JP3710189B2 - Imaging device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an imaging apparatus including a progressive scanning solid-state imaging device in which color filters are arranged in a so-called Bayer array.
[0002]
[Prior art]
In recent years, with the advance of semiconductor technology, solid-state imaging devices (all-pixel reading solid-state imaging devices, hereinafter referred to as all-pixel imaging devices) capable of reading signals by progressive scanning have been developed. These image sensors have less blur even with respect to a moving subject and can capture a high-resolution image compared to conventional interlaced scan (interlaced scan) image sensors. That is, in the interlace imaging device, an image of one frame is composed of two fields that are shifted by one field period. Therefore, when a moving subject is imaged, motion blur occurs due to the time difference between the fields. Further, when an image of one field is extracted, the vertical resolution is halved although motion blur is eliminated. On the other hand, since the all-pixel image sensor can capture one frame in one field period, the above-described problem does not occur.
Utilizing such characteristics, the all-pixel imaging device is expected to be applied to a still image capturing camera, an EDTV II camera, and the like.
[0003]
Most of the all-pixel imaging devices currently used are ones in which R, G, and B filters are arranged in a so-called Bayer array as shown in FIG. In this color arrangement, G is used as a luminance signal. Since the all-pixel image sensor reads out one frame of image data in one field period, the charge transfer rate is doubled as compared with the conventional vertical two-pixel addition readout image sensor. Therefore, as shown in FIG. 4, the output from the image sensor is the same as that of the conventional one-line readout, and the odd-numbered and even-numbered scanning line signal outputs are, for example, the first and second signal outputs, respectively. It is desirable to read.
[0004]
Next, the circuit configuration of a conventional 2-pixel readout all-pixel imaging device will be described with reference to FIGS. In the image pickup apparatus of FIG. 4, all pixel image pickup device 1 performs vertical register transfer and horizontal register transfer according to timing signals t1 and t2 given from TG 9, and odd and even scanning line signals are respectively sent from ch1 and ch2. Is output. Output signals from ch1 and ch2 correspond to odd lines and even lines of the all-pixel imaging device 1, respectively, and are sent to the CDS circuits 31 and 32 via the buffers 21 and 22, respectively. The CDS circuits 31 and 32 remove reset noise from the image sensor and send video signals to the AGC circuits 41 and 42, respectively. The AGC circuits 41 and 42 give a designated gain to the video signal based on the gain difference of each channel detected by the camera process circuit 6. The A / D converters 51 and 52 convert analog video signals into digital signals and send the digital signals to the camera process circuit 6.
[0005]
In FIG. 5, the camera process circuit 6 first receives output signals from the ch1 and ch2 from the input terminals 600-1 and 600-2, respectively. The color separation circuit 601 performs G, B, and R image data sent in a dot-sequential manner with the pixel arrangement as shown in FIG. Is interpolated so that each channel has data for all pixel positions by means of average value interpolation in the horizontal and vertical directions. The main line G is output without being interpolated as shown in FIG. Further, Gv-high of the vertical contour compensation signal system is output after being interpolated in the horizontal direction as shown in FIG.
The output G from the color separation circuit 601 is zero-inserted and interpolated by an SPC-FIL (two-dimensional filter) 612 having a spatial frequency characteristic as shown in FIG. 8, for example, and 1 / 2fs_h and 1 / 2fs_v are trapped. (Fs_h and fs_v are sampling frequencies in the horizontal and vertical directions) The purpose of using SPC-FIL here is to eliminate the moire of 1 / 2fs_o (fs_o is the sampling frequency in the diagonal direction) that appears in the diagonal direction of the luminance signal. The sum of the two sets of checkered coefficients is designed to be equal so as to be suitable for zero insertion interpolation of the dot sequential signal.
[0006]
In FIG. 4, when gain is given to the AGC circuits 41 and 42, the AGC circuits 41 and 42 change the video signal by a predetermined gain. However, the AGC circuits 41 and 42 have individual operation variations. Even if the same gain is specified for 41 and 42, the video signal level will be different. When such a phenomenon occurs, a difference occurs in the output level between the odd / even lines and the odd / even pixels. Therefore, when a uniform image (for example, a white chart) is imaged as shown in FIG. As described above, variation in the gain of the AGC circuit results in grid-like interference on the screen, which significantly degrades the image. However, since 1 / 2fs_v is trapped, the output level difference between the odd-numbered and even-numbered lines is eliminated because it is replaced by the frequency oscillation of 1 / 2fs_v. The output level difference of the even pixels is eliminated because it is replaced by the frequency vibration of 1/2 fs_h.
[0007]
Thereafter, G is branched into a main line signal and a horizontal contour compensation signal, and a horizontal contour signal is extracted by an H-HPF (horizontal high-pass filter) 604. Further, Gv-high of the vertical contour compensation signal system extracts a vertical contour signal by a V-HPF (vertical high-pass filter) 603 having a frequency characteristic as shown in FIG. After the horizontal and vertical contour compensation signals are added, the contour compensation signal is added to the main line signal and sent to the color conversion matrix circuit 607 via the γ processing circuit 606. G-low, B-low, and R-low are white balanced by the white balance circuit 605 and then sent to the color conversion matrix circuit 607 via the γ processing circuit 606. The color conversion matrix circuit 607 receives G, G-low, B-low, and R-low, and performs linear matrix processing and RGB-YCbCr conversion. Outputs Y, Cb, and Cr from the color conversion matrix circuit 607 are output from the output terminal 609Y as the luminance signal Y, and the color difference signals Cb and Cr are subjected to band compression and point sequentialization by the point sequentialization circuit 608. Output from the output terminal 609C.
Then, when the entire screen becomes dark due to shooting at low illuminance or the like, in order to appropriately increase the video signal level, it is sent to the microcomputer 8 of information c1 obtained from G that has passed through an H-LPF (horizontal low-pass filter) 610.
[0008]
The microcomputer 8 determines the gains of the AGC circuits 41 and 42 based on the information c1 received from the camera process circuit 6, and sends a control signal c2. The AGC circuits 41 and 42 receive the control signal c2 and adjust it to a desired gain.
When gain is given to the AGC circuits 41, 42, the AGC circuits 41, 42 change the video signal by a predetermined gain, but the AGC circuits have individual operation variations, and the same gain designation is given to the AGC circuits 41, 42. Even so, the video signal level will be different. Even if such a phenomenon occurs, the SPC-FIL 612 eliminates the trapping of 1 / 2fs_h and 1 / 2fs_v, and a suitable image can be obtained.
[0009]
In this method, the gain variation of the first and second AGC circuits 41 and 42 is replaced with the Nyquist frequency vibration on the horizontal and vertical frequency axes, so that the design of the two-dimensional filter traps 1/2 fs_h, and 1 By adopting a configuration that traps / 2fs_v, it becomes possible to effectively eliminate variations in gains of the first and second AGC circuits without increasing the amount of hardware.
In addition, by effectively suppressing the oblique luminance carrier with a two-dimensional filter, a suitable image with sharpness can be obtained without narrowing the luminance band at the time of image formation with an O-LPF (optical filter). It becomes possible to provide.
[0010]
[Problems to be solved by the invention]
However, in the above-described method, since the luminance signal obtained by performing horizontal interpolation in the color separation circuit 601 is used when the vertical contour signal is obtained, variations in the gains of the first and second AGC circuits occur. In this case, as shown in FIG. 7B, the gain variation of the AGC circuit becomes a horizontal stripe-like component, and the vertical high-frequency component appears concentrated on the Nyquist frequency 1 / 2fs_v. For this reason, there has been a problem that a high-frequency component perpendicular to the entire screen is placed, and the image is remarkably deteriorated.
[0011]
The present invention has been made to solve the above problem, and an object of the present invention is to obtain an imaging apparatus capable of suppressing image quality deterioration due to gain variations of two AGC circuits.
[0012]
[Means for Solving the Problems]
An imaging apparatus according to the present invention includes a color filter arranged in a Bayer array, reads out signals of all pixels by non-addition, and simultaneously outputs odd and even scanning line signals as first and second video signals. Means, first and second gain control means for controlling the gain of the first and second video signals, and separating the G signal as a luminance signal from the gain-controlled first and second video signals. Separating means for interpolating in the horizontal direction and outputting, and a filter for removing a frequency component of ½ of the sampling frequency in the vertical direction of the imaging means from the separated signal interpolated in the horizontal direction to form a vertical contour signal Means.
[0013]
[Action]
According to the configuration of the present invention, the variability of the gains of the first and second gain control means can be obtained without increasing the hardware amount by trapping 1 / 2fs_v by the filter means provided in the vertical contour compensation signal system. It is possible to effectively remove the influence of the above.
[0014]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings. Note that in the imaging apparatus according to the embodiment of the present invention, the processing common to the conventional example of FIG. 4 is performed from the video signal output of the solid-state imaging device to A / D conversion, and therefore the description of FIG. 4 is omitted. Only the camera process circuit 6 according to the present invention in FIG. 1 will be described.
The first and second digital video signals obtained in FIG. 4 are sent to the camera process circuit 6 according to the present invention in FIG. In the camera process circuit 6, first, output signals from the ch 1 and ch 2 are received from the input terminals 600-1 and 600-2 and input to the color separation circuit 601.
[0015]
The color separation circuit 601 performs G, B, and R pixel data sent in a dot-sequential manner with the pixel arrangement as shown in FIG. Is an average value interpolation in the horizontal and vertical directions, and interpolates so that each channel has data for all pixel positions. The main line G is output without being interpolated as shown in FIG. Further, Gv-high of the vertical contour compensation signal system is output after being interpolated in the horizontal direction as shown in FIG.
The output G from the color separation circuit 601 is zero-inserted and interpolated by an SPC-FIL 612 having a spatial frequency characteristic as shown in FIG. 8, for example, and 1 / 2fs_h and 1 / 2fs_v are trapped. Here, SPC-FIL is used for the purpose of eliminating moire of 1 / 2fs_o appearing in the diagonal direction of the luminance signal, and two sets of checkered coefficients are suitable for zero insertion interpolation of the dot sequential signal. Are designed to be equal.
[0016]
When gain is given to the AGC circuits 41 and 42, the AGC circuits 41 and 42 change the video signal by a predetermined gain. However, as described above, these AGC circuits 41 and 42 have individual operation variations. Even if the same gain is designated for the circuits 41 and 42, the video signal level will be different. When such a phenomenon occurs, a difference occurs in the output level between the odd / even lines and the odd / even pixels. Therefore, when a uniform image (for example, a white chart) is imaged as shown in FIG. As described above, variation in the gain of the AGC circuit results in grid-like interference on the screen, which significantly degrades the image. However, since 1 / 2fs_v is trapped, the output level difference between the odd-numbered and even-numbered lines is eliminated because it is replaced by the frequency oscillation of 1 / 2fs_v. The output level difference of the even pixels is eliminated because it is replaced by the frequency vibration of 1/2 fs_h.
[0017]
Thereafter, G branches to a main line signal and a horizontal contour compensation signal, and a horizontal contour signal is extracted by the H-HPF 604. Also, for Gv-high of the vertical contour compensation signal system, for example, a vertical contour signal is extracted by a V-BPF (vertical bandpass filter) 611 having a frequency characteristic as shown in FIG. 2, and 1/2 fs_v is trapped. After the horizontal and vertical contour compensation signals are added, the contour compensation signal is added to the main line signal and sent to the color conversion matrix circuit 607 via the γ processing circuit 606. G-low, B-low, and R-low are white balanced by the white balance circuit 605 and then sent to the color conversion matrix circuit 607 via the γ processing circuit 606.
[0018]
The color conversion matrix circuit 607 receives G, G-low, B-low, and R-low, and performs linear matrix processing and RGB-YCbCr conversion. Outputs Y, Cb, and Cr from the color conversion matrix circuit 607 are output from the output terminal 609Y as the luminance signal Y, and the color difference signals Cb and Cr are subjected to band compression and point sequentialization by the point sequentialization circuit 608. Output from the output terminal 609C.
When the entire screen becomes dark due to shooting at low illuminance or the like, information c1 obtained from G that has passed through the H-LPF 610 is sent to the microcomputer 8 in order to appropriately increase the video signal level.
[0019]
The microcomputer 8 determines the gains of the AGC circuits 41 and 42 based on the information c1 received from the camera process circuit 6, and sends a control signal c2. The AGC circuits 41 and 42 receive the control signal c2 and adjust it to a desired gain.
When gain is given to the AGC circuits 41, 42, the AGC circuits 41, 42 change the video signal by a predetermined gain, but the AGC circuits have individual operation variations, and the same gain designation is given to the AGC circuits 41, 42. Even so, the video signal level will be different. Even if such a phenomenon occurs, the SPC-FIL 612 eliminates the trapping of 1 / 2fs_h and 1 / 2fs_v, and a suitable image can be obtained.
[0020]
In this method, the gain variation of the first and second AGC circuits 41 and 42 is replaced with the Nyquist frequency oscillation on the horizontal and vertical frequency axes, so that the design of the two-dimensional filter traps 1/2 fs_h, and 1 By adopting a configuration that traps / 2fs_v, it becomes possible to effectively eliminate variations in gains of the first and second AGC circuits without increasing the amount of hardware.
In addition, by effectively suppressing the luminance carrier in the oblique direction with the two-dimensional filter, it is possible to provide a suitable image with sharpness without narrowing the luminance band at the time of image formation with the O-LPF. It becomes possible.
[0021]
Here, as described above, when the vertical contour signal is obtained, the luminance signal subjected to the horizontal interpolation is used, so when the gain variation of the first and second AGC circuits 41 and 42 occurs. As shown in FIG. 7B, the gain variation of the AGC circuit becomes a horizontal stripe-like component, and the vertical high-frequency component appears concentrated on the Nyquist frequency 1 / 2fs_v. However, in the present invention, the V-BPF 611 designed to trap 1/2 fs_v is used as a filter for extracting the vertical high-frequency component, so that it is suitable even if the gain variation of the AGC circuit occurs. A contour compensation signal can be obtained, and a sharp image can be maintained.
[0022]
【The invention's effect】
According to the configuration of the present invention, the filter means for trapping 1/2 fs_v in the vertical contour compensation signal system to the Nyquist frequency oscillation on the horizontal and vertical frequency axes is provided for the gain variation of the first and second AGC circuits. Thus, variation in gains of the first and second gain control means in the vertical contour compensation signal system can be effectively removed, and thus a suitable image with sharpness can be provided.
[Brief description of the drawings]
FIG. 1 is a block diagram showing an embodiment of the present invention.
2 is a characteristic diagram showing filter characteristics of the V-BPF of FIG. 1. FIG.
FIG. 3 is a configuration diagram illustrating a color filter array to which the present invention can be applied.
FIG. 4 is a block diagram illustrating an imaging apparatus to which the present invention can be applied.
5 is a block diagram showing a conventional example of the camera process circuit of FIG. 4. FIG.
FIG. 6 is a configuration diagram showing a method of vertical and horizontal interpolation.
FIG. 7 is a configuration diagram showing an interpolation method of a vertical contour compensation signal system.
FIG. 8 is a characteristic diagram showing characteristics of a two-dimensional filter.
FIG. 9 is a characteristic diagram showing characteristics of a vertical HPF.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 All pixel readout image pick-up element 41 1st AGC circuit 42 2nd AGC circuit 6 Camera process circuit 611 V-BPF

Claims (2)

An imaging means having a color filter arranged in a Bayer array, reading out signals of all pixels by non-addition, and simultaneously outputting odd and even scanning line signals as first and second video signals;
First and second gain control means for controlling the gain of the first and second video signals;
Separating means for separating the G signal as a luminance signal from the first and second gain-controlled video signals, interpolating in the horizontal direction, and outputting the separated signal;
An image pickup apparatus comprising: filter means for removing a frequency component ½ of the sampling frequency in the vertical direction of the image pickup means from the separated signal interpolated in the horizontal direction to form a vertical contour signal. .   2. The image pickup apparatus according to claim 1, wherein the separating unit is a color separating unit that separates R, G, and B color signals.

Images