CN101312537A - Drive method for solid-state imaging device, solid-state imaging device, and imaging apparatus - Google Patents

Abstract

A driving method for a solid-state imaging device having a slanted pixel pattern, comprising the steps of: in an area having n pixels adjacent in the horizontal direction and n pixels adjacent in the vertical direction, for odd-numbered rows and For an even-numbered row, add x pixels in the horizontal direction and y pixels in the vertical direction that have the same color, where n is an odd number of three or greater and n≥x ≥y; and repeatedly add the x pixels and the y pixels, and simultaneously move the n×n region by m pixels in the vertical or horizontal direction, where m is an odd number of three or more. The n×n area of the odd-numbered row is shifted by m pixels relative to the n×n area of the even-numbered row in the oblique direction of the oblique pixel pattern.

Description

Solid-state imaging device driving method, solid-state imaging device, and imaging apparatus

This application is a divisional application of an invention patent application with an application date of December 27, 2005, an application number of 200510135801.X, and an invention title of "a solid-state imaging device driving method, a solid-state imaging device, and an imaging device".

Cross References to Related Applications

The present invention contains subject matter related to Japanese Patent Application JP 2004-375405 filed in Japan Patent Office on December 27, 2004, Japanese Patent Application JP 2005-107034 filed in Japan Patent Office on April 4, 2005 , and Japanese Patent Application JP 2005-107037 filed in the Japan Patent Office on April 4, 2005, the entire contents of which are hereby incorporated by reference.

technical field

The present invention relates to a driving method of a solid-state imaging device, a solid-state imaging device, and an imaging apparatus. More specifically, the present invention relates to a driving method for an X-Y addressing solid-state imaging device (a typical example of which is a complementary metal oxide semiconductor (CMOS) device image sensor), a solid-state imaging device realizing the above driving method, and a solid-state imaging device using the solid-state imaging device. An imaging device of an imaging device.

The present invention also relates to solid-state imaging devices and imaging devices, and more particularly to color filters in which a color filter having a primary color component for generating a luminance (Y) component and other color components is arranged on the surface of a pixel A solid-state imaging device, and also an imaging device using a solid-state imaging device as an imaging means.

Background technique

As disclosed in, for example, Japanese Unexamined Patent Application Publication No. 2004-266369, in order to increase the frame rate of a solid-state imaging device, generally, the amount of pixel information is reduced by adding information on a plurality of pixels.

Examples of the above techniques are as follows. In the color coding of the Bayer pattern shown in Fig. 1, pixels of the same color in two columns and two rows are extracted and added from a 3×3 pixel area, and at the same time, by Changing the color space repeating pattern keeps the original pixel pattern without changing the pixel pitch ratio in the vertical, horizontal, and oblique directions, while shifting the 3×3 pixel area by three pixels.

Red (R) pixels 311 , 313 , 331 , and 333 located in odd-numbered rows are added, and then, the resulting added R signal is positioned at centroid A. Similarly, green (G) pixels 314, 316, 334, and 336 are summed by shifting three pixels horizontally from R pixels 311, 313, 331, and 333, and then, the resulting added G signal is positioned in the shape Heart B. The R signals 317, 319, 337, and 339 are summed by further shifting three pixels horizontally from the G pixels 314, 316, 334, and 336, and then the resulting added R signal is positioned at the centroid C.

Then, by vertically shifting three pixels from the R pixels 311, 313, 331, and 333, the G pixels 341, 343, 361, and 363 located in the even-numbered rows are summed, and then, the resulting added G signal is positioned At the centroid D. The blue (B) pixels 344, 346, 364, and 366 are summed by shifting three pixels horizontally from the G pixels 341, 343, 361, and 363, and the resulting added B signal is positioned at the centroid E. In this way, by summing color pixels as described above over the entire pixel area, pixels of the same color can be summed while maintaining the original color pattern without changing the color space repeat pattern or changing the vertical , horizontal, and pixel pitch ratios in the oblique directions.

Among imaging devices such as digital still cameras and video cameras, the number of pixels of solid-state imaging devices used as imaging means is increasing, and solid-state imaging devices having several million pixels are gradually being widely used. The use of multi-pixel imaging devices is aimed at obtaining high-resolution images. However, there remains a need for solid-state imaging devices demonstrating higher resolution.

In a single-panel digital camera, the color pattern of a color filter used in a solid-state imaging device is very important for obtaining high resolution. A typical example of this color pattern is the well known and widely used Bayer pattern.

Bayer pattern

The Bayer pattern is a color pattern as shown in FIG. 2 in which GR lines alternately having G pixels and R pixels and GB lines alternately having G pixels and B pixels are alternately arranged in the horizontal direction (also in the vertical direction) . This Bayer pattern is characterized in that the pixels are arranged in a square lattice (square lattice) at regular pixel intervals d (pixel pitch) in the vertical and horizontal directions; and the G:R ratio of the GRB color in this square lattice pattern is :B ratio is 2:1:1.

The spatial frequency characteristics of the RGB colors in the Bayer pattern will now be described by separately considering the characteristics of the G color and other colors (ie, R and B colors) which are the primary colors for generating the luminance (Y) component.

In general, the luminance signal Y is generated according to equation (1).

Y＝0.6G+0.3R+0.1B (1)

Equation (1) is based on the fact that the human eye is more sensitive to G color and less sensitive to R and B colors. That is, if the luma (Y) component requires high resolution, it is important to increase the resolution of the G color component, while the other R and B color components require less high resolution.

3A and 3B illustrate a G pattern in which only G pixels are extracted from the Bayer pattern. Now consider the spatial frequency characteristics in conjunction with Figures 3A and 3B. If the pixel sampling rate (sampling rate) is set to the pixel pitch d, the sampling rate of G pixels is equal to the pixel pitch d in the vertical and horizontal directions, and, according to the sampling theorem, the highest frequency (1/2) fs can be collected (fs (=1/d): the signal component of the sampling frequency). That is, according to the theoretical threshold it is possible to collect the signal components indicated by the half-tone column (half-tone column) and the empty column (voided column) shown in Figure 3A, but it is impossible to collect higher frequencies with frequencies exceeding this threshold signal components.

所以根据抽样定理可以收集最高的信号分量。As for the 45° oblique direction, since the sampling rate of G pixels is

Therefore, according to the sampling theorem, the highest signal components.

Similarly, the spatial frequency characteristics of R and B pixels are considered below. In this case, since the pixel pitches of the R and B pixels are the same, only the spatial frequency characteristic of the R pixel will be described below.

这样，根据抽样定理可以收集频率最高

的信号分量。3C and 3D show R patterns in which only R pixels are extracted from the Bayer pattern. For the spatial frequency characteristics of R pixels, since the sampling rate of R pixels in the vertical and horizontal directions is 2d, the signal components with the highest frequency of 1/4fs can be collected according to the sampling theorem. In the oblique direction, the sampling rate of R pixels is

In this way, according to the sampling theorem, the highest frequency can be collected

signal components.

In FIGS. 3A to 3D , threshold frequency components that can be collected in vertical, horizontal, and oblique directions are indicated by empty columns and halftone columns.

的分辨率，并且，R像素的空间频率特性展示在垂直和水平方向上最高1/4fs的分辨率和在倾斜45°的方向上最高

的分辨率，即，可以收集最高上述阈值频率的信号分量。The spatial frequency characteristics of G, R, and B pixels are shown in FIG. 4 . Fig. 4 shows that when the sampling rate is set to the pixel pitch d (= 1/fs), the spatial frequency characteristic of G pixels exhibits a resolution of up to 1/2 fs in the vertical and horizontal directions and in the direction inclined by 45° on the highest

, and the spatial frequency characteristics of R pixels exhibit a resolution of up to 1/4fs in the vertical and horizontal directions and a resolution of up to 45° in the

resolution, i.e., signal components up to the above-mentioned threshold frequency can be collected.

Bayer pixel shift pattern

In addition to the above-mentioned Bayer pattern, as disclosed in Japanese Unexamined Patent Application Publication No. 10-262260, patterns shown in FIGS. 3A to 3D such as those shown in FIGS. 6A to 6D have been proposed. A Bayer pattern shifted by 45°, that is, a trimmed Bayer pattern in which pixels are arranged vertically and horizontally shifted by half a pixel pitch.

其是拜尔图案的抽样率的

倍，所以可以获得比拜尔图案的分辨率更高的分辨率。A color pattern generated by shifting the Bayer pattern by 45° is hereinafter referred to as a "Bayer pixel shifted pattern". In this Bayer pixel shifting pattern, since the sampling rate turns out to be

which is the sampling rate of the Bayer pattern

times, so a higher resolution than that of the Bayer pattern can be obtained.

From another point of view, if the same resolution is required in the Bayer pattern and the Bayer pixel shift pattern, the sampling rate of the Bayer pixel shift pattern can be increased by times. In other words, by using the Bayer pixel shift pattern, the same resolution can be obtained with a smaller number of pixels than in the Bayer pattern. As a result, the pixel aperture can be increased, so that the light sensitivity of the pixel can be increased, thereby obtaining a signal with a high signal-to-noise (S/N) ratio.

However, Bayer shifting patterns can only exhibit high resolution for achromatic subjects. The reason for this is as follows.

Figure 5 illustrates the color coding of a Bayer pixel shift pattern.

6A and 6B show G patterns in which only G pixels are extracted from the Bayer shift pattern. Since the sampling rate of G pixels in the vertical and horizontal directions is It is larger than the sampling rate d of G pixels in the Bayer pattern in the vertical and horizontal directions, so the resolution in the Bayer pixel shift pattern is lower than that in the Bayer pattern. On the other hand, since the sampling rate d of the G pixel in the oblique direction of 45° is higher than that in the oblique direction of 45° in the Bayer pattern Small, so the resolution is higher than in the Bayer pattern.

Similarly, consider the resolution of R pixels and B pixels. Since the pixel pitches of R pixels and B pixels are the same, only the resolution of R pixels will be described below.

并且R像素在倾斜方向上的抽样率是2d。6C and 6D show R patterns in which only R pixels are extracted from the Bayer shift pattern. The sampling rate of R pixels in the vertical and horizontal directions is

And the sampling rate of R pixels in the oblique direction is 2d.

In FIGS. 6A to 6D , threshold frequency components that can be collected in vertical, horizontal, and oblique directions can be indicated by empty columns and halftone columns.

The spatial frequency characteristics of G, R, and B pixels are shown in FIG. 7 . When comparing FIG. 7 and FIG. 4 , it can be seen that the spatial frequency characteristics of the Bayer pixel shift pattern are the same as those shifted by 45° from that of the Bayer pattern.

Contents of the invention

In order to improve the effective integrity of pixels including photoelectric transducers, some solid-state imaging devices use the next so-called "slanted pixel pattern". In this oblique pixel pattern, from the matrix pixel pattern, even-numbered column pixels are shifted in the column direction relative to odd-numbered column pixels by about 1/2 of the pixel pitch, and even-numbered row pixels are shifted row-wise Upwards are shifted about 1/2 the pixel pitch relative to odd-numbered row pixels. When a color filter is arranged on a solid-state imaging device having such an oblique pixel pattern, as shown in FIG. 8 , the color coding of the Bayer pattern is shifted by 45°.

In a CMOS image sensor having an oblique pixel pattern, when pixel signals are line-sequentially read, in a pixel region 101 in which pixels 100 are arranged obliquely, as shown in FIG. 9, each of them is driven by a vertical selection circuit 106 The lines are all connected to the horizontal pixel driving line group 105 of the pixel 100 with two zigzag lines, and the signal of the pixel 100 of the zigzag line selected via the horizontal pixel driving line group 105 is stored in the column processing circuit 103, each The column processing circuits 103 are arranged for one column via the group of vertical signal lines 102 , and each vertical signal line is arranged for one pixel column. Then, the signals of the pixels 100 stored in the column processing circuit 103 are sequentially read out to the horizontal signal line 108 via the horizontal selection switch group 107 which is sequentially selected by the horizontal selection circuit 104 .

In this reading method, since many pixel signals can be read by one reading operation, the reading speed is fast, but on the other hand, it is necessary to read out the pixel signals in two adjacent rows at the same time, and its flexibility is more Small. Therefore, when pixel addition is performed in the color coding of the oblique pixel pattern shown in FIG. 8 generated by shifting the Bayer pattern by 45°, unlike pixel addition in the color coding of the Bayer pattern, the addition The resulting color pattern of the signal becomes different from the original color pattern, while it is found difficult to maintain the same color space repeating pattern and the same pitch ratio in vertical, horizontal, and oblique directions.

In yet another reading method in a CMOS image sensor having an oblique pixel pattern, as shown in FIG. The horizontal pixel driving line group 205 arranged for one pixel row, and the signal of the pixel 200 of the row selected via the horizontal pixel driving line group 205 is stored in the column processing circuit 203, and each column processing circuit 203 passes through the vertical signal line group 202 are arranged in two zigzag columns, and each signal line is connected to the pixel 200 in the same two zigzag columns. Then, the signals of the pixels 200 stored in the column processing circuit 203 are sequentially read out to the horizontal signal line 208 via the horizontal switch group 207 which is sequentially selected by the horizontal selection circuit 204 .

In this reading method, since only pixel signals can be read out line by line, it is difficult to achieve a fast reading operation. In addition, pixel signals in adjacent odd-numbered and even-numbered rows are read out via the same vertical signal line in the vertical signal line group 202 and processed in the same column circuit 203 . Therefore, when pixel addition is performed in color coding by a pixel pattern as shown in FIG. Keep the same color space repeat pattern and the same pitch ratio in vertical, horizontal, and oblique directions.

Differences between the Bayer pixel shift pattern and the Bayer pattern, which are typical examples of known color patterns, are described below.

In the Bayer pixel shifting pattern, since the sampling rate is the sampling rate of the Bayer pattern times, so as long as the Bayer pixel shift pattern is used for the non-color object, the pixel information of the Bayer pattern can be obtained twice as much as the pixel information. That is, higher resolutions can be achieved in Bayer pixel shift patterns. In other words, the use of the Bayer pixel shift pattern exhibits the same resolution as that of a Bayer pixel with a smaller number of pixels, which makes it possible to increase the aperture of the pixel, thereby increasing the light sensitivity of the pixel, that is, the S/N ratio .

The sampling rate of the Bayer pixel shift pattern in the vertical and horizontal directions is larger than that of the Bayer pattern only from the viewpoint of G pixels for generating a primary component of the luminance (Y) component. This means that the resolution of the G pixels in the Bayer pattern in the vertical and horizontal directions is higher than the corresponding resolution of the Bayer pixel shift pattern. In other words, the Bayer pixel shift pattern is inferior to the Bayer pattern as long as attention is paid to the resolution of the G pixels in the vertical and horizontal directions.

从而在所有的垂直、水平、和倾斜方向上实现比拜尔图案更高的分辨率。To overcome this, when imaging non-color objects, the RGB balance is adjusted in the camera signal processing system, ie gains are applied so that the RGB levels become the same. Then, assuming that the R and B levels are equal to the G level, the luminance (Y) component is generated, and the sampling rate of the Y component is processed as

This results in higher resolution than Bayer patterns in all vertical, horizontal, and oblique directions.

However, the above processing is effective only for achromatic objects, and it is difficult to obtain high resolution if the same processing is performed on chromatic objects. Also, when the level balance is deviated, if processing is performed assuming that the R and B levels are equal to the G level, it is difficult to perform correct interpolation processing in the camera signal processing system, resulting in the occurrence of wrong colors.

In view of the above-mentioned background, it is necessary to provide a driving method for a solid-state imaging device, a solid-state imaging device, and an imaging device in which, after adding pixels in an oblique pixel pattern, the original color pattern can be maintained without changing Color space repeating pattern or pitch ratio in vertical, horizontal, and oblique directions.

It is also necessary to provide a solid-state imaging device and an imaging device that realize high resolution for both achromatic objects and colored objects without causing false colors.

According to an embodiment of the present invention, there is provided a driving method for a solid-state imaging device having an oblique pixel pattern in which pixels are arranged obliquely. The driving method includes the following steps: in an area having n pixels adjacent in the horizontal direction and n pixels adjacent in the vertical direction, for odd-numbered rows and even-numbered rows, x pixels in the horizontal direction Add y pixels in the vertical direction, the x pixels and y pixels have the same color, where n is an odd number of three or more and n≥x≥y; and the x pixels in the horizontal direction and y pixels in the vertical direction are repeatedly added, and at the same time, in the vertical or horizontal direction, the area with adjacent n pixels in the horizontal direction and n pixels adjacent in the vertical direction is moved by m pixels, where m is three or a larger odd number. Between the area having n pixels adjacent in the horizontal direction and n pixels adjacent in the vertical direction of the odd-numbered lines and the area having n pixels adjacent in the horizontal direction and n pixels adjacent in the vertical direction of the even-numbered lines The spatial positional relationship among them is such that they are shifted by m pixels from each other in the oblique direction of the oblique pixel pattern.

By adding pixels as described above, the original color pattern can be maintained without changing the color space repeat pattern or the pitch ratio in the vertical, horizontal, and oblique directions.

According to yet another embodiment of the present invention, there is provided a solid-state imaging device including: a pixel including photoelectric transducers two-dimensionally arranged in a matrix; and a color filter including component of the primary color component and other color components, the color filter is arranged on the surface of the pixel. The primary color components and other color components are arranged such that the primary color components surround the other color components. This solid-state imaging device can be used as an imaging device in an imaging device such as a digital still camera or a video camera.

The above-configured solid-state imaging device or an imaging device using the solid-state imaging device as an imaging means has a color pattern in which a color component (for example, a G color component) that is a main component for generating a luminance component surrounds other color components ( such as R and B color components). In this color pattern, G components exist in all rows and all columns. Therefore, the spatial frequency characteristics of the G component to which human eyes have higher sensitivity can be increased, and processing for adjusting the level balance of the RGB components becomes unnecessary.

According to an embodiment of the present invention, in a solid-state imaging device having an oblique pixel pattern, after adding pixels, the original color pattern can be maintained without changing the color space repeating pattern or the pitch in the vertical, horizontal, and oblique directions. distance ratio. As a result, by spatially equalsampling, high-quality addition signals can be obtained. At the same time, the added signal can be output in the same color pattern as that obtained by the progressive (full pixel) reading method for independently reading pixel signals, thereby facilitating the subsequent stage. signal processing.

In addition, it is possible to increase the spatial frequency characteristics of a color component that is a main component for generating a luminance component. In this way, high resolution can be obtained not only for achromatic objects but also for colored objects. Furthermore, processing for adjusting the level balance of RGB pixels becomes unnecessary, thereby preventing the occurrence of wrong colors.

Description of drawings

Figure 1 illustrates the color coding of the Bayer pattern;

Figure 2 illustrates the color coding of the Bayer pattern shown in Figure 1;

3A to 3D illustrate the relationship between patterns of individual colors in a Bayer pattern and sampling rates;

FIG. 4 illustrates frequency characteristics of a Bayer pattern;

Figure 5 illustrates the color coding of the Bayer pixel shift pattern;

6A to 6D illustrate the relationship between the pattern of the individual colors in the Bayer pixel shift pattern and the sampling rate;

FIG. 7 illustrates frequency characteristics of a Bayer pixel shift pattern;

Figure 8 illustrates the color coding of a color pattern shifted 45° from a Bayer pattern;

FIG. 9 is a schematic diagram illustrating an example of a configuration of a CMOS image sensor having an oblique pixel pattern;

10 is a block diagram illustrating still another example of the configuration of a CMOS image sensor having an oblique pixel pattern;

FIG. 11 illustrates the idea of adding pixels by the driving method according to the first embodiment of the present invention;

FIG. 12 illustrates the idea of adding pixels by a driving method according to a second embodiment of the present invention;

13 illustrates the idea of adding pixels by a driving method according to a third embodiment of the present invention;

FIG. 14 illustrates the idea of adding pixels by a driving method according to a fourth embodiment of the present invention;

FIG. 15 illustrates the idea of adding pixels by a driving method according to a fifth embodiment of the present invention;

16 is a schematic diagram illustrating a configuration of a CMOS image sensor according to an embodiment of the present invention;

17 is a circuit diagram illustrating an example of the configuration of an odd-numbered column processing circuit;

FIG. 18 is a timing chart illustrating the operation of the CMOS image sensor shown in FIG. 16;

19 is a block diagram illustrating an example of the configuration of an imaging device according to an embodiment of the present invention;

20 is a block diagram illustrating an example of the configuration of an imaging device according to an embodiment of the present invention;

21 illustrates color patterns of color filters according to a sixth embodiment of the present invention;

22A and 22B illustrate G patterns in which only G pixels are extracted from the color patterns of the sixth embodiment;

22C and 22D illustrate R patterns in which only R pixels are extracted from the color patterns of the sixth embodiment;

FIG. 23 illustrates spatial frequency characteristics of color patterns according to the sixth embodiment;

24 illustrates color patterns of color filters according to a seventh embodiment of the present invention;

25A and 25B illustrate G patterns in which only G pixels are extracted from the color patterns of the seventh embodiment;

25C and 25D illustrate R patterns in which only R pixels are extracted from the color patterns of the seventh embodiment;

FIG. 26 illustrates spatial frequency characteristics of a color pattern according to a seventh embodiment;

27 illustrates color patterns of color filters according to an eighth embodiment of the present invention;

28A and 28B illustrate G patterns in which only G pixels are extracted from the color patterns of the eighth embodiment;

28C and 28D illustrate R patterns in which only R pixels are extracted from the color patterns of the eighth embodiment;

FIG. 29 illustrates spatial frequency characteristics of color patterns according to the eighth embodiment;

30A illustrates comparison results between spatial frequency characteristics of G pixels and spatial frequency characteristics of known color patterns according to sixth to eighth embodiments;

30B illustrates comparison results between the spatial frequency characteristics of R and B pixels and the spatial frequency characteristics of known color patterns according to the sixth to eighth embodiments;

Figure 31 illustrates the characteristics of a band-limited low-pass filter (LPF); and

Fig. 32 illustrates the idea of the decimation process.

Detailed ways

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

The driving method for performing pixel addition in the solid-state imaging device is discussed below through the description of the first to fifth embodiments.

It is now assumed that the solid-state imaging device using this driving method discussed below has the next oblique pixel pattern. In this pixel pattern, many pixels including photoelectric converters are arranged in a matrix, and even-numbered column pixels are shifted from odd-numbered column pixels by about 1/2 of the pixel pitch in the column direction, and even-numbered Numbered row pixels are shifted in the row direction by approximately 1/2 of the pixel pitch from odd-numbered row pixels.

In the next embodiment of the present invention, in the solid-state imaging device, pixels including photoelectric converters are arranged two-dimensionally, and colors having as primary color components for generating luminance (Y) components are arranged on the surface of the pixels Color filters for component and other colors. In this solid-state imaging device, the color pattern of the color filter is very important.

Therefore, in the next three embodiments, the color patterns of the color filters are mainly described. The solid-state imaging device having the color filter according to each of the three embodiments may be: a charge-transfer solid-state imaging device, a typical example of which is a charge-coupled device (CCD) solid-state imaging device; or an X-Y addressing solid-state imaging device, which A typical example is a MOS solid-state imaging device.

In the description that follows, it is assumed that, in a color filter, a color component that is a primary color for generating a luminance (Y) component is a G component, and other color components are R and B components.

However, in the present invention, these color components are merely examples, for example, white, cyan, and yellow components may be used as primary colors for generating the Y component, and, for example, purple, cyan, and yellow components may be used as other color components.

In the above-described solid-state imaging device having an oblique pixel pattern, image addition can be performed while maintaining the original color pattern without changing the color space repetition pattern or the pitch ratio in vertical, horizontal, and oblique directions.

first embodiment

FIG. 11 illustrates the idea of adding pixels by the driving method according to the first embodiment of the present invention. In color coding of a pixel pattern shifted by 45° from a Bayer pattern according to an oblique pixel pattern, from a 3×3 pixel area, pixels of the same color in two columns and two rows are extracted and added while the 3×3 The 3-pixel area is moved by three pixels, that is, three pixels are used as a unit. The process of such pixel addition will be discussed in detail below.

In FIG. 11 , the R pixels 111 , 113 , 151 , and 153 located in odd-numbered rows are added, and then, the resulting added R signal is positioned at the centroid A. Similarly, B pixels 114 , 116 , 154 , and 156 are summed by shifting three pixels horizontally from R pixels 111 , 113 , 151 , and 153 , and the resulting added B signal is positioned at centroid B. The R signals 117 , 119 , 157 and 159 are then summed by shifting three pixels further horizontally from the B pixels 114 , 116 , 154 and 156 , and the resulting summed R signal is positioned at the centroid C.

G pixels 142, 144, 182, and 184 located in even-numbered rows are summed by shifting three pixels from the R signals 117, 119, 157, and 159 by tilting, and the resulting added G signal is positioned in the Heart D. G pixels 145 , 147 , 185 , and 187 are summed by shifting three pixels horizontally from G pixels 142 , 144 , 182 , and 184 , and the resulting added G signal is positioned at centroid E. FIG.

In this way, by adding color pixels as described above over the entire pixel area, pixels of the same color can be added while maintaining the original color pattern without changing the color space repeating pattern or the vertical, horizontal, and The pitch ratio in the oblique direction. In this oblique pixel pattern, it is necessary that the number of columns of pixels corresponding to odd-numbered rows is different from that of pixels corresponding to even-numbered rows. As a result, by spatially equal sampling, high-quality addition signals can be obtained. Meanwhile, the added signal can be output in the same color pattern as that obtained by the progressive (full pixel) reading method for independently reading pixel signals, thereby facilitating signal processing in a subsequent stage.

second embodiment

FIG. 12 illustrates the idea of adding pixels by a driving method according to a second embodiment of the present invention. In color coding of a pixel pattern shifted by 45° from a Bayer pattern according to an oblique pixel pattern, from a 5×5 pixel area, pixels of the same color in three columns and three rows are extracted and added, while the 5×5 A 5 pixel area is shifted by three pixels. The process of such pixel addition will be discussed in detail below.

In FIG. 12 , R pixels 211 , 213 , 215 , 251 , 253 , 255 , 291 , 293 , and 295 located in odd-numbered rows are added, and the resulting added R signal is positioned at the centroid A. B pixels 214, 216, 218, 254, 256, 258, 294, 296, and 298 are summed by shifting horizontally three pixels from R pixels 211, 213, 215, 251, 253, 255, 291, 293, and 295 , and then localize the resulting summed B signal at the centroid B. R pixels 217, 219, 21b, 257, 259, 25b, 297, 299, and 29b are shifted horizontally by a further three pixels from B pixels 214, 216, 218, 254, 256, 258, 294, 296, and 298 , then localize the resulting summed R signal at the centroid C.

G pixels 242 , 244 , 246 , 282 , 284 , 286 , 2c2 , 2c4 , and 2c6 located in even-numbered rows are summed and the resulting added G signal is positioned at centroid D by shifting three pixels by tilt. Adding G pixels 245, 247, 249, 285, 287, 289, 2c5, 2c7, and 2c9 by shifting horizontally three pixels from G pixels 242, 244, 246, 282, 284, 286, 2c2, 2c4, and 2c6, The resulting summed G signal is then localized at the centroid E.

In this way, by adding color pixels as described above over the entire pixel area, pixels of the same color can be added while maintaining the original color pattern without changing the color space repeating pattern or vertical, horizontal, and oblique The pitch ratio in the direction. As a result, by spatially equal sampling, high-quality addition signals can be obtained. Meanwhile, the added signal can be output in the same color pattern as that obtained by the progressive reading method for independently reading pixel signals, thereby facilitating signal processing in a subsequent stage.

third embodiment

FIG. 13 illustrates the idea of adding pixels by a driving method according to a third embodiment of the present invention. In the oblique pixel pattern without a color filter, from a 3x3 pixel area, adjacent pixels in two columns and two rows are extracted and added, while shifting the 3x3 pixel area by three pixels. The process of such pixel addition will be discussed in detail below.

In FIG. 13 , pixels 311 , 312 , 331 , and 332 located in odd-numbered rows are added, and then, the resulting added signal is positioned at the centroid A. Pixels 314 , 315 , 334 , and 335 are added by shifting three pixels horizontally from pixels 311 , 312 , 331 , and 332 , and then, the resulting added signal is positioned at centroid B. Pixels 317 , 318 , 337 , and 338 are added by shifting a further three pixels from 314 , 315 , 334 , and 335 , and the resulting added signal is then positioned at centroid C.

Then, the pixels 342 , 343 , 362 , and 363 located in even-numbered rows are added by shifting three pixels in a tilt, and then, the resulting added signal is positioned at the centroid D. Pixels 345 , 346 , 365 , and 366 are added by shifting three pixels horizontally from pixels 342 , 343 , 362 , and 363 , and the resulting added signal is positioned at centroid E.

In this way, by adding pixels as described above over the entire pixel area, the pixels can be added while maintaining the original pattern without changing the pitch ratio in the vertical, horizontal, and oblique directions. As a result, by spatially equal sampling, high-quality addition signals can be obtained. Meanwhile, the addition signal can be output in the same pattern as that obtained by the progressive reading method for independently reading pixel signals, thereby facilitating signal processing in a subsequent stage.

Fourth embodiment

FIG. 14 illustrates the idea of adding pixels by a driving method according to a fourth embodiment of the present invention. In the oblique pixel pattern without a color filter, from a 3×3 pixel area, adjacent pixels in three columns and three rows are extracted and added while shifting the 3×3 pixel area by three pixels. The process of such pixel addition will be discussed in detail below.

In FIG. 14 , pixels 411 , 412 , 413 , 431 , 432 , 433 , 451 , 452 , and 453 located in odd-numbered rows are added, and then, the resulting added signal is positioned at the centroid A. Add pixels 414, 415, 416, 434, 435, 436, 454, 455, and 456 by shifting horizontally three pixels from pixels 411, 412, 413, 431, 432, 433, 451, 452, and 453, and then , positioning the resulting summed signal at the centroid B. Add pixels 417, 418, 419, 437, 438, 439, 457, 458, and 459 by moving a further three pixels from pixels 414, 415, 416, 434, 435, 436, 454, 455, and 456, and The resulting summed signal is positioned at the centroid C.

Then, pixels 442 , 443 , 444 , 462 , 463 , 464 , 482 , 483 , and 484 located in even-numbered rows are added by shifting three pixels by tilting, and the resulting added signal is positioned at the centroid D. Add pixels 445, 446, 447, 465, 466, 467, 485, 486, and 487 by shifting horizontally three pixels from pixels 442, 443, 444, 462, 463, 464, 482, 483, and 484, and add The resulting summed signal is localized at the centroid E.

In this way, by adding pixels as described above over the entire pixel area, the pixels can be added while maintaining the original pattern without changing the pitch ratio in the vertical, horizontal, and oblique directions. As a result, by spatially equal sampling, high-quality addition signals can be obtained. Meanwhile, the added signal can be output in the same pattern as that obtained by the progressive reading method for independently reading pixel signals, thereby facilitating signal processing in a subsequent stage.

fifth embodiment

FIG. 15 illustrates the idea of adding pixels by a driving method according to a fifth embodiment of the present invention. In dedicated color coding as an example of dedicated color coding, pixels of the same color in two columns and two rows are extracted and added from a 5×5 pixel region while shifting the 5×5 pixel region by three pixels. The process of such pixel addition will be discussed in detail below.

In FIG. 15 , R pixels 511 , 513 , 551 , and 553 located in odd-numbered rows are added, and then, the resulting added R signal is positioned at the centroid A. Similarly, G pixels 514 , 516 , 554 , and 556 are summed by shifting three pixels horizontally from R pixels 511 , 513 , 551 , and 553 , and the resulting added G signal is positioned at centroid B. The R image signals 517 , 519 , 557 , and 559 are added by shifting further three pixels horizontally from the G pixels 514 , 516 , 554 , and 556 , and the resulting added R signal is positioned at the centroid C.

The G pixels 542 , 544 , 582 , and 584 located in the even-numbered rows are summed and the resulting added G signal is positioned at the centroid D by shifting three pixels by tilt. G pixels 545 , 547 , 585 , and 587 are summed by shifting three pixels horizontally from G pixels 542 , 544 , 582 , and 584 , and the resulting added G signal is positioned at centroid E. FIG.

In this way, by adding color pixels as described above over the entire pixel area, pixels of the same color can be added while maintaining the original color pattern without changing the color space repeating pattern or the vertical, horizontal, and oblique directions pitch ratio. As a result, by spatially equal sampling, high-quality addition signals can be obtained. At the same time, the addition signal can be output in the same color pattern as that obtained by the progressive reading method for independently reading pixel signals, thereby facilitating signal processing in a subsequent stage.

Basic form

Through the first to fifth embodiments, the driving method for adding pixels by using various patterns has been discussed. The general form of adding pixels is described below.

In a solid-state imaging device including a pixel region (pixel array) having a slanted pixel pattern, in an n ² region of adjacent n×n pixels (n is an odd number of three or more), extracting and adding the values in the x columns and y (n≥x≥y) rows of pixels of the same color while shifting the n×n area by m pixels (m is an odd number of three or more) in the vertical or horizontal direction. In this case, pixels are added for odd-numbered and even-numbered rows, respectively. In this case, the spatial positional relationship between the adjacent pixels in the adjacent n×n pixels of the odd-numbered row and the adjacent pixels in the adjacent n×n pixels of the even-numbered row is such that they are in are shifted by m pixels from each other in the oblique direction.

According to the driving method for performing the above-described pixel addition, pixels can be added while maintaining the original color pattern without changing the color space repetition pattern or the pitch ratio in the vertical, horizontal, and oblique directions.

As a result, by spatially equal sampling, high-quality addition signals can be obtained. At the same time, the added signal can be output in the same pattern as that obtained by the progressive reading method for independently reading pixel signals, thereby facilitating signal processing in a subsequent stage. In addition, since n≥x≥y, as described below, it is not necessary to increase sampling and holding capacitors for adding pixels.

solid state imaging device

A description will now be given of an example of the configuration of a solid-state imaging device implementing the driving method for adding pixels according to the first to fifth embodiments.

FIG. 16 schematically illustrates the configuration of an X-Y addressing solid-state imaging device (for example, a CMOS image sensor) according to an embodiment of the present invention.

The CMOS image sensor shown in Figure 16 includes: pixel 10, pixel array 11, color filter 12, odd-numbered row drive line group 13, even-numbered row drive line group 14, odd-numbered column signal line (vertical signal line) group 15 , even numbered column signal line (vertical signal line) group 16, odd numbered row vertical selection circuit 17, even numbered row vertical selection circuit 18, odd numbered row and column processing circuit 19, even numbered row and column processing circuit 20, odd numbered row horizontal switch 21 , odd-numbered horizontal signal lines 22, odd-numbered horizontal selection circuits 23, output amplifiers 24 and 28, even-numbered horizontal switches 25, even-numbered horizontal signal lines 26, even-numbered horizontal selection circuits 27, and timing generating circuits 29.

In FIG. 16 , pixels 10 having photoelectric converters arranged in a matrix form a pixel array (pixel region) 11 . In the pixel array 11, the even-numbered column pixels 10 are shifted from the odd-numbered column pixels 10 in the column direction (vertical direction in FIG. In the direction (horizontal direction in FIG. 16 ), the pixels 10 of odd-numbered rows are shifted by about 1/2 of the pixel pitch. That is, the pixel array 11 forms a slanted pixel pattern.

In the pixel array 11 having an oblique pixel pattern, color filters 12 having color codes shifted by 45° from the Bayer pattern (see FIG. 8 ) are arranged on the pixels 10 . The odd-numbered row driving lines of the odd-numbered row driving line group 13 are provided for each odd-numbered coded row of the pixel 10 , and the even-numbered row driving lines of the even-numbered row driving line group 14 are provided for each even-numbered coded row of the pixel 10 . The odd-numbered column signal lines of the odd-numbered column signal line group 15 are connected to each odd-numbered column pixel 10 , and the even-numbered column signal lines of the even-numbered column signal line group 16 are connected to each even-numbered column pixel 10 .

One end of each odd-numbered row driving line of the odd-numbered row driving line group 13 is connected to the corresponding output terminal of the odd-numbered row vertical selection circuit 17 . One end of each even-numbered row driving line of the even-numbered row driving line group 14 is connected to the corresponding output terminal of the even-numbered row vertical selection circuit 18 . The odd-numbered row vertical selection circuit 17 and the even-numbered row vertical selection circuit 18 form a circuit for selecting odd-numbered rows that are not adjacent to each other in the pixel array 11 via the odd-numbered row drive line group 13 and the even-numbered row drive line group 14, respectively. Row selectors for pixels 10 of numbered and even numbered rows.

One end of each odd-numbered column signal line of the odd-numbered column signal line group 15 is connected to the corresponding odd-numbered row-column processing circuit arranged on one side of the pixel array 11 (in this embodiment, the lower side in FIG. 16 ). 19 inputs. The odd-numbered row-column processing circuit 19 stores pixel signals in odd-numbered rows, and adds pixel signals in every other column.

Connect one end of each even-numbered column signal line of the even-numbered column signal line group 16 to the corresponding even-numbered row-column process arranged on the other side of the pixel array 11 (in this embodiment, the upper side in FIG. 16 ). input to circuit 20. The even-numbered row-column processing circuit 20 stores pixel signals in even-numbered rows, and adds pixel signals in all other columns.

Specific circuit configurations of the odd-numbered row-column processing circuit 19 and the even-numbered row-column processing circuit 20 are described below. The symbols A, B, and C attached to the odd-numbered row-column processing circuit 19 and the even-numbered row-column processing circuit 20 in FIG. 16 are for distinguishing the switches shown in FIG. 17 .

The output ends of the odd-numbered row and column processing circuits 19 are connected to odd-numbered horizontal signal lines 22 via corresponding odd-numbered row horizontal switches 21 . The odd-numbered row horizontal switches 21 are sequentially selected by the odd-numbered row horizontal selection circuit 23 to read out the signal added in the odd-numbered row-column processing circuit 19 to the odd-numbered row horizontal signal line 22 . The signal read out to the odd-numbered row horizontal signal line 22 is amplified in the output amplifier 24 and then output.

The output terminals of the even-numbered row-column processing circuit 20 are connected to the even-numbered row-horizontal signal lines 26 via corresponding even-numbered row-level switches 25 . The even-numbered row horizontal switches 25 are sequentially selected by the even-numbered row horizontal selection circuit 27 to read out the signal added in the even-numbered row-column processing circuit 20 to the even-numbered row horizontal signal line 26 . The signal read out to the even-numbered row horizontal signal line 26 is amplified in the output amplifier 28 and then output.

The odd-numbered row horizontal selection circuit 23 and the even-numbered row horizontal selection circuit 27 form a column selector for selecting columns such that column numbers in odd-numbered rows do not coincide with column numbers in even-numbered rows. The odd-numbered row vertical selection circuit 17, the even-numbered row vertical selection circuit 18, the odd-numbered row-column processing circuit 19, the even-numbered row-column processing circuit 20, and the odd-numbered row horizontal selection circuit are executed based on various timing signals output from the timing generation circuit 29. 23, and the drive control of the horizontal selection circuit 27 for even-numbered rows.

17 is a circuit diagram illustrating an example of the configuration of the odd-numbered row-column processing circuit 19 assuming that pixels of the same color in two columns and two rows are added in color coding shifted by 45° from the Bayer pattern (see FIG. 8 ). The even-numbered row-column processing circuits 20 basically have the same configuration.

In Fig. 17, clamp (clamp) pulse line 31, clamp voltage line 32, recording control line 33, addition control line 34, A switch line 35, B switch line 36, C switch line 37, and clamp The voltage line 38 is connected to each of the odd-numbered column processing circuits 19A, 19B, and 19C (corresponding to the column processing circuits A, B, and C shown in FIG. 16 ) forming one unit.

The column processing circuits 19A, 19B, 19C basically have the same circuit configuration. The column processing circuit 19A includes: a clamp capacitor 41A; first, second, and third switches 42A, 43A, and 44A; and a sampling and holding capacitor 45A. The column processing circuit 19B includes: a clamp capacitor 41B; first, second, and third switches 42B, 43B, and 44B; and a sample-and-hold capacitor 45B. The column processing circuit 19C includes: a clamp capacitor 41C; first, second, and third switches 42C, 43C, and 44C; and a sample-and-hold capacitor 45C. N-channel MOS transistors may be used as the first, second, and third switches.

A more specific configuration of the column processing circuit 19 is discussed below by taking the column processing circuit 19A as an example. One end of the clamp capacitor 41A is connected to one end of the odd-numbered column signal line (vertical signal line) corresponding to the odd-numbered column signal line group 15 . The first switch 42A is connected between the other end of the clamp capacitor 41A and the clamp voltage line 32 , and the gate of the first switch 42A is connected to the clamp pulse line 31 . One main electrode of the second switch 43A is connected to the other end of the clamp capacitor 41A, and its gate electrode is connected to the recording control line 33 .

One main electrode of the third switch 44A is connected to the other main electrode of the second switch 43A, and its gate electrode is connected to the A switch line 35 . In the column processing circuit 19B, the gate electrode of the third switch 44B is connected to the B switch line 36 . In the column processing circuit 19C, the gate electrode of the third switch 44C is connected to the C switch line 37 . One end of the sample-and-hold capacitor 45A is connected to the other main electrode of the third switch 44A, and the other end thereof is connected to the clamp voltage line 38 .

In the column processing circuits 19A, 19B, and 19C configured above, in this embodiment, the pixel signals in the column processing circuits 19A and 19C are added. Furthermore, between the column processing circuits 19A and 19C, more specifically, between the main electrodes of the third switches 44A and 44C in the column processing circuits 19A and 19C, an addition switch 46 is connected. As the adding switch 46, an N-channel MOS transistor can be used. The gate of the addition switch 46 is connected to the addition control line 34 .

The operation of the CMOS image sensor with the oblique pixel pattern including the above-configured column processing circuits 19A, 19B, and 19C according to this embodiment will be described below with reference to the timing chart of FIG. 18 .

18 shows the timing relationship between the following pulses: a reset pulse for resetting the potential in the floating diffusion region in the pixel 10; a charge transfer pulse for transferring the signal charge photoelectrically converted in the photoelectric transducer to the floating diffusion region; Diffusion area; clamping pulse, is supplied to clamping pulse line 31; recording control pulse, is supplied to recording control line 33; addition control pulse, is supplied to addition control line 34; A switch pulse, is supplied to A The switch line 35 ; the B switch pulse is supplied to the B switch line 36 ; and the C switch pulse is supplied to the C switch line 37 .

When operating the CMOS image sensor having the inclined pixel pattern of this embodiment in the addition read mode, the addition control pulse is changed to H level to put the addition switch 46 in the ON state. If the addition operation is not being performed, the addition control pulse is changed to L level so that the addition switch 46 is in an "OFF" state. In this case, dummy switches are preferably arranged in the column processing circuits 19 and 20 so as to be connected to the vertical signal lines in the odd-numbered row signal line group 15 and the vertical signal lines in the even-numbered row signal line group 16. The change in the load capacitance of the adding switch 46 of the line becomes insignificant.

In FIG. 16, the first row is selected by vertical scanning by the odd-numbered row vertical selection circuit 17, and the fourth row is selected by the even-numbered row vertical selection circuit 18 by vertical scanning. By selecting rows in this way, odd-numbered rows and even-numbered rows that are not adjacent to each other are simultaneously read, which is a feature of the embodiment of the present invention. For ease of understanding, this is described below together with the first embodiment (FIG. 11).

Reset pulses are first supplied to pixels in selected two rows (first and fourth rows), and then reset voltages reflecting the reset levels of the pixels appear in corresponding signal lines of the vertical signal line groups 15 and 16 . In this case, the clamping pulse is changed to H level so that the first switches 42A, 42B and 42C in FIG. Stored in the clamp capacitors 41A, 41B, and 41C (first reset voltage reading period). Then, the clamp pulse is changed to L level to turn "OFF" the first switches 42A, 42B, and 42C.

Subsequently, charge transfer pulses are supplied to selected two rows (first and fourth rows), and then signal voltages reflecting optical signal levels of pixels 10 appear in corresponding signal lines of vertical signal line groups 15 and 16 . In this case, since a difference between the signal voltage and the reset voltage stored in the clamp capacitors 41A, 41B, and 41C is generated, fixed pattern noise is eliminated from the pixel 10 (first signal voltage read period).

Then, the recording control pulse is changed to H level, and the A switch pulse is changed to H level, so that the third switch 44A (hereinafter, simply referred to as "A switch 44A") is connected to the A switch line. The signals in the column of 35 are added to the signals in the column in which the third switch 44C (hereinafter, simply referred to as “C switch 44C”) is connected to the C switch line 37 .

As a result, the signal in the column connected to the A switch 44A and the signal in the column connected to the C switch 44C are added via the adding switch 46, and the added signal is stored in the sampling and holding capacitor 45A of the column processing circuit 19A. middle. More specifically, in FIG. 11 , in odd-numbered rows, the signal of the R pixel 111 and the signal of the R pixel 113 are added, the signal of the G pixel 114 and the signal of the G pixel 116 are added, and so on. In even-numbered rows, the signal of the G pixel 142 and the signal of the G signal 144 are added, the signal of the B signal 145 and the signal of the B signal 147 are added, and so on.

Then, both the recording control pulse and the A switching pulse are changed to L level. Thereafter, the floating diffusions (FDs) of the selected two rows are reset. The period of these operations so far is the first read period including the first reset voltage read period and the first signal voltage read period.

Subsequently, in FIG. 16 , the fifth row is selected by vertical scanning by the odd-numbered row vertical selection circuit 17 , and the eighth row is selected by the even-numbered row vertical selection circuit 18 by vertical scanning. Odd-numbered rows and even-numbered rows that are not adjacent to each other are simultaneously read out.

Reset pulses are first supplied to pixels in selected two rows (fifth and eighth rows), and then reset voltages reflecting the reset levels of the pixels appear in corresponding signal lines of vertical signal line groups 15 and 16 . Then, the clamp pulse is changed to the H level to turn ON the first switches 42A, 42B, and 42C, and the reset voltage of the signal lines in the vertical signal line groups 15 and 16 is stored in the clamp capacitors 41A, 41B. , and 41C (the second reset voltage reading period). Then, the clamp pulse is changed to L level to turn "OFF" the first switches 42A, 42B, and 42C.

Subsequently, charge transfer pulses are supplied to the selected two rows (fifth and eighth rows), and signal voltages reflecting the optical signal levels of the pixels appear in the corresponding signal lines of the vertical signal line groups 15 and 16 . Since the difference between the signal voltage and the reset voltage stored in the clamp capacitors 41A, 41B, and 41C is generated, fixed pattern noise can be eliminated from the pixel 10 (second signal voltage read period).

Then, the recording control pulse is changed to H level, and the C switch pulse is changed to H level, so that the signal in the column connected to the A switch 44A is added to the signal in the column connected to the C switch 44C, and The added signal is stored in the sample-hold capacitor 45C of the column processing circuit 19C.

As a result, the signal in the column connected to the A switch 44A and the signal in the column connected to the C switch 44C are added via the adding switch 46, and the added signal is stored in the sample-and-hold circuit 45C of the column processing circuit 19C middle. More specifically, in FIG. 11 , in odd-numbered rows, the signal of the R pixel 151 and the signal of the R pixel 153 are added, the signal of the B pixel 154 and the signal of the B pixel 156 are added, and so on. In even-numbered rows, the signal of the G pixel 182 and the signal of the G pixel 184 are added, the signal of the G pixel 185 and the signal of the G pixel 187 are added, and so on.

Subsequently, the recording control pulse and the C switching pulse are changed to L level, and the floating diffusion regions of the pixels of the selected two rows are reset. Then, the A switch pulse and the C switch pulse are changed to H level to turn on the A switch 44A and the C switch 44C. As a result, the horizontal addition signals stored in the sampling and holding capacitors 45A and 45C of the column processing circuits 19A and 19C are vertically added (vertical addition period).

More specifically, in odd-numbered rows, the horizontal addition signals (the signals of the R pixel 111 and the R pixel 113, the signals of the B pixel 114 and the B pixel 116, etc.) stored in the sampling and holding capacitor 45A and the horizontal Addition signals (signals of the R pixel 151 and R pixel 153, signals of the B pixel 154 and B pixel 156, etc.) are vertically added.

In even-numbered rows, the horizontal addition signal (the signal of the G pixel 142 and the G pixel 144, the signal of the G pixel 145 and the G pixel 147, etc.) and the horizontal addition signal ( The signals of the G pixel 182 and the G pixel 183, the signals of the G pixel 185 and the G pixel 187, etc.) are added vertically.

According to the addition operation described above, in the color coding of the first embodiment shifted by 45° from the Bayer pattern, from the 3×3 pixel area, pixels of the same color in two columns and two rows are extracted and added. A signal obtained by adding these pixels can be read out from the sampling and holding capacitor 45A or 45C. The period of the operation so far is the second read period including the second reset voltage read period, the second signal voltage read period, and the vertical addition period.

In the aforementioned reading method, the added signals of horizontally x pixels (in this embodiment, x=2) are stored in a single sampling and holding capacitor 45 (45A, 45B or 45C), and, This operation is repeated for the same number of vertical y columns (in this embodiment, y=2). In this case, if x≧y, the sampling and holding capacitors 45 provided for the same number of columns of the original pixel pattern are sufficient. In this way, the addition operation does not require an additional sample-and-hold capacitor

The first read period and the second read period form a horizontal blanking period. In one horizontal blanking period, odd-numbered rows and even-numbered rows that are not adjacent to each other are read out multiple times (in this embodiment, since two pixels (two rows) vertically are added, two times), which is a feature of this embodiment of the invention. After the horizontal blanking period, the horizontal read period begins.

If the added signal is read out from the sample-and-hold capacitor 45A of the column processing circuit 19A in the vertical addition period, the first column, the first column, the first column, and the second column can be selected in the odd-numbered row horizontal selection circuit 23 according to the configuration shown in FIG. Fourth column, seventh column, and so on. Then, the added signals in the odd-numbered rows can be extracted. Similarly, the second column, the fifth column, the eighth column, etc. can be selected by the even-numbered row horizontal selection circuit 27, and then, the added signal in the even-numbered row can be extracted.

In this case, columns corresponding to odd-numbered rows and columns corresponding to even-numbered rows are selected so that they are shifted from each other. That is, the column number corresponding to the odd-numbered row selected by the odd-numbered row horizontal selection circuit 23 does not match the column number corresponding to the even-numbered row selected by the even-numbered row horizontal selection circuit 27 . The pattern of the added signal needs to match the original pattern of the pixel signal.

As described above, the solid-state imaging device having an oblique pixel pattern includes: vertical selection circuits 17 and 18 for simultaneously selecting odd-numbered rows and even-numbered rows that are not adjacent to each other a plurality of times in one horizontal blanking period; a column processing circuit 19 (19A, 19B, and 19C) and 20 (20A, 20B, and 20C), respectively having odd-numbered row horizontal switches 21 and even-numbered row horizontal switches 25; and horizontal selection circuits 23 and 27 for selecting the switches 21 and 25, respectively . With this configuration, the column corresponding to the selected odd-numbered row and the column corresponding to the even-numbered row are shifted from each other. Therefore, after the pixels are added, the original color encoding can be maintained without changing the color space repeat pattern or the pixel pitch ratio in the vertical, horizontal, and oblique directions. As a result, by spatially equal sampling, high-quality addition signals can be obtained. At the same time, the addition signal can be output in the same color pattern as that obtained by the progressive reading method for independently reading pixel signals, thereby facilitating signal processing in a subsequent stage.

Application example

The above-described X-Y addressing solid-state imaging device (a typical example of which is a CMOS image sensor) having an oblique pixel pattern is suitably used as an imaging device for an imaging device (camera module) such as a digital still camera or a video camera.

Fig. 19 is a block diagram illustrating an example of the configuration of an imaging device according to an embodiment of the present invention. The imaging apparatus shown in FIG. 19 includes a lens 51 , an imaging device 52 , a signal processing circuit 52 , a mode setting unit 54 , and a device driving circuit 55 .

The lens 51 forms an image corresponding to light reflected by a subject on an imaging plane of an imaging device 52 . Under the control of the device driving circuit 55 , the imaging device 52 converts the optical image formed on the imaging surface through the lens 51 into an electrical signal in units of pixels, and outputs the resulting image signal. An X-Y addressing solid-state imaging device (a typical example of which is a CMOS image sensor) including an oblique pixel pattern according to the above-described embodiments is used as the imaging device 52 . The device driving circuit 55 may be mounted on the imaging device 52 by an in-chip technique.

The signal processing circuit 53 performs various signal processing operations on the image signal output from the imaging device 52 . The mode setting unit 54 selectively sets a progressive reading mode for reading all pixel signals or an addition reading mode for performing pixel addition as the operation mode of the imaging device 52 in response to a user's instruction.

The device driving circuit 55 is formed of, for example, the timing generating circuit 29 shown in FIG. 16 , and controls driving of the timing device 52 in response to a mode signal supplied from the mode setting unit 54 .

More specifically, when the progressive reading mode is specified in the imaging device (CMOS image sensor) shown in FIG. , and the signals read from the pixels of the selected row are sequentially selected by the horizontal selection circuits 23 and 27, respectively.

When the addition read mode is specified, the device driving circuit 55 performs control so that odd-numbered rows and even-numbered rows that are not adjacent to each other are simultaneously selected multiple times by the vertical selection circuits 17 and 18 in one horizontal blanking period, respectively. , and the signals of the pixels of the selected row are horizontally added by column processing circuits 19 (19A, 19B, and 19C) and 20 (20A, 20B, and 20C) having horizontal switches 21 and 25, respectively, and The horizontal selection circuits 23 and 27 sequentially read out the added signals.

As discussed above, the X-Y addressable solid-state imaging device (a typical example of which is a CMOS image sensor) according to one of the above-described embodiments is loaded in an imaging device such as a digital still camera or a video camera. Therefore, the imaging device can cope with both the progressive reading mode and the additive reading mode. In addition, the imaging device may output an addition signal having the same color pattern as a signal obtained by the progressive reading mode. As a result, by spatially equal sampling, a high-quality added signal can be obtained, and signal processing in the signal processing circuit 53 can be facilitated.

Sixth embodiment

FIG. 21 illustrates color patterns of a color filter according to a sixth embodiment of the present invention. In the solid-state imaging device including the color pattern according to the sixth embodiment, as shown in FIG. The equal intervals (pixel pitch) d are arranged in a square grid.

In this square grid sub-pixel pattern, in the first row, RGRB is repeatedly arranged in the horizontal direction in units of four pixels, in the second row, only G pixels are arranged, and in the third row, BGRG is arranged in units of Four pixels are repeatedly arranged in the horizontal direction in units, and in the fourth row, only G pixels are arranged. Thereafter, these four rows are repeatedly arranged.

In the color pattern of this embodiment, as seen from FIG. 21, a color component (G component in this embodiment) and other components (G component in this R and B components in ) are arranged such that the G component surrounds the R and B components. Also, in this color pattern, R and B pixels are arranged at regular intervals 4d in the vertical and horizontal directions.

而对R和B像素的抽样率是

In the color pattern configured above, if the pixel sampling rate in the vertical and horizontal directions is set to the pixel pitch d, the sampling rate for G pixels is d, and the sampling rate for R and B pixels is 2d. That is, R pixels or B pixels are arranged every other column (in this embodiment, odd-numbered rows) and every other row (in this embodiment, odd-numbered columns) such that the R and B pixels are vertically and horizontally The sampling rate becomes 1/2 of the sampling rate for G pixels. Therefore, the resolution of R and B pixels is 1/2 of the resolution of G pixels. In the 45° oblique direction, the sampling rate for G pixels is

while the sampling rate for R and B pixels is

如图22B所示，所以根据抽样定理可以收集频率最高

的信号分量。22A and 22B show G patterns in which only G pixels are extracted from the color patterns shown in FIG. 21 . The spatial frequency characteristic of the G pixel is considered in conjunction with FIGS. 22A and 22B. In the vertical and horizontal directions, since the sampling rate of G pixels is d, as shown in FIG. 22A, the signal component with the highest frequency (1/2) fs (fs: sampling frequency) can be collected according to the sampling theorem. In the 45° oblique direction, since the sampling rate of G pixels is

As shown in Figure 22B, so according to the sampling theorem, the highest frequency can be collected

signal components.

Similarly, the spatial frequency characteristics of R and B pixels are considered below. In this case, since the pixel pitches of the R and B pixels are the same, only the spatial frequency characteristic of the R pixel will be described below.

22C and 22D show R patterns in which only R pixels are extracted from the color patterns shown in FIG. 21 . In FIGS. 22A to 22D , threshold frequency components that can be collected in the vertical, horizontal, and oblique directions are indicated by empty columns and halftone columns.

如图22D所示，所以根据抽样定理可以收集频率最高

的信号分量。In the vertical and horizontal directions, since the sampling rate of R pixels is 2d, as shown in FIG. 22C , the signal components with the highest frequency of 1/4fs can be collected according to the sampling theorem. In the 45° inclined direction, since the sampling rate for R pixels is

As shown in Figure 22D, so according to the sampling theorem, the highest frequency can be collected

signal components.

的信号分量。From the above points, FIG. 23 indicates the spatial frequency characteristics of the color pattern of the sixth embodiment. Regarding the spatial frequency characteristics of G pixels, in the vertical and horizontal directions, the signal components with the highest frequency (1/2) fs can be collected, while in the oblique direction, the highest frequency can be collected signal components. Regarding the spatial frequency characteristics of R and B pixels, in the vertical and horizontal directions, the signal components with the highest frequency (1/4) fs can be collected, while in the oblique direction, the highest frequency can be collected

signal components.

Seventh embodiment

FIG. 24 illustrates color patterns of a color filter according to a seventh embodiment of the present invention. Like the solid-state imaging device using the color pattern of the sixth embodiment shown in FIG. 21 , the solid-state imaging device using the color filter of this embodiment has a checkered pixel pattern.

In the color pattern of this embodiment, in the first row, RGGG is repeatedly arranged in units of four pixels in the horizontal direction, in the second row, only G pixels are arranged, and in the third row, GGBG The arrangement is repeated in the horizontal direction in units of four pixels, and in the fourth row, only G pixels are arranged. Thereafter, these four rows are repeatedly arranged.

In the color pattern of this embodiment, as seen from FIG. 24, a color component (G component in this embodiment) and other components (G component in this R and B components in the example) are arranged such that the G component surrounds the R and B components. Also, in this color pattern, like the color pattern in the sixth embodiment, R and B pixels are arranged at regular intervals 4d in the vertical and horizontal directions.

In the color pattern configured above, if the pixel sampling rate in the vertical and horizontal directions is set to the pixel pitch d, the sampling rate for G pixels is d, and the sampling rate for R and B pixels is 4d. That is, R pixels or B pixels are arranged every other column (in this embodiment, odd-numbered rows) and every other row (in this embodiment, odd-numbered columns) so that the sampling of R and B pixels in the vertical and horizontal directions The rate becomes 1/4 of the sampling rate for G pixels. Therefore, the resolution of R and B pixels is 1/4 of that of G pixels. In the 45° inclined direction, the sampling rate for G pixels is The sampling rate for R and B pixels is

如图25B所示，所以根据抽样定理可以收集频率最高的信号分量。25A and 25B show G patterns in which only G pixels are extracted from the color patterns shown in FIG. 24 . The spatial frequency characteristic of the G pixel is considered in conjunction with FIGS. 25A and 25B. In the vertical and horizontal directions, since the sampling rate of the G pixel is d, as shown in FIG. 25A, the signal component with the highest frequency (1/2) fs (fs: sampling frequency) can be collected according to the sampling theorem. In the direction inclined at 45°, since the sampling rate of G pixels is

As shown in Figure 25B, so according to the sampling theorem, the highest frequency can be collected signal components.

Similarly, the spatial frequency characteristics of R and B pixels are considered below. In this case, since the pixel pitches of the R and B pixels are the same, only the spatial frequency characteristic of the R pixel will be described below.

25C and 25D show R patterns in which only R pixels are extracted from the pixel patterns shown in FIG. 24 . In FIGS. 25A to 25D , threshold frequency components that can be collected in the vertical, horizontal, and oblique directions are indicated by empty columns and halftone columns.

如图25D所示，所以根据抽样定理可以收集频率最高

的信号分量。In the vertical and horizontal directions, since the sampling rate for R pixels is 4d, as shown in Fig. 25C, the signal components with the highest frequency of 1/8fs can be collected according to the sampling theorem. In the 45° inclined direction, since the sampling rate for R pixels is

As shown in Figure 25D, so according to the sampling theorem, the highest frequency can be collected

signal components.

的信号分量。关于R和B像素的空间频率特性，在垂直和水平方向上，可以收集频率最高(1/8)fs的信号分量，而在倾斜方向上，可以收集频率最高

的信号分量。From the above points, FIG. 26 indicates the spatial frequency characteristics of the color pattern of the seventh embodiment. Regarding the spatial frequency characteristics of G pixels, in the vertical and horizontal directions, the signal components with the highest frequency (1/2) fs can be collected, while in the oblique direction, the highest frequency can be collected

signal components. Regarding the spatial frequency characteristics of R and B pixels, in the vertical and horizontal directions, the signal components with the highest frequency (1/8) fs can be collected, while in the oblique direction, the highest frequency can be collected

signal components.

Eighth embodiment

并且，将偶数编号列像素从奇数编号列像素在行和列方向上移位大约像素节距的1/2。就是说，在每一行，在水平方向以

的节距布置像素，而在每一列，在垂直方向以

的节距布置像素。FIG. 27 illustrates color patterns of a color filter according to an eighth embodiment of the present invention. As shown in FIG. 27, a solid-state imaging device using a color pattern according to this embodiment has an oblique pixel pattern in which the pixel pitches in the vertical and horizontal directions are set to be the same as those of the checkered sub-pixel pattern shown in FIG. different

And, the even-numbered column pixels are shifted from the odd-numbered column pixels in the row and column directions by approximately a pixel pitch 1/2 of. That is, in each row, in the horizontal direction with

Pixels are arranged at a pitch, and in each column, in the vertical direction with

Pixels are arranged at a pitch of .

In the color pattern with this oblique pixel pattern, in the first row, G and R pixels are alternately arranged, in the second row, only G pixels are arranged, in the third row, B and G pixels are alternately arranged, and in the In the fourth row, only G pixels are arranged. Thereafter, these four rows are repeatedly arranged.

In the color pattern of this embodiment, as seen from FIG. 27, a color component (G component in this embodiment) and other components (G component in this R and B components in the example) are arranged such that the G component surrounds the R and B components. In this color pattern, at regular intervals vertically and horizontally R and B pixels are laid out. This color pattern is equal to a pattern shifted by 45° from the color pattern in the checkered sub-pixel pattern of the sixth embodiment shown in FIG. 21 .

而对于R和B像素的抽样率是

就是说，隔列(在这个实施例中，奇数编号的行)和隔行(在这个实施例中，奇数编号的列)布置R像素或B像素，使得在垂直和水平方向上对R和B像素的抽样率变成对G像素的抽样率的1/4。因此，R和B像素的分辨率是G像素的分辨率的1/4。在45°倾斜方向上，对G像素的抽样率是d，而对R和B像素的抽样率是2d。In the color pattern configured above, in the vertical and horizontal directions, the sampling rate for G pixels is

while the sampling rate for R and B pixels is

That is, R pixels or B pixels are arranged every other column (in this embodiment, odd-numbered rows) and every other row (in this embodiment, odd-numbered columns) such that the R and B pixels are vertically and horizontally The sampling rate becomes 1/4 of the sampling rate for G pixels. Therefore, the resolution of R and B pixels is 1/4 of that of G pixels. In the 45° oblique direction, the sampling rate for G pixels is d, and the sampling rate for R and B pixels is 2d.

28A and 28B show G patterns in which only G pixels are extracted from the color patterns shown in FIG. 27 . The spatial frequency characteristics of G pixels are considered in conjunction with FIGS. 28A and 28B. In the vertical and horizontal directions, since the sampling rate of G pixels is As shown in Figure 28A, so according to the sampling theorem, the highest frequency can be collected (fs: sampling frequency) signal component. In the oblique direction of 45°, since the sampling rate for G pixels is d, as shown in FIG. 28B, the signal component with the highest frequency (1/4) fs can be collected according to the sampling theorem.

Similarly, the spatial frequency characteristics of R and B pixels are considered below. In this case, since the pixel pitches of the R and B pixels are the same, only the spatial frequency characteristic of the R pixel will be described below.

28C and 28D show R patterns in which only R pixels are extracted from the pixel patterns shown in FIG. 27 . In FIGS. 28A to 28D , threshold frequency components that can be collected in vertical, horizontal, and oblique directions are indicated by empty columns and halftone columns.

的信号分量。在45°倾斜方向上，由于R像素的抽样率是2d，如图28D所示，所以根据抽样定理可以收集频率最高(1/2)fs的信号分量。In the vertical and horizontal directions, since the sampling rate for R pixels is As shown in Figure 28C, so according to the sampling theorem, the highest frequency can be collected

signal components. In the 45° oblique direction, since the sampling rate of the R pixel is 2d, as shown in FIG. 28D , the signal component with the highest frequency (1/2)fs can be collected according to the sampling theorem.

的信号分量，而在倾斜方向上，可以收集频率最高(1/4)fs的信号分量。关于R和B像素的空间频率特性，在垂直和水平方向上，可以收集频率最高

的信号分量，而在倾斜方向上，可以收集频率最高(1/2)fs的信号分量。From the above points, FIG. 29 indicates the spatial frequency characteristics of the color pattern of the eighth embodiment. Regarding the spatial frequency characteristics of G pixels, in the vertical and horizontal directions, the highest frequency can be collected

, while in the oblique direction, the signal component with the highest frequency (1/4) fs can be collected. Regarding the spatial frequency characteristics of R and B pixels, in the vertical and horizontal directions, the highest frequency can be collected

, while in the oblique direction, the signal component with the highest frequency (1/2) fs can be collected.

In the above-described embodiments, the spatial frequency characteristics of the G pixels shown in FIGS. 23, 26, and 29 are theoretical frequency characteristics, assuming that the camera signal processing system discussed below (the camera signal processing circuit 1014 shown in FIG. 20) performs Interpolation process for interpolating G pixels at R and B spatial positions.

30A shows the spatial frequency characteristics (spatial resolution) of G pixels of color patterns according to the sixth to eighth embodiments and the spatial frequency of G pixels of known color patterns (Bayer pattern and Bayer pixel shift pattern) Comparison results between properties. 30B shows the spatial frequency characteristics (spatial resolution) of R and B pixels of color patterns according to the sixth to eighth embodiments and R and B pixels of known color patterns (Bayer pattern and Bayer pixel shift pattern). The result of the comparison between the spatial frequency characteristics of the pixels.

30A shows that the spatial frequency characteristics of the G pixels of the color patterns of the sixth and seventh embodiments in the 45° oblique direction are the same as the corresponding spatial frequency characteristics of the Bayer pixel shift pattern, wherein the corresponding spatial frequency characteristics of the Bayer pixel shift pattern The frequency characteristics are higher than the corresponding spatial frequency characteristics of the Bayer pattern; and the spatial frequency characteristics of the G pixels of the color patterns of the sixth and seventh embodiments in the vertical and horizontal directions are higher than the corresponding spatial frequency characteristics of the Bayer pattern, wherein Bayer The corresponding spatial frequency characteristic of the Bayer pattern is higher than that of the Bayer pixel shift pattern.

30A also shows that the spatial frequency characteristics of the G pixels of the color pattern of the eighth embodiment in the vertical and horizontal directions are the same as the corresponding spatial frequency characteristics of the Bayer pattern, wherein the corresponding spatial frequency characteristics of the Bayer pattern are larger than those of the Bayer pixels. The corresponding spatial frequency characteristic of the shift pattern is high; while the spatial frequency characteristic of the G pixel of the eighth embodiment in the 45° oblique direction is higher than the corresponding spatial frequency characteristic of the Bayer pixel shift pattern, wherein the corresponding spatial frequency characteristic of the Bayer pixel shift pattern The characteristic is higher than the corresponding spatial frequency characteristic of the Bayer pattern.

30B shows: the spatial frequency characteristics of the R and B pixels of the color pattern of the sixth embodiment in the vertical and horizontal directions are the same as the corresponding spatial frequency characteristics of the Bayer pattern, wherein the corresponding spatial frequency characteristics of the Bayer pattern are larger than those of the Bayer pattern. The corresponding spatial frequency characteristics of the pixel shift pattern are low; while the spatial frequency characteristics of the R and B pixels of the color pattern of the sixth embodiment in the 45° oblique direction are lower than the corresponding spatial frequency characteristics of the Bayer pixel shift pattern, wherein the Bayer pixel The corresponding spatial frequency characteristic of the moving pattern is higher than that of the Bayer pattern.

30B also shows that the spatial frequency characteristics of the R and B pixels of the color pattern of the seventh embodiment in the vertical and horizontal directions are lower than the corresponding spatial frequency characteristics of the Bayer pattern, wherein the corresponding spatial frequency characteristics of the Bayer pattern are lower than those of the Bayer pattern. The corresponding spatial frequency characteristics of the Er pixel shift pattern are low; while the spatial frequency characteristics of the R and B pixels of the color pattern of the seventh embodiment in the 45° oblique direction are the same as those of the sixth embodiment.

30B also shows that the spatial frequency characteristics of the R and B pixels of the color pattern of the eighth embodiment in the vertical and horizontal directions are lower than the corresponding spatial frequency characteristics of the Bayer pattern, wherein the corresponding spatial frequency characteristics of the Bayer pattern are lower than those of the Bayer pattern. The corresponding spatial frequency characteristics of the Bayer pixel shift pattern are low; while the spatial frequency characteristics of the R and B pixels of the color pattern of the eighth embodiment in the 45° oblique direction are lower than those of the Bayer pixel shift pattern.

These comparison results show that: the G pixels of the color patterns of the sixth and seventh embodiments in the vertical and horizontal directions, and the G pixels of the color pattern of the eighth embodiment exhibit a larger color pattern than the known color patterns in the 45° oblique direction. (Bayer pattern and Bayer pixel shift pattern) higher spatial frequency characteristics. Since the G pixel is the base pixel used to generate the luminance (Y) component, the higher spatial frequency characteristic of the G pixel is very helpful for improving the resolution.

In contrast, the spatial frequency characteristics of the R and B pixels are lower than those of known color patterns (Bayer patterns and Bayer pixel shift patterns). However, as mentioned above, the human eye is less sensitive to R and B colors, so the lower characteristics of R and B colors do not seriously affect the resolution improvement.

As discussed above, pixels including photoelectric transducers therein are two-dimensionally arranged in a matrix, and color filters including a G component and other R and B components as main components for generating a luminance (Y) component In a solid-state imaging device arranged on the surface of pixels, color filters are arranged so that G pixels surround R and B pixels. Therefore, the G pixels are arranged in all the rows and all the columns of the color filter, thereby increasing the spatial frequency characteristic (spatial resolution) of the G component to which the human eye has high sensitivity.

More specifically, the resolution of the G pixels in the vertical and horizontal directions is higher than or equal to the corresponding resolution of the Bayer pattern, wherein the corresponding resolution of the Bayer pattern is higher than the corresponding resolution of the Bayer pixel shift pattern; and the G pixel The resolution in the 45° oblique direction is higher than or equal to the corresponding resolution of the Bayer pixel shift pattern, wherein the corresponding resolution of the Bayer pixel shift pattern is higher than that of the Bayer pattern. As a result, higher resolution can be exhibited not only for achromatic objects but also for colored objects.

In addition, unlike the known Bayer pixel shift pattern, it is not necessary to adjust the RGB balance in the camera signal processing system described below to overcome the lower resolution of the G pixel, thereby preventing the deviation in the level balance. Causes the occurrence of wrong colors.

Specifically, in the color pattern of the sixth embodiment, R and B pixels are arranged in alternate rows and columns so that the sampling rate of R and B pixels in the vertical and horizontal directions becomes 1/2 of that of G pixels. Therefore, the resolution of the G pixels in the vertical and horizontal directions is higher than the corresponding resolution of the Bayer pattern, wherein the corresponding resolution of the Bayer pattern is higher than that of the Bayer pixel shift pattern; while the G pixel is tilted at 45° The resolution in the direction is comparable to the resolution of the Bayer pixel shift pattern, wherein the corresponding resolution of the Bayer pixel shift pattern is higher than that of the Bayer pattern.

In the color pattern of the seventh embodiment, R and B pixels are arranged in alternate rows and columns so that the sampling rate of R and B pixels in the vertical and horizontal directions becomes 1/4 of that of G pixels. Therefore, like the color pattern of the sixth embodiment, the resolution of the G pixels in the vertical and horizontal directions is higher than the corresponding resolution of the Bayer pattern, wherein the corresponding resolution of the Bayer pattern is higher than that of the Bayer pixel shift pattern. resolution; while the resolution of the G pixels in the 45° oblique direction is comparable to the corresponding resolution of the Bayer pixel shift pattern, wherein the corresponding resolution of the Bayer pixel shift pattern is higher than that of the Bayer pattern.

In the color pattern of the eighth embodiment, in the oblique pixel pattern in which even-numbered pixels are shifted from odd-numbered pixels in the row and column directions by 1/2 pixel pitch, R and R are arranged every other row and column. B pixels, so that the sampling rate of R and B pixels in the vertical and horizontal directions becomes 1/4 of the sampling rate of G pixels. Therefore, like the color pattern of the eighth embodiment, the resolution of the G pixels in the vertical and horizontal directions is comparable to the corresponding resolution of the Bayer pattern, which is higher than that of the Bayer pixel shift pattern. The corresponding resolution; while the resolution of the G pixel in the direction inclined at 45° is twice the corresponding resolution of the Bayer pattern.

倍。因此，可以获得方格子像素图案的两倍多的像素信息，即，与方格子像素图案相比可以获得更高的分辨率。如果需要与方格子像素图案分辨率相同的分辨率，则可以在更宽的节距上布置像素。结果，可以增加像素孔径(pixel aperture)，使得可以增加像素的光感度，从而获得具有高S/N比的信号。Specifically, in the color pattern of the eighth embodiment with a slanted pixel pattern, the sampling rate is that of the Bayer pattern

times. Therefore, more than twice as much pixel information as the checkered sub-pixel pattern can be obtained, ie, higher resolution can be obtained compared with the checkered sub-pixel pattern. If the same resolution as the checkered sub-pixel pattern is required, the pixels can be arranged on a wider pitch. As a result, the pixel aperture can be increased, so that the light sensitivity of the pixel can be increased, thereby obtaining a signal with a high S/N ratio.

In the color pattern of the sixth embodiment, FIG. 23 shows that the spatial frequency characteristics of the R and B pixels are 1/2 of the spatial frequency characteristics of the G pixel in the vertical and horizontal directions, and that in the 45° oblique direction , which is 1/4 of the spatial frequency characteristic of a G pixel. In the color pattern of the seventh embodiment, FIG. 26 shows that the spatial frequency characteristics of the R and B pixels are 1/4 of that of the G pixel in the vertical and horizontal directions, and that in the 45° oblique direction, It is 1/4 of the spatial frequency characteristic of the G pixel. In the color pattern of the eighth embodiment, FIG. 29 shows that the spatial frequency characteristics of the R and B pixels are 1/4 of that of the G pixel in the vertical and horizontal directions, and that in the 45° oblique direction, It is 1/2 of the spatial frequency characteristic of the G pixel.

However, human visibility characteristics demonstrate that the human eye tends to recognize high luminance resolutions but struggles to recognize high color resolutions. Therefore, the above-mentioned spatial frequency characteristics of the R and B pixels are sufficient.

For example, in a general television signal format, the frequency band of a chroma (C) signal is 1/4 of that of a luminance (Y) signal, which is sufficient for color resolution. The color patterns of the sixth to eighth embodiments make the best use of this characteristic.

That is, according to the color patterns of the sixth to eighth embodiments, the color space frequency characteristics are reduced to the lowest level that does not give the human eye an unnatural feeling, and instead, are comparable to the known color patterns. Greatly increases luminance spatial frequency characteristics.

In addition, in the color patterns of the sixth to eighth embodiments, G pixels are arranged in all rows and all columns. This is very helpful for increasing interpolation accuracy when performing interpolation processing for interpolating G pixels at R and B spatial positions in the camera signal processing system, as will be discussed below.

A solid-state imaging device including a color filter having the color patterns of the sixth to eighth embodiments is suitable for use as an imaging device such as a digital still camera and a video camera, and is particularly suitable as an imaging device in a video camera.

imaging device

FIG. 20 is a block diagram illustrating an example of the configuration of an imaging device according to an embodiment of the present invention. In FIG. 20 , an optical system such as a lens 1011 forms an image corresponding to light reflected by an object (not shown) on an imaging surface of an imaging device 1012 . As the imaging device 1012, a pixel in which a pixel including a photoelectric transducer is arranged two-dimensionally, and a color filter having a color component as a main component for generating a luminance component and other color components is arranged on the surface of the pixel is used solid-state imaging devices. In this case, a color filter having a color pattern according to the sixth, seventh, or eighth embodiment is used.

The imaging apparatus shown in FIG. 20 includes a lens 1011 , an imaging device (solid-state imaging device) 1012 , an analog-to-digital (A/D) converter 1013 , and a camera signal processing circuit 1014 . The camera signal processing circuit 1014 includes an optical system correction circuit 1021, an interpolation processing circuit 1022, a luminance (Y) signal processing circuit 1023, a color (C) signal processing circuit 1024, a band limiting low-pass filter (LPF) 1025, and decimation processing circuit 1026.

Of the light incident on the imaging device 1012, only the light components corresponding to the color components of the color filters pass through the color filters and are incident on the pixels. Then, the light incident on the pixel is converted into an electrical signal by a photoelectric transducer such as a photodiode, and read as an analog signal. Then, the analog signal is converted into a digital signal by the A/D converter 1013 , and the digital signal is input to the camera signal processing circuit 1014 .

The optical system correction circuit 1021 performs corrections to the sensor or optical system, such as for correcting digital clamping for black level adjustment or for defect correction of the imaging device 1012 for defect correction, and for correcting for edge illumination obscuration (eclipse ) to correct the shading correction of the lens 1011.

The interpolation processing circuit 1022 generates three plane signals from RGB signals that are spatially out of phase with each other to generate RGB signals having the same spatial position, and supplies the RGB signals to the Y signal processing circuit 1023 and the C signal processing circuit 1024 . The Y signal processing circuit 1023 generates a luminance (Y) signal according to the above formula (1). The C signal processing circuit 1024 generates color difference signals Cr(R-Y) and Cb(B-Y) according to equation (2).

Cr=(R-Y)=R-(0.3R+0.6G+0.1B)

Cb＝(B-Y)＝B-(0.3R+0.6G+0.1B) (2)

The band-limited LPF 1025 is a filter whose cutoff frequency is equal to 1/8 of the sampling frequency fs as shown in Figure 31. This filter attenuates (drops) Cr and Cb color difference signal. Output the Cr and Cb color difference signals of the band-limited LPF 1025 according to the TV signal format. If the Cr and Cb color-difference signals are output without performing band limitation, a signal with a frequency of 1/8 fs or higher is output as an error signal. As shown in FIG. 32, the extraction processing circuit 1026 extracts Cr and Cb color difference signals.

The reason for decimating the Cr and Cb color-difference signals is that the Cr and Cb color-difference signals require only 1/4 of the frequency band of the luminance (Y) signal. This is determined by the format of the television signal, and also because the human eye readily recognizes high-frequency luminance (Y) signals but has difficulty recognizing high-frequency chrominance (C) signals.

When observing the difference between Y:Cr:Cb=4:4:4 (the output frequency bands of Y and C signals are the same) and Y:Cr:Cb=4:1:1 as the output signal, except when using red Except for special objects such as objects with blue point light sources, it is difficult to recognize the difference in the output image of the object. That is, if the C signal has 1/4 of the frequency band of the Y signal, a sufficient level of resolution can be obtained, which is determined by the television signal format.

This means that the color pattern according to the sixth, seventh, or eighth embodiment can be used as a color filter pattern of a color filter serving as the imaging device 1012 of a solid-state imaging device.

More specifically, the spatial frequencies of the R and B color components, which are different from the G component that is the primary color component used to generate the luminance component, are as follows. In the color pattern of the sixth embodiment, as shown in FIG. 23, the spatial frequencies of the R and B color components are 1/2 in the vertical and horizontal directions and 1/4 in the 45° direction. In the color pattern of the seventh embodiment, as shown in FIG. 26, the spatial frequencies of the R and B color components are 1/4 in the vertical and horizontal directions and 1/4 in the 45° direction. In the color pattern of the eighth embodiment, as shown in FIG. 29, the spatial frequencies of the R and B color components are 1/4 in the vertical and horizontal directions and 1/2 in the 45° direction. Therefore, any color pixel satisfies the condition of Y:Cr:Cb=4:1:1.

As described above, by using the color pattern of the sixth, seventh, or eighth embodiment as the imaging device 1012 for an imaging device such as a digital still camera or a video camera (especially a video camera), it is possible to increase human Spatial frequency characteristics (spatial resolution) of the G component to which the eye has high sensitivity. As a result, high luminance resolution can be obtained not only for achromatic objects but also for colored objects.

Additionally, unlike known Bayer pixel shifting patterns, there is no need to adjust the level balance of the RGB pixels in the camera signal processing circuit 1014 to overcome the lower resolution of the G pixels. In addition, circuit operations can be simplified, and occurrence of wrong colors caused by deviations in level balance can be prevented.

In any of the color patterns of the sixth, seventh, and eighth embodiments, G pixels surround R and B pixels. Therefore, since the G pixels are arranged in all the rows and all the columns, the interpolation accuracy when the G pixels are interpolated at the R and B spatial positions in the interpolation processing circuit 1022 of the camera signal processing circuit 1014 is greatly improved .

It should be understood by those skilled in the art that various changes, combinations, sub-combinations and alterations may occur depending on design requirements and other factors insofar as they are within the scope of the appended claims or the equivalents thereof.

Claims (10)

1, a kind of solid-state imaging apparatus comprises:

Pixel, comprise by bidimensional be arranged in photoelectric transformer in the matrix; And

Colour filter comprises primary color component and other color component as the fundamental component that is used to generate luminance component, and this colour filter is disposed on the surface of pixel,

Wherein primary color component and other color component are arranged such that primary color component surrounds other color component.

2, according to the solid-state imaging apparatus of claim 1, wherein other color component is arranged by interlacing with every row, make the sampling rate of other color component on vertical and horizontal direction become primary color component sampling rate 1/2.

3, according to the solid-state imaging apparatus of claim 2, wherein other color component has two kinds of colors, and the row of going and comprising primary color component and other color component that only comprises primary color component is by arranged alternate.

4, according to the solid-state imaging apparatus of claim 1, wherein other color component is arranged by interlacing with every row, make the sampling rate of other color component on vertical and horizontal direction become primary color component sampling rate 1/4.

5, according to the solid-state imaging apparatus of claim 4, wherein, other color component has two kinds of colors; And

The row that only comprises primary color component arranged by interlacing, and the row of other color component of going and comprising primary color component and another kind of color of other color component that comprises primary color component and a kind of color is by arranged alternate.

6, according to the solid-state imaging apparatus of claim 4, wherein, every capable pixel is by from adjacent lines pixel shift 1/2 pixel pitch, and every row pixel is by from adjacent column pixel shift 1/2 pixel pitch.

7, according to the solid-state imaging apparatus of claim 6, wherein, other color component has two kinds of colors; And

The row that only comprises primary color component arranged by interlacing, and the row of other color component of going and comprising primary color component and another kind of color of other color component that comprises primary color component and a kind of color is by arranged alternate.

8, a kind of imaging device comprises:

Solid-state imaging apparatus, it comprises having as the primary color component of the fundamental component that is used to generate luminance component and the colour filter of other color component, this primary color component and other color component are arranged such that primary color component surrounds other color component;

Optical system, be configured on the imaging surface of solid-state imaging apparatus to form corresponding to from the image of light of object; And

Signal processing circuit is configured to handle the output signal of solid-state imaging apparatus.

9, a kind of solid state image pickup device comprises:

Pel array, has the oblique pixel pattern that pixel is wherein arranged with being tilted, the capable vertical signal line of odd-numbered in the capable vertical signal line group of odd-numbered is connected to each row of the capable pixel of odd-numbered, and the capable vertical signal line of even-numbered in the capable vertical signal line group of even-numbered is connected to each row of the capable pixel of even-numbered;

Row selector is configured to select respectively the odd-numbered of oblique pixel pattern capable and even-numbered is capable;

Odd-numbered ranks treatment circuit group, it comprises column processing circuit and is connected to the capable vertical signal line group of odd-numbered, is used for the signal plus with the capable pixel of odd-numbered between a plurality of row;

Even-numbered ranks treatment circuit group, it comprises column processing circuit and is connected to the capable vertical signal line group of even-numbered, is used for the signal plus with the capable pixel of even-numbered of pixel column; And

Column selector is configured to select the column processing circuit of odd-numbered ranks treatment circuit group and the column processing circuit of even-numbered ranks treatment circuit group.

10, a kind of imaging device comprises:

Solid state image pickup device, has the oblique pixel pattern that pixel is wherein arranged with being tilted, on having horizontal direction on an adjacent n pixel and the vertical direction in the zone of an adjacent n pixel, capable and even-numbered is capable and with y pixel repeated addition on pixel of the x on the horizontal direction and the vertical direction at the odd-numbered of oblique pixel pattern respectively, simultaneously will have on the horizontal direction a regional mobile m pixel of an adjacent n pixel on the adjacent n pixel and vertical direction on the horizontal or vertical direction, this x pixel has identical color with y pixel, wherein, n is three or bigger odd number, m is three or bigger odd number, and n 〉=x 〉=y, odd-numbered is capable has on the horizontal direction the capable spatial relation between the zone of an adjacent n pixel that has on the horizontal direction on the adjacent n pixel and vertical direction in the zone of an adjacent n pixel and even-numbered on the adjacent n pixel and vertical direction make their m pixels that be shifted each other on the incline direction of oblique pixel pattern;

Mode setting unit, be configured to optionally to be provided for independently to read all pixels in the oblique pixel pattern signal gradual read mode and be used for pixel addition read mode with the pixel addition of oblique pixel pattern; And

Driver element is configured to drive solid state image pickup device according to gradual read mode or pixel addition read mode by the mode setting unit setting.

Images