JP2015198290A - Stitching exposure processing correction method and imaging apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a stitching exposure processing correction method which performs correction so as to reduce a level difference for each region in an output value from a produced imaging device even if a difference is generated in a circuit pattern for every exposure by a change of an ambient environment or the like, and an imaging apparatus.SOLUTION: The stitching exposure processing correction method includes initial profile generation processing and profile update processing. In the initial profile generation processing, in pixel output from an imaging device 11 that images a luminance change pattern, a gain value is calculated for each of corresponding pixels in both side regions of a stitching column, and the pixel value and the gain value are stored in a table section 15. In the profile update processing, a gain value of a correction value (c) at a position (x) is then calculated by a calculation section 14 in such a manner that a gradient in a luminance change direction at each pixel position meets a conditional expression (1) regarding each of regions at both sides of the stitching column each time each frame is imaged, and in the case where the gain value exceeds a predetermined threshold, the pixel value of the pixel in the table section 15 is multiplied by the gain value of the correction value (c).

Description

  The present invention relates to a stitching exposure processing correction method and an image pickup apparatus, and more particularly to correction for an image output from an image pickup device manufactured using stitching employed when exposing a large size image sensor. The present invention relates to a stitching exposure processing correction method and an imaging apparatus that perform processing.

  Conventionally, when an IC element such as an image sensor is manufactured using a lithography process, a circuit pattern formed on an original plate such as a mask or a reticle is applied to a resist-coated wafer or glass plate via a projection optical system. A projection exposure apparatus that performs exposure transfer on a substrate (hereinafter referred to as a wafer or the like) is known. As this type of projection exposure apparatus, a step-and-repeat type reduction projection exposure apparatus (so-called stepper) or the like is known as a typical one.

  On the other hand, particularly when the image sensor has a large size (for example, the outer width is 40 mm), depending on the pattern to be formed on the wafer or the like, the exposure apparatus may be more than the range exposed at one time. The pattern area to be exposed may be larger. In such a case, an exposure method by stitching (aperture synthesis) is known in which a pattern to be formed on a wafer or the like is divided into a plurality of parts, and the divided patterns are sequentially exposed and connected to adjacent portions on the wafer or the like. (See Non-Patent Document 1).

A Large Format CMOS Image Sensors (Suat Utku Ay; pp186-187; 7.6 Special CMOS Fabrication Process: Stitching)

  However, in stitching exposure processing, the circuit pattern differs for each exposure due to changes in the surrounding environment and various conditions, so that the formed circuit (particularly the function as an amplifying transistor) is different for each region corresponding to divided exposure. The circuit operation is not uniform. For example, when exposure is performed on both sides of the center line of the image sensor in the horizontal direction (horizontal direction in normal use), the image sensor has different sensitivity in each area and is imaged. The output image may have a luminance level difference (hereinafter, simply referred to as a level difference) in each region.

  The present invention has been made in view of such circumstances, and in the stitching exposure processing at the time of manufacturing the imaging device, a difference occurred in the projection circuit pattern for each exposure due to changes in the surrounding environment and various conditions. To provide a stitching exposure processing correction method and an imaging apparatus that can correct a level difference for each region in an image output from the image sensor even if it is a case. With the goal.

The stitching exposure processing correction method of the present invention includes:
In a manufacturing process, a stitching exposure processing correction method for correcting an image output from an imaging device manufactured using stitching exposure processing,
First, a luminance change pattern whose luminance changes in a predetermined direction is imaged by the imaging device, and the corresponding pixels in each region located on both sides of the stitching column in the image output from the imaging device are compared, and this Store the comparison result in the profile table, perform the initial profile generation process,
Next, for each of the regions located on both sides of the stitching column, for each imaging of a predetermined number of frames, the gradient grad (x, y in the direction in which the luminance of the luminance change pattern changes at each pixel position (x, y). The correction value c (x, y) at the position (x, y) is obtained so that y) satisfies a predetermined arithmetic expression, and the value related to the correction value c (x, y) has a predetermined threshold value. If it exceeds, the pixel value at the corresponding pixel position in the profile table is multiplied by the value related to the correction value c (x, y), and the profile update process is performed.
Here, the “value related to the correction value c (x, y)” means, for example, a gain value in the correction value c (x, y).
Further, in this specification, the “gain value” means a multiplication value (magnification) for an original value (input value) for changing the original value (input value) to a new value (output value).

Further, areas having a predetermined width located on both sides of the stitching column in the luminance change pattern image are set as respective comparison target areas, and a gain value corresponding to the input pixel value is obtained for each pixel value in these comparison target areas. Based on the obtained gain value, it is preferable to regard the gain value for each pixel value in the entire region located on both sides of the stitching column.
Further, in order to obtain the gain value at the correction value c (x, y) at the position (x, y) in the profile update process, before calculating the gain value in the comparison target area, the luminance of the luminance change pattern It is preferable to perform a low-pass filter process in the direction in which the angle changes.

Moreover, it is preferable that the predetermined arithmetic expression is represented by the following conditional expression (1).
grad (x, y) = {grad (x-1, y) + grad (x + 1, y)} / 2 (1)
Further, prior to the initial profile generation process, the output level difference between the regions located on both sides of the stitching column when imaged with light shielding is measured and calculated, and the output level difference calculated and calculated is reduced. It is preferable to perform offset correction.
The predetermined number of frames is preferably 1.

Furthermore, the imaging device of the present invention includes:
In the manufacturing process, an image pickup apparatus that corrects image output from an image pickup device manufactured using stitching exposure processing,
A pixel value comparison / determination unit that compares pixel values output from regions located on both sides of a stitching column in an image output from the image sensor in accordance with imaging of a luminance change pattern whose luminance changes in a predetermined direction;
Profile table initial setting means for storing a comparison determination result of the pixel value comparison determination means in a profile table associated with each pixel value;
For each region that is located on both sides of the stitching column for every predetermined number of frames, the gradient grad (x, y) in the direction in which the luminance of the luminance change pattern changes at each pixel position (x, y). Correction value calculating means for obtaining a correction value c (x, y) at the position (x, y) so as to satisfy a predetermined arithmetic expression;
Threshold value comparing means for comparing the value relating to the correction value c (x, y) calculated by the correction value calculating means with a predetermined threshold value;
When the threshold value comparison unit determines that the value related to the correction value is larger than a predetermined threshold value, the correction value c (x, y) is set to the pixel value at the corresponding pixel position in the profile table. Profile table updating means for performing profile update processing by multiplying the value,
It is characterized by comprising.

According to the stitching exposure processing correction method and the imaging apparatus of the present invention, the pixel output from the imaging device is subjected to correction processing divided into initial profile generation processing and profile update processing, and at predetermined timings. The above update process is performed. As a result, in the case of manufacturing a large-sized image sensor that requires stitching, even if there is a difference in the captured image quality for each of the separately exposed areas at the time of manufacture, Corrections can be made to reduce differences in image quality.
Particularly in the case of moving images, it is necessary to switch and display frame images at a predetermined speed. Therefore, the method and apparatus of the present invention that can quickly respond to changes in ambient temperature and voltage are suitable.

1 is a block diagram illustrating an imaging apparatus according to an embodiment of the present invention. It is the schematic which shows an example of the pattern formed in the original profile production | generation original plate. It is the schematic which shows an example of the image output from the imaging device which imaged the pattern shown in FIG. It is a figure showing the correction table which shows the correspondence of the image pick-up element output value i, the final output value c (i), and a gain value in a profile table. It is the schematic for demonstrating the concept of the gradient used in a correction value calculation means. It is a flowchart for demonstrating the update process of a profile table.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
(Embodiment)
FIG. 1 is a block diagram showing a part of an imaging apparatus 10 according to an embodiment of the present invention. In a signal processing unit 12 provided at a subsequent stage of the imaging element 11, a correction processing unit 13 and a profile calculation unit are provided. 14 shows that the profile table unit 15 and the changeover switch unit 16 are organically connected.

By the way, in the stitching exposure process in the manufacturing process of the image pickup device (image sensor such as CMOS or CCD) 11 described above, a difference occurs in the circuit pattern for each exposure (divided region) due to changes in the surrounding environment and various conditions. Therefore, the circuit operation is not uniform. Therefore, in the imaging device 10 of the present embodiment, imaging is performed using the manufactured imaging device 11, and predetermined correction is performed on each pixel output of the image obtained by imaging, so that each stitching region After the image luminance level difference is reduced, the video signal is output.
The predetermined correction is performed using data that is set in advance in the profile table unit 15 and is sequentially updated.

The level difference of the video signal output for each stitching area is considered to be caused by the offset and the nonlinear gain.
(Level difference due to offset)
Regarding the level difference caused by the offset, the signal output obtained in a state where the image sensor is shielded from light is measured, and the profile calculation unit 14 determines the level difference in the corresponding pixels in the regions on both sides of the stitching column (stitching line). Calculation is performed for each pixel, and a predetermined value is added to each pixel so that the level difference is eliminated in the profile table unit 15.
The correction for reducing the level difference due to the offset is preferably performed prior to the correction for reducing the non-linear level difference described below, but may be performed simultaneously with the correction for reducing the non-linear level difference.

(Non-linear level difference)
On the other hand, among the level differences for each stitching area generated in the stitching exposure process, the level difference of the nonlinear gain characteristic depending on the light amount (particularly caused by changes in the ambient temperature and applied voltage) is caused by various factors. It is difficult to calculate by simple calculation.

Therefore, in the present embodiment, an initial profile generation process for generating an initial profile and a profile update process for sequentially updating the profile for each subsequent frame are performed, and correction is performed using a data table that stores the sequentially updated data. By performing the processing, the level difference of the non-linear gain characteristic depending on the light amount is reduced.
In the following, processing for reducing the level difference of the light amount-dependent nonlinear gain characteristics will be described separately for initial profile generation processing and profile update processing.

A. Initial Profile Generation Processing First, as shown in FIG. 2, a luminance change pattern recorded on an original plate such as a reticle or a glass plate is imaged by the image sensor 11 for a predetermined time. In this luminance change pattern, the luminance gradually changes in a predetermined direction (the vertical direction in FIG. 2).
Note that the pixel output as it is output from the image sensor 11 after the above imaging is expressed as shown in FIG. That is, the areas on both sides of the stitching column (stitching line) 21 are different in luminance from each other, and it is obvious that the image quality deteriorates as it is.

  Next, an average value for each pixel is calculated for several frames in the time axis direction for the captured video. Thereby, random noise components and steep level changes can be suppressed. This calculation process is performed in the profile calculation unit 14.

  Here, the areas of several columns on both the left and right sides of the stitching column 21 are referred to as comparison target areas 22.

Thereafter, for each pixel corresponding to each other in the comparison target area 22 formed on both the left and right sides of the stitching column 21, a gain value for each input pixel value is calculated and compared.
Based on the gain value at each pixel in the comparison target area 22, the gain value for the pixel at another location in the area is also calculated by using a process such as interpolation for a pixel value that has not yet been calculated (each area It is assumed that there is a predetermined correlation between the pixels in the image). These calculation processes are performed in the profile calculation unit 14.

  Next, the profile table unit 15 creates a data table in which the relationship between each pixel value and the corresponding gain value is associated. The profile table unit 15 stores correspondence data for each region for two regions if stitching is performed once, and for three regions if stitching is performed twice. Among the correspondence data, each pixel value and a gain value at the pixel value are sent to the correction processing unit 13.

B. Profile update processing Temperature changes, voltage fluctuations, and the like may occur during image capture execution. In response to this, the profile table section 15 is updated for each frame. Thereby, the correction process corresponding to the situation rapidly can be performed.

  In the above update process, the luminance change direction (vertical direction in FIG. 3: the direction in which the luminance of the luminance change pattern changes (hereinafter referred to as the luminance change pattern) for the purpose of reducing random noise components and steep level changes in the comparison target area 22. It is preferable to apply noise cut filter processing (low-pass filter processing), for example, addition average filter processing to the same)). As a result, the correlation between both regions sandwiching the stitching column can be enhanced.

Next, in the region where correction is performed, the gradient grad (x, y) in the luminance change direction at the pixel position (x, y) across the stitching column is
grad (x, y) = (grad (x-1, y) + grad (x + 1, y)) / 2 (1)
A gain value for the correction value c (x, y) of the position (x, y) is obtained so as to satisfy the following conditional expression (1).

  FIG. 5 conceptually represents the above equation (1). In each position of the curve representing the change of the pixel value, the gradient (grad (x−1, y) and This shows that the slope grad (x, y) can be obtained by calculating the average value of the slope grad (x + 1, y), that is, information on the areas on both sides of the stitched line. Based on this, the inclination on the stitching line is obtained and used as input information to the correspondence table, whereby the correlation between the two regions can be successively improved.

Next, the obtained gain value is compared with a predetermined threshold value, and if it exceeds the threshold value, the stored data at the corresponding position in the profile table unit 15 is multiplied by this gain value to obtain a profile table. The update process ends.
Actually, the gain value at the corrected value c ′ (x, y) of the current frame is divided by the gain value at the value c (x, y) of the previous frame, and the result is the above threshold. Must be compared with the value.

(Configuration of imaging device)
Next, the configuration of the imaging apparatus 10 that performs the initial profile generation process and the profile update process will be described in detail along the flow of the process with reference to FIG. 1 again.

  First, in the initial profile generation process, the pixel output from the image sensor 11 is input to the changeover switch unit 16 and also to the profile table unit 15.

  At the time of initial profile generation processing, since the changeover switch unit 16 is connected to the terminal a, the pixel output from the image sensor 11 (first image frame) is output from the terminal a of the changeover switch unit 16 as a video signal. Are input to the profile calculation unit 14 and directly output to the outside without passing through the correction processing unit 13.

In the profile calculation unit 14, the input pixel output is calculated, and the numerical value calculated thereby is sent to the profile table unit 15.
The profile table unit 15 stores a new correspondence between each pixel value and a corresponding gain value based on the input pixel output and the numerical value calculated by the profile calculation unit 14. For example, as shown in FIG. 4, the correspondence relationship is such that the pixel output value i (image sensor output value i) from the image sensor 11 and the pixel output value (correction value) c (i) (table from the profile table unit 15. Partial output value c (i)) and gain value c (i) / i.

  Thereafter, each pixel value and the gain value at the pixel value are sent to the correction processing unit 13. The correction processing unit 13 corrects the pixel output for the next image frame based on the relationship between each pixel value and the gain value in the pixel value.

  Even during the profile update process, the pixel output from the image sensor 11 is input to the changeover switch unit 16 and also to the profile table unit 15.

  At the time of the profile update process, since the changeover switch unit 16 is connected to the terminal b, the pixel output from the image sensor 11 is input from the changeover switch unit 16 to the correction processing unit 13. Immediately before that, each pixel value input from the profile table unit 15 and the gain value in the pixel value are corrected, and output from the correction processing unit 13 to the outside as a video signal output. The output value from the correction processing unit 13 is also input to the profile calculation unit 14 and the profile table unit 15, respectively.

  On the other hand, in the profile table unit 15, based on the pixel output input to the profile table unit 15, the numerical value calculated by the profile calculation unit 14, and the output value from the correction processing unit 13, each pixel value and its corresponding A new correspondence with the gain value to be stored is stored. In this storing process, the gain value of the correction value c (x, y) for the position (x, y) that can satisfy the conditional expression (1) has been determined, and the gain value has a predetermined threshold. Based on the determination that the value has been exceeded, the numerical value at a predetermined position in the profile table unit 15 is multiplied by the new gain value.

  Thereafter, each pixel value and the gain value at the pixel value are sent to the correction processing unit 13. The correction processing unit 13 corrects the pixel output for the next image frame based on the relationship between each pixel value and the gain value in the pixel value, and outputs the corrected pixel signal to the outside as a video signal output.

  Next, a part of the stitching exposure processing correction method according to the present embodiment will be described with reference to the flowchart shown in FIG.

First, as an initial position (x, y), a stitching column position is set in x, and 0 is set in y (S10).
Note that the various calculation processes in step 10 (S10) and the following steps are performed in the profile calculation unit 14.
Next, the gain value of the correction value (output value from the correction processing unit 13) c (x, y) for the previous image frame with respect to the image sensor output value i at the (x, y) position (hereinafter, simply referred to as “j”). c (x, y)) is acquired (S11).

Next, the gradient grad (x, y) in the y direction (vertical direction in FIG. 3) at the c (x, y) position is calculated (S12).
Next, a search is made for a gain value (hereinafter simply referred to as c ′ (x, y)) of the correction value c ′ (x, y) that satisfies the conditional expression (1) (S13).

Next, c (x, y) is divided by c '(x, y) to obtain a magnification (gain magnification: the same applies hereinafter) G (S14).
Next, it is compared whether or not the magnification G obtained in step 14 (S14) exceeds a predetermined threshold value (S15). As a result of the comparison, if it is determined that the magnification G exceeds a predetermined threshold value, the output value (correction value) for i is updated in the profile table unit 15 (S16), and the target setting position (x , y) is updated (y is incremented) (S17). As a result of the comparison, if it is determined that the magnification G is equal to or less than the predetermined threshold value, step 16 (S16) is omitted and the target setting position (x, y) in step 17 (S17) is updated (y is incremented). (S17).

  Thereafter, it is determined whether or not y at the target setting position (x, y) has been updated to the final line (S18). If y is not updated to the final line as a result of the determination, the process returns to step 11 above. The subsequent steps are repeated, and if y has been updated to the last line, this flow processing is terminated.

  The stitching exposure processing correction method and the imaging apparatus according to this embodiment have been described above. However, the stitching exposure processing correction method and the imaging apparatus according to the present invention are not limited to those described in the above embodiment. However, other various modifications can be made.

  For example, in the above-described embodiment, it is described that stitching is performed only in one direction in two dimensions, but stitching may be performed in two orthogonal directions. In this case, it is preferable that the pattern for generating the initial profile has gradation in two orthogonal directions.

  In the above embodiment, the profile update process is performed for each frame, but may be appropriately performed at other timings (for example, every two frames).

In the above embodiment, the gain value at the correction value c (x, y) at the position (x, y) is obtained using the conditional expression (1). An equation can be used.
For example, the gradient grad (x, y) at pixel position (x, y) is changed to grad (x-2, y), (grad (x-1, y), (grad (x + 1, y) and grad You may make it obtain | require by the average value of the gradient of (x + 2, y).

DESCRIPTION OF SYMBOLS 10 Imaging device 11 Image sensor 12 Signal processing means 13 Correction processing part 14 Profile calculation part 15 Profile table part 16 Changeover switch part 21 Stitching column 22 Comparison object area (enclosed part)

Claims (7)

In a manufacturing process, a stitching exposure processing correction method for correcting an image output from an imaging device manufactured using stitching exposure processing,
First, a luminance change pattern whose luminance changes in a predetermined direction is imaged by the imaging device, and the corresponding pixels in each region located on both sides of the stitching column in the image output from the imaging device are compared, and this Store the comparison result in the profile table, perform the initial profile generation process,
Next, for each of the regions located on both sides of the stitching column, for each imaging of a predetermined number of frames, the gradient grad (x, y in the direction in which the luminance of the luminance change pattern changes at each pixel position (x, y). The correction value c (x, y) at the position (x, y) is obtained so that y) satisfies a predetermined arithmetic expression, and the value related to the correction value c (x, y) has a predetermined threshold value. A stitching exposure processing correction method, wherein when it exceeds, the pixel value at the corresponding pixel position in the profile table is multiplied by a value related to the correction value c (x, y), and the profile update processing is performed.   A region having a predetermined width located on both sides of the stitching column in the image of the luminance change pattern is set as each comparison target area, and a gain value corresponding to the input pixel value is obtained for each pixel value of these comparison target areas. 2. The stitching exposure processing correction method according to claim 1, wherein the stitching exposure processing correction method is regarded as a gain value relating to each pixel value of all regions located on both sides of the stitching column on the basis of the obtained gain value.   In order to obtain a gain value for the correction value c (x, y) of the position (x, y) in the profile update process, the luminance of the luminance change pattern changes before calculating the gain value in the comparison target area. 3. The stitching exposure processing correction method according to claim 2, wherein low-pass filter processing is performed in a direction in which the stitching exposure processing is performed. The stitching exposure processing correction method according to claim 1, wherein the predetermined arithmetic expression is represented by the following conditional expression (1).
grad (x, y) = {grad (x-1, y) + grad (x + 1, y)} / 2 (1)   Prior to the initial profile generation process, the output level difference of each region located on both sides of the stitching column when imaged with light shielding is measured and calculated, and the offset is calculated so as to reduce the measured output level difference. 5. The stitching exposure processing correction method according to claim 1, wherein correction is performed.   6. The stitching exposure processing correction method according to claim 1, wherein the predetermined number of frames is one. In the manufacturing process, an image pickup apparatus that corrects image output from an image pickup device manufactured using stitching exposure processing,
A pixel value comparison / determination unit that compares pixel values output from regions located on both sides of a stitching column in an image output from the image sensor in accordance with imaging of a luminance change pattern whose luminance changes in a predetermined direction;
Profile table initial setting means for storing a comparison determination result of the pixel value comparison determination means in a profile table associated with each pixel value;
For each region that is located on both sides of the stitching column for every predetermined number of frames, the gradient grad (x, y) in the direction in which the luminance of the luminance change pattern changes at each pixel position (x, y). Correction value calculating means for obtaining a correction value c (x, y) at the position (x, y) so as to satisfy a predetermined arithmetic expression;
Threshold value comparing means for comparing the value relating to the correction value c (x, y) calculated by the correction value calculating means with a predetermined threshold value;
When the threshold value comparison unit determines that the value related to the correction value is larger than a predetermined threshold value, the correction value c (x, y) is set to the pixel value at the corresponding pixel position in the profile table. Profile table updating means for performing profile update processing by multiplying the value,
An imaging apparatus comprising: