then, as shown in FIG. 7 (b), for example, the shade value of the first peak 81 on the side where the shade value is lower is calculated as the comparison threshold value. After calculating the approximate curve 75 by calculating the two-dimensional polynomial 52 of the least square method shown in FIG. 4 (S6), it is determined whether or not the approximate curve 75 is a mountain having a maximum (S7). .

  At this time, when it is determined that the approximate curve 75 is not a mountain shape, a valley indicating a minimum 76 exists between the temporary threshold and the comparison threshold as shown in FIG. Therefore, it can be seen that the comparison threshold value is not a valley. For this reason, since the temporary threshold value at this time can be determined as the minimum valley that is the valley having the lowest brightness, the gray value at the minimum valley is selected as the threshold (S9), and this routine is terminated.

  As described above, the minimum valley which is the first valley having the lowest brightness can be detected from the wafer image 13, and the gray value at the minimum valley can be set as the threshold value.

  For this reason, it is possible to automatically select an appropriate threshold value as compared with the conventional case where the operator has selected the threshold value.

  When the threshold value is used to binarize the wafer image 13 to form a binarized image, it is possible to accurately grasp the presence / absence and quality of a chip from the binarized image.

  In addition, the valley position can be detected from the histogram 21 by using Otsu's formula 51 which is generally used.

  Then, the data having a gray value larger than the threshold value obtained by Otsu's equation 51 is discarded, the histogram 21 is updated, and the detection operation for detecting the valley position by Otsu's equation 51 is further repeated, whereby the lightness and shade at the minimum valley bottom is repeated. A value can be detected and used as the threshold value.

  This makes it possible to reliably obtain the gray value at the minimum valley bottom while using Otsu's equation 51, which tends to obtain the valley position at the intermediate position as a threshold value.

It is a block diagram which shows one embodiment of this invention. (A) is a figure which shows the wafer image of the embodiment, (b) is a figure which shows a histogram. It is a flowchart which shows the operation | movement of the embodiment. (A) is a figure which shows Otsu's type | formula used in the embodiment, (b) is a figure which shows the two-dimensional polynomial used in the embodiment. (A) is explanatory drawing which shows operation | movement of the embodiment, (b) is explanatory drawing which shows the operation | movement following (a) of the figure. (A) is explanatory drawing which shows the operation | movement following (b) of FIG. 5, (b) is explanatory drawing which shows the operation | movement following (a) of the figure. (A) is explanatory drawing which shows the operation | movement following (b) of FIG. 6, (b) is explanatory drawing which shows the operation | movement following (a) of the figure. It is a figure which shows the wafer image and histogram in a prior art example. (A) is a figure which shows the binarized image of the same prior art example, (b) is a figure which shows the binarized image on other conditions.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Bonding apparatus 3 Wafer sheet 4 Chip 5 Wafer 13 Wafer image 14 Image processing part 21 Histogram 41 Distribution curve 42 1st trough 43 2nd trough 71 Approximate curve

Claims (2)

In a bonding apparatus in which a threshold value is used when binarizing each pixel of a wafer image based on shading,
Histogram forming means for classifying each pixel of the wafer image for each gray value and forming a histogram indicating the distribution of the number of pixels for each gray value;
In the distribution curve formed by the histogram, a minimum with a small number of pixels is detected as a trough, and a trough value indicating a darkest shading value among the detected troughs is set as the threshold value, and a gray value larger than the threshold value Threshold selection means for truncating data and updating the histogram;
A bonding apparatus comprising: The threshold selection means includes
A temporary threshold calculating means for calculating a temporary threshold from the histogram formed by the histogram forming means using the method of Otsu;
Histogram updating means for updating the histogram by truncating data having a gray value larger than the temporary threshold calculated by the temporary threshold calculating means;
A comparison threshold value calculating means for calculating a comparison threshold value from the updated histogram updated by the histogram updating means using the method of Otsu;
An approximate curve calculation means for obtaining an approximate curve between the temporary threshold calculated by the temporary threshold calculation means and the comparison threshold calculated by the comparison threshold calculation means;
Determination means for determining whether or not the approximate curve obtained by the approximate curve calculation means is a mountain having a maximum;
When the determination unit determines that the approximate curve is a mountain shape, a repetition unit that shifts to the histogram update unit using the comparison threshold as the temporary threshold; and
A threshold value determining unit that sets the temporary threshold value as the threshold value when the determining unit determines that the approximate curve is not a mountain shape;
The bonding apparatus according to claim 1, further comprising:

Images