JPH11328361A - Position fine adjusting method of input image in image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide the position fine adjusting method of an input image without requiring high-precision mechanical adjustments. SOLUTION: On a green sheet 1, pieces 2 arranged in matrix are printed. Cameras 4-1 to 4-5 are fitted to a camera table 5 so that are arranged on one straight line parallel to an X direction at the same intervals as the pieces 2, but if a fitting error is present, identical patterns in the respective pieces 2 which are picked up by the cameras 4-1 to 4-5 have different coordinate values. An image recognizing device 6, when inputting two-dimensional images from the cameras 4-1 to 4-5 adjusts the horizontal and vertical input timing, camera by camera, so that the respective images have the same horizontal and vertical coordinates.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus such as a pattern inspection apparatus which captures an image with a plurality of cameras, and more particularly to a horizontal and vertical coordinate system for each image by correcting a mounting error of each camera. And a method for finely adjusting the position of a captured image.

[0002]

2. Description of the Related Art Conventionally, PGA (Pin Grid Array) has been known as a mounting technique suitable for a demand for increasing the number of pins of ICs and LSIs. In PGA, a ceramic substrate is used as a base of a package for attaching a chip, and wiring is performed to a lead wire extraction position. To make this ceramic substrate, a so-called green sheet made by kneading alumina powder with a liquid binder is used, and a paste containing a high melting point metal is screen-printed on the green sheet. By firing such a sheet, so-called simultaneous firing, in which the green sheet is sintered and the paste is metallized, is performed.

[0003] As another mounting technique, TAB is used.
(Tape Automated Bonding) is known. TAB method is a polyimide film carrier (TAB tape)
The copper foil pattern formed thereon is joined to the electrode of the IC chip to form an external lead. The copper foil pattern is formed by attaching a copper foil to a film with an adhesive and etching this.

In such a green sheet or film carrier, a pattern is visually inspected by a human using a microscope after the pattern is formed. However, visually inspecting a fine pattern requires skill and has a problem of overworking the eyes. Therefore, as an alternative to the visual inspection, a technique has been proposed in which a pattern formed on a film carrier or the like is imaged by a TV camera and automatically inspected (for example, Japanese Patent Application Laid-Open No. 6-34196).
No. 0).

In the pattern inspection apparatus disclosed in Japanese Patent Application Laid-Open No. 6-341960, in order to reduce the inspection time,
2. Description of the Related Art Images captured using a plurality of cameras are processed in parallel. In this case, it is necessary to arrange a plurality of cameras at predetermined intervals according to the inspection work. The reason is that the same pattern is usually formed at regular intervals on the inspection work, and each input image captured by multiple cameras is compared with the reference master pattern. This is because the image needs to be handled in a single coordinate system.

[0006]

With the miniaturization of circuit patterns accompanying the recent miniaturization of electronic products, the size of defects that must be detected is reduced to 10 μm or less, and the resolution of an image for detecting these defects is reduced. About several μm is required. For this reason, in the above-described pattern inspection apparatus using a plurality of cameras, it is necessary to mechanically arrange the cameras at equal intervals and with high accuracy of about several μm. Adjustment is an extremely difficult task, and even if such adjustment is possible, there is a problem that if the inspection work changes, the adjustment must be performed again. The present invention has been made to solve the above problems, and has as its object to provide a position fine adjustment method of a captured image that does not require high-precision mechanical adjustment.

[0007]

According to the present invention, when taking in a plurality of two-dimensional images picked up by a plurality of cameras, the horizontal coordinates of each image coincide with each other. In this way, the horizontal capture timing is adjusted for each camera, and the vertical capture timing is adjusted for each camera so that the vertical coordinates of each image match. Thus, horizontal,
By adjusting the capture timing in the vertical direction for each camera, if a plurality of cameras are arranged at a predetermined interval on a straight line parallel to the X direction (horizontal direction of the image), for example, The same points on a plurality of imaging targets (inspection workpieces) arranged at predetermined intervals on a straight line have the same X, Y coordinates. In this way, it is possible to correct an error in the X direction of each camera and an error in the Y direction (vertical direction of the image) from the straight line.

[0008]

Next, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram of a pattern inspection apparatus showing an embodiment of the present invention. In FIG. 1, reference numeral 1 denotes a green sheet serving as an inspection work;
Is an independent pattern group piece formed on the green sheet 1, 3 is an XY table on which the green sheet 1 is placed, 4-1, 4-2, 4-3, 4-4, and 4-5 are green. 5 line sensor cameras for imaging the sheet 1,
Reference numeral 5 denotes a camera table for supporting the line sensor cameras 4-1 to 4-5, and reference numeral 6 denotes an image of the pattern to be measured imaged by the line sensor cameras 4-1 to 4-5, which is compared with a reference master pattern. An image recognition device for inspecting the pattern to be measured, a host computer 7 for controlling the entire device, and a display device 8 for displaying the inspection result.

On the green sheet 1, pieces 2 arranged in a matrix of n rows × m columns (n and m are integers of 1 or more) are printed. Usually, the pieces 2 have the same pattern, and one piece corresponds to, for example, one IC.
Such a green sheet 1 is placed on the XY table 3 and imaged by five line sensor cameras 4-1 to 4-5.

The line sensor cameras 4-1 to 4-5 are X
It is attached to the camera table 5 so as to be arranged at a predetermined interval (in this embodiment, at an interval between the pieces 2) on a straight line parallel to the direction. The number of pixels of the line sensor cameras 4-1 to 4-5 in which pixels are arranged in the X direction is, for example, 5000.
Pixel. When the host computer 7 controls the XY table 3, the XY table 3
Move in the Y direction perpendicular to the X direction.

As a result, a two-dimensional image signal is output from each of the line sensor cameras 4-1 to 4-5. Note that X
Needless to say, instead of moving the Y table 3 in the Y direction, the camera table 5 may be moved in the Y direction. The image recognition device 6 includes a line sensor camera 4-1.
When capturing a two-dimensional image output from 4-5, the horizontal and vertical capture timings described below are adjusted.
In the horizontal direction, 4096 out of the 5000 pixels
Pixels are captured, and 8192 pixels are captured in the vertical direction.

Thus, the two-dimensional image signal of 4096 pixels in width × 8192 pixels in height captured by the image recognition device 6 is
The image data is digitized in the image recognition device 6 and stored in an image memory (not shown). The image recognition device 6 compares the pattern stored in the image memory with a master pattern created from non-defective products of the green sheet 1 to check whether the pattern of the green sheet 1 has a defect. Transfer to The host computer 7
The inspection result is displayed on the display device 8.

In the above pattern inspection apparatus,
If the mounting error exists in the five line sensor cameras 4-1 to 4-5, the same pattern in each piece 2 captured by the cameras 4-1 to 4-5 has different coordinate values.
When each image captured by the cameras 4-1 to 4-5 is compared with the same master pattern, a case may be erroneously recognized as a defect. Thus, the image recognition device 6 performs fine adjustment of the position of the captured image by the following processing.

First, a sheet 11 as shown in FIG. 2A on which a mark M serving as a pattern for position fine adjustment is formed is prepared. The sheet 11 is placed on the XY table 3 so that the sheet 11 comes directly below the line sensor camera 4-1, and the sheet 11
Is captured by the camera 4-1. Then, the image recognition device 6
FIG. 2 is a view taken from the line sensor camera 4-1.
The X and Y coordinates (x1, y1) of the center of gravity of the mark M in the image as shown in (b) are obtained and stored. Note that the coordinate values at this time are calculated using, for example, the upper left of the captured image as the origin.

Subsequently, the XY table 3 is moved in the X direction by the above-mentioned predetermined interval so that the sheet 11 is directly below the line sensor camera 4-2, and the sheet 11 is
Image is taken at -2. Then, the image recognizing device 6 uses the X and Y coordinates (x2, y2) of the center of gravity of the mark M in the image as shown in FIG.
And memorize it.

Similarly, while the XY table 3 is sequentially moved in the X direction by a predetermined interval, the sheet 11 is sequentially imaged by the cameras 4-3, 4-4, and 4-5.
FIG. 2D and FIG. 2 taken from 3, 4-4 and 4-5.
(E), X and Y coordinates (x3, y3), (x4, y4), (x4) of the center of gravity of the mark M in the image as shown in FIG.
5, y5) are sequentially calculated and stored.

As described above, the line sensor cameras 4-1 to 4-1
Due to the mounting error of 4-5, the coordinate values of the mark M captured by the cameras 4-1 to 4-5 are shifted. Therefore, the image recognition device 6 adjusts the horizontal and vertical capture timings for each camera based on the stored coordinate values of the center of gravity of the mark M. FIG. 3 is a horizontal view of the image recognition device 6.
FIG. 9 is a diagram for explaining adjustment of a fetch timing in a vertical direction. In FIG. 3, NH indicates the time for 4096 pixels in the horizontal direction, and NV indicates the time for 8192 pixels in the vertical direction.

For example, with reference to the coordinates (x3, y3) of the mark M in the image fetched from the line sensor camera 4-3, the coordinates (x1, y3) of the mark M in the image fetched from the line sensor camera 4-1. y1) has a small Y coordinate. This is because the position of the camera 4-1 in the Y direction is shifted downward in FIG.

Therefore, as shown in FIG. 3, the image recognizing device 6 sets the vertical capture timing SV1 when capturing the two-dimensional image 12 from the camera 4-1 to the camera 4-3.
This is earlier than at the same timing SV3 when the image 12 is taken in from. Thereby, the coordinates (x1, y1) can be matched with the coordinates (x3, y3). Since there is no deviation in the X coordinate, the image 1
2 in the horizontal direction when SH is taken
1 is set to be the same as the same timing SH3 when capturing the image 12 from the camera 4-3.

The coordinates (x2, y2) of the mark M in the image captured from the line sensor camera 4-2 have a smaller X coordinate than the coordinates (x3, y3). This is because the position of the camera 4-2 in the X direction is shifted to the right in FIG. Therefore, the image recognition device 6 sets the horizontal capture timing SH2 when capturing the two-dimensional image 12 from the camera 4-2 earlier than the same timing SH3 when capturing the image 12 from the camera 4-3. Thereby, the coordinates (x2, y2) can be matched with the coordinates (x3, y3). Since there is no deviation in the Y coordinate, the vertical capture timing SV2 when capturing the image 12 from the camera 4-2 is set to be the same as the same timing SV3 when capturing the image 12 from the camera 4-3.

On the other hand, the coordinates (x4, y4) of the mark M in the image captured from the line sensor camera 4-4 have a larger X coordinate than the coordinates (x3, y3). This is because the position of the camera 4-4 in the X direction is shifted to the left in FIG. Therefore, the image recognition device 6 delays the horizontal capture timing SH4 when capturing the two-dimensional image 12 from the camera 4-4 as compared to the same timing SH3 when capturing the image 12 from the camera 4-3. Thereby, the coordinates (x4, y4) can be matched with the coordinates (x3, y3). Since there is no deviation in the Y coordinate, the vertical capture timing SV4 when capturing the image 12 from the camera 4-4 is set to be the same as the same timing SV3 when capturing the image 12 from the camera 4-3.

The coordinates (x5, y5) of the mark M in the image fetched from the line sensor camera 4-5 have a larger Y coordinate than the coordinates (x3, y3). This is because the position of the camera 4-5 in the Y direction is shifted upward in FIG. Therefore, the image recognition device 6 uses the camera 4-5.
The capture timing SV5 in the vertical direction when capturing the two-dimensional image 12 from the camera 4-3 is delayed as compared with the same timing SV3 when capturing the image 12 from the camera 4-3. Thereby, the coordinates (x5, y5) can be matched with the coordinates (x3, y3). Since there is no deviation in the X coordinate, the horizontal capture timing SH5 when capturing the image 12 from the camera 4-5 is set to be the same as the same timing SH3 when capturing the image 12 from the camera 4-3.

As described above, the line sensor camera 4
The mounting error of -1 to 4-5 can be corrected to about the moving error (several μm) of the XY table 3 when the image of the sheet 11 is taken. In the present embodiment, a line sensor camera has been described as an example.
If the number of pixels captured by the image recognition device 6 is smaller than the number of images in the Y direction (if the number of pixels has a margin for adjusting the capturing timing), the present invention is applied to a two-dimensional area sensor. Needless to say, it can be applied.

Also, in the present embodiment, the XY table 3 is sequentially moved in the X direction, and the sheet 11 is sequentially imaged. However, a sheet in which a plurality of marks are arranged at intervals of the pieces 2 is imaged. You may. In this case, the imaging for the position fine adjustment can be completed only once. However,
When the interval between the pieces 2 changes, it is necessary to recreate a sheet corresponding to the change.

[0025]

According to the present invention, by adjusting the horizontal and vertical capturing timings for each camera, even if each camera has an attachment error of, for example, about several tens to several hundreds of micrometers, this attachment error can be prevented. Can be corrected. As a result, there is no need to perform high-precision mechanical adjustment when attaching each camera.

[Brief description of the drawings]

FIG. 1 is a block diagram of a pattern inspection apparatus showing an embodiment of the present invention.

FIG. 2 is a diagram showing a mark for fine-tuning the position of a captured image and an image in which the mark is captured by each line sensor camera.

FIG. 3 is a diagram for explaining adjustment of horizontal and vertical capturing timings by the image recognition device.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Green sheet, 2 ... Piece, 3 ... XY table, 4-1-4-5 ... Line sensor camera, 5 ... Camera table, 6 ... Image recognition device, 7 ... Host computer, 8 ... Display device.

Claims (1)

[Claims] When capturing a plurality of two-dimensional images captured by a plurality of cameras, the capturing timing in the horizontal direction is adjusted for each camera so that the horizontal coordinates of each image coincide with each other. Wherein the vertical capturing timing is adjusted for each camera such that the vertical coordinates of the captured image coincide with each other.