JP3410638B2 - Video camera system - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a single-panel color video camera using a CCD image pickup device. 2. Description of the Related Art FIG. 2 shows a configuration of a conventional camera system using a CCD image pickup device having a primary color array color filter. In a conventional camera system, a lens (1)
The image of the subject from which the light enters is attenuated by the optical filter (2) to attenuate the high frequency components contained in the image, and then the CC
The photoelectric conversion is performed by the D imaging device (3), and a signal is output for each pixel. [0004] The CCD output is converted into a digital signal by an AD converter (5) after a video signal is separated and a signal level is adjusted by a CDS / AGC circuit (4). A digital signal processing circuit (6) creates a luminance signal and a chrominance signal from the video signal converted into a digital value, and outputs it from terminals (24) and (25). In the digital signal processing circuit (6), the black level value included in the signal is fixed to a specified value by the LCLIP (7) which is a black level clamp circuit, and then the white balance circuit (8) supplies the signal to the white balance circuit (8). White (hereinafter referred to as R), green (hereinafter referred to as G), and blue (hereinafter referred to as B) pixel signals are subjected to white balance adjustment. The output of the white balance circuit (8) is 1
1HDL (9), (1
0) and (11). First, processing of the luminance signal is performed by (8), (9),
With respect to the output signals of (10) and (11), the CCD output is switched to G, R and B, R for each line, so that a vertical low-pass filter YVLPF ( 12), the horizontal low-pass filter YHLPF (13) removes the color signal carrier component in the horizontal direction. The output of the YHLPF (13) is gamma-corrected by a luminance γ correction circuit YGAM (14), and a synchronization signal is added by a sync signal addition circuit SYNCMIX (15). In the color signal processing, gamma correction is applied to each of the synchronized pixel signals (8), (9) and (10) by the color γ correction circuit CGAM (17), and then the color difference is calculated. In the block of CG (18) which is a signal generation circuit, R / G and B are determined by R / G and B / G adjacent in the vertical direction.
-G is created. Further, the RGB (18) output RGB, B
A color difference matrix operation is performed on the -G signal by the matrix MATRIX (22), and RY and BY are calculated.
Is created. The color difference signal data RY and BY are CGAI
After the gain adjustment is performed in the N (23) block, the burst flag mix circuit BFMIX (24)
The burst signal data is superimposed on CMOD (25)
Is subjected to color signal modulation. In this system, since the output of the CCD image pickup device is not interlaced, INTRCONV (2
2) and (23) perform non-interlace-interlace conversion and output in a predetermined video signal format. In FIG. 2, a vertical low-pass filter YVLPF (12) is applied to each pixel signal of RGB with white balance to create a luminance signal.
And a horizontal low-pass filter YHLPF (13),
At this time, in order to secure the horizontal resolution as much as possible, a filter is formed based on the Nyquist sampling theorem so that the pass band is almost up to half the band of the pixel frequency. FIG. 3 shows the horizontal frequency spectrum of the luminance signal at this time. FIG. 3A shows the signal value of each color filter when an achromatic subject is imaged, FIG. 3B shows the luminance signal spectrum in this case, and FIG. Signal value of each color filter when capturing a small number of subjects,
FIG. 3D shows the luminance signal spectrum in this case.
FIG. 3E shows signal values of the respective color filters when a subject having only R components is captured, and FIG. 3F shows a luminance signal spectrum in this case. As shown in FIG. 3A, when the subject is an achromatic subject and the white balance is maintained, the RGB pixel signal values of the color filter are the luminance signal values at each pixel position of the image. Shows true. Therefore, when the horizontal pixel period is T, the horizontal luminance signal spectrum of the CCD is (2π / T).
A luminance signal spectrum is generated around a frequency that is an integral multiple of. FIG. 3B shows this spectrum.
And the signal band without aliasing distortion is in the range of (1). The band of the horizontal low-pass filter YHLPF for creating a luminance signal is represented by (2) in the figure. As shown in FIG. 3 (e), when the pixel signal has only the R component, the sampling position of the pixel has a period of 2T in the horizontal direction. As a result, in the horizontal frequency spectrum of the image, a luminance signal spectrum is generated around a frequency that is an integral multiple of (2π / 2T). When the same horizontal low-pass filter YHLPF as that shown in FIG. 3B is used for this signal, a component (5) generated at (2π / 2T) as shown in (6) is generated in the baseband signal. It overlaps with the high frequency band, and this becomes aliasing distortion, and a luminance false signal is generated, thereby deteriorating the image quality. Further, as in the case where the white balance is shifted in FIG. 3C, the signal value of R is slightly larger than that of GB, that is, the value of the ratio of R / G or B / G is slightly different from 1. In the case of, as shown in the frequency spectrum of FIG. 3D, according to the difference between the signal values of GB and R, (2π /
2T), the amount of the component (3) changes, and similarly, the amount of the aliasing distortion (4) also changes. As described above, when the characteristics of the YHLPF are expanded to near (2π / 2T) in order to obtain a high-resolution luminance signal, aliasing distortion occurs in a chromatic object, and the image quality is greatly deteriorated. Will be. Also, this aliasing distortion is the difference between the signal values of R and GB or B and RG
It can be seen that the larger the difference between the signal values of and, the more remarkable it appears. Although a method of narrowing the band of the YHLPF to prevent the occurrence of the aliasing distortion is conceivable, a sufficient resolution cannot be obtained and a blurred image cannot be obtained. According to the present invention, a color filter array of a CCD image pickup device of a color video camera is arranged in the vertical direction of (green, red), (blue, green) in two horizontal pixels, In a color video camera system using a CCD image sensor having a function of outputting a signal converted by a photoelectric conversion element of a CCD independently on a pixel basis with a color array having a periodicity of two pixels in a vertical direction, A means for generating a luminance signal having a wide band, a means for generating a luminance signal having a narrow horizontal spatial frequency band, and a method of normalizing the signal values of the color difference signals RG and BG, and then summing the respective squares , And a multiplier and an adder that change the addition ratio of the horizontal broadband luminance signal and the horizontal narrowband luminance signal according to the value of the sum of squares. The object of the present invention is to prevent the image quality of the luminance component from being degraded by the aliasing distortion caused when the chromatic object or white balance is shifted due to the color filter arrangement of the CCD. The achromatic portion of the image is adjusted by adjusting the addition ratio of the band luminance signal and the high-band luminance signal that provides high resolution when the achromatic subject and the white balance are optimally adjusted according to the coloring amount of the subject. A video camera characterized by being able to obtain a high-resolution luminance signal and a luminance signal having no aliasing distortion in a chromatic portion, and to obtain an image with high resolution and little image quality deterioration due to aliasing distortion. It is. FIG. 4 shows a frequency spectrum characteristic of a luminance signal according to the present invention. FIG. 4A shows YHLP.
A spectrum of a high-band luminance signal created by F, including aliasing distortion in the case of a chromatic subject. FIG. 2B shows the YL block (19) shown in FIG.
T) Interpolation processing is performed by a low-pass filter (3) having a band equal to or smaller than T), and a spectrum of a low-pass luminance signal created with an addition ratio of YL = 0.30 × R + 0.59 × G + 0.11 × B ( 2)
And does not include folding distortion. As shown in FIG. 3, in the case of an achromatic color, the ratio of R / G and B / G is both 1. At this time, (2) in (a)
The aliasing component (4) generated around the frequency of (π / 2T) is 0, and in a chromatic object, the component (4) increases as the R / G or B / G ratio moves away from 1. Therefore, a coefficient relating to the color density is obtained from the color difference signal for each pixel, and when the color is light, the spectrum (a)
Is increased to create a high-resolution luminance signal. When the color is dark, the low-frequency luminance signal component of the spectrum (b) without aliasing distortion is increased, and the luminance signal is calculated by the sum of both luminance signal components. create. (C), (d) and (e) show the frequency spectrum of the luminance signal changed according to the color density.
(C) is an achromatic color, that is, a luminance signal of a subject whose R / G and B / G ratios are both 1, and (e) is a chromatic R / G signal.
Alternatively, the spectrum of the luminance signal when the B / G ratio is far from 1 is shown, and (d) is the spectrum when the ratio is intermediate. From this figure, it can be seen that the resolution decreases in the order of (c), (d) and (e), but the aliasing distortion becomes very small, so that an image with little image quality deterioration can be obtained. FIG. 1 shows the configuration of the present invention. In FIG. 1, an image of a subject incident from a lens (1) is subjected to photoelectric conversion by a CCD image pickup device (3) after attenuating high frequency components contained in the image by an optical filter (2). , A signal is output for each pixel. After the CCD output is separated by the CDS / AGC circuit (4) and the signal level is adjusted, the CCD output
The signal is converted into a digital signal by the converter (5). A digital signal processing circuit (6) creates a luminance signal and a chrominance signal from the video signal converted into the digital value,
It is output from the terminals (27) and (28). In the digital signal processing circuit (6), LC
After the value of the black level included in the signal is fixed to a specified value in the LIP (7), red (hereinafter R), green (hereinafter G),
White balance adjustment is performed on each pixel signal of blue (hereinafter, referred to as B). The output of the white balance circuit (8) is 1
1HDL (9), (1
By 0) and (11), signals of three lines adjacent in the vertical direction are synchronized. The processing of the luminance signal is first performed by (8), (9),
With respect to the output signals of (10) and (11), the CCD output is switched to G, R and B, R for each line, so that a vertical low-pass filter YVLPF ( After the application of 12), the horizontal color signal carrier component is removed by the wideband horizontal low-pass filter YHLPF (13). In the color signal processing, gamma correction is applied to each of the synchronized pixel signals (8), (9) and (10) by the color γ correction circuit CGAM (17).
RG and BG are created by R, G, and B adjacent in the vertical or horizontal direction in the -G (18) block. The RG and BG signals output from the CG (18) are subjected to a color difference matrix operation in a matrix MATRIX (22) to create RY and BY. The color difference signal data RY and BY are CGAIN (2
3) After the gain is adjusted in the block, BFMIX
In (24), the burst signal data is superimposed and CMOD
The color signal is modulated in (25). The output of the CGAM (17) block is input to a narrow-band luminance signal generation circuit YL (19) block to generate a low-frequency luminance signal. Further, CG (1
The luminance signal addition ratio operation circuit YH / YLGAIN (20) block outputs the YHLPF (13) wide-band luminance signal output and Y based on the color difference component RG / BG signal output from 8).
A value of the addition ratio of the narrowband luminance signal output of L (19) is created. Broadband luminance signal output of YHLPF (13) and YL
The narrow-band luminance signal output of (19) is output from the luminance signal addition circuit YH.
/ YH / YLGAIN (20) at / YLMIX (21)
After being multiplied by the respective coefficients with the output values of, the signals are added to generate a luminance signal. The output of YH / YLMIX (21) is luminance γ
The gamma correction is performed by the correction circuit YGAM (14), and then the synchronization signal is added by the synchronization signal addition circuit SYNCMIX (15). In this system, since the output of the CCD image pickup device is not interlaced, INTRCONV (1
6) and (26) perform non-interlace-interlace conversion and output in a predetermined video signal format. FIG. 5 is a diagram for explaining the details of the YHLPF (13) block. In this block, YVLPF (1
The output signal of 2) is passed through a wideband low-pass filter to output a wideband luminance signal. (B) shows an example of the frequency characteristics of the filter. The circuit configuration is constituted by an FIR filter as shown in FIG. FIG. 6 shows details of YL (19). D output from the color γ correction circuit CGAM (17)
The signal data of 0H, D1H, and D2H are as shown in (a), (b), and (c) of the timing chart of (B). The signals (a) and (c) of D0H and D2H are added by the adder (1) to become a signal (d). (2),
The selector (3) selects the signal (d) or (b) for each horizontal scanning period, and the output (e) of (2) is R
/ G pixel signal line, output (f) of (3) is B
Only the line having the / G pixel signal is output. Next, the selectors (4), (5) and (6) output the R pixel signal to the output of the selector (4), the B pixel signal to the selector (5), and the G pixel signal to the selector (6). Is output to At this time, since data exists only for every other pixel for R and B, 0 is inserted at a pixel position where there is no data. The output signals are shown in (g), (h) and (i).
The signals of (g), (h) and (i) are (10), (11),
(12), (16), (17), and (18) adders and (7), (8), (9), (13), (14), (1)
Interpolated by a horizontal low-pass filter composed of a delay circuit for one pixel clock of (5) (j), (k),
It becomes the signal of (l). The horizontal filter characteristic at this time is shown in FIG. Due to the characteristic of (C), the band of each RGB color signal is (2π
/ 4T) and contains almost no aliasing component. The interpolated RGB signals (j), (k), and (l) are added to the narrow-band luminance signal (m) at an addition ratio of YL = 0.30 × R + 0.59 × G + 0.11 × B. ) To create. This operation is performed by using the multipliers of (19), (20), and (21), and
This is performed by the adders (22) and (23). FIG. 7 shows a luminance signal addition ratio calculation circuit YH / YL.
It is a detailed view of GAIN (20). In the block diagram (A), (1),
In the divider of (2), input RG, BG signals (a),
(B) is standardized by coefficients R-Gmax and B-Gmax. These outputs (c) and (d) are output from multipliers (3),
(4) is squared. The outputs (e) and (f) of the multipliers are added by the adder (5) to become a coefficient YLgain indicating the gain of the narrow-band luminance signal YL. Also, the code converter (6),
By subtracting the value of (g) from 1 in the block of the adder (7), the coefficient YHg indicating the gain of the broadband luminance signal is obtained.
Create ain. FIG. 7B is a color difference vector diagram with RG and BG as axes. In the case of an achromatic pixel in the image, the pixel is located at the center point (a) of the vector diagram, and the color difference vector moves to the position shown by the circle (b) as the color density increases. Therefore, after normalizing the RG and BG signals so that the maximum value becomes 1, each value is set to 2
When the power is raised, the characteristic on the paraboloid which approaches 0 at the center point (a) of the vector diagram and approaches 1 toward the periphery is obtained as shown in FIG. By setting this as the addition coefficient YLgain of the narrowband luminance signal and YHgain = 1−YLgain, the addition coefficient YHgain of the wideband luminance signal is obtained.
Create FIG. 8 shows a luminance signal adding circuit YH / YLMIX.
It is a detailed view of (21). YHLPF (13) output is Y
H input (a) is input, and YL (19) output is YL
Leads to input (b). In addition, YH / YLGAIN (2
0) are output as (c) and (d), respectively. The arithmetic expression in this block is represented by Y = YH × YHgain + YL × YLgain, and the value of Y is output as OUT (e). At this time, since YHgain + YLgain = 1 is always established from the processing contents of the block of YH / YLGAIN (20), the DC components are YHgain and YLga.
Even if the value of “in” changes, the luminance signal does not change due to the change of the addition coefficient, and only the high frequency component changes. As described above, according to the present invention, when a conventional method is used to obtain a high-resolution luminance signal, aliasing distortion occurs in chromatic colors, thereby deteriorating the image quality and causing aliasing distortion. The use of a narrow-band luminance signal that does not provide an image can improve the problem that only an image without sharpness can be obtained, and can provide an image with sharpness and without image quality deterioration due to aliasing distortion in a chromatic image region.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 shows an embodiment of the present invention. FIG. 2 shows a configuration of a conventional single-panel color video camera. FIG. 3 is an explanatory diagram of a frequency spectrum change of a luminance signal depending on the color depth of a subject. FIG. 4 shows frequency spectrum characteristics in the present system. FIG. 5 shows a configuration of a wideband horizontal low-pass filter circuit for generating a wideband luminance signal. FIG. 6 shows a configuration of a narrow-band luminance signal generation circuit. FIG. 7 shows a configuration of a luminance signal addition ratio calculation circuit. FIG. 8 shows a configuration of a luminance signal addition circuit. [Description of Signs] 1 lens 2 optical filter 3 CCD image sensor 4 CDS / AGC circuit 5 A / D converter 6 digital signal processing circuit 7 LCLIP (black level clamp circuit) 8 WB (white balance circuit) 9, 10, 11 1 HDL (delay circuit for one horizontal scanning period) 12 YVLPF (luminance vertical LPF) 13 YHLPF (broadband horizontal low-pass filter) 14 YGAM (luminance γ correction circuit) 15 SYNCMIX (synchronous signal addition circuit) 16, 26 INTRCONV (non-inter 17-CGAM (color-gamma correction circuit) 18 CG (color-difference signal generation circuit) 19 YL (narrow-band luminance signal generation circuit) 20 YH / YLGAIN (luminance signal addition ratio calculation circuit) 21 YH / YLMIX (Luminance signal addition circuit) 22 MATRIX ( Matrix circuit) 23 CGAIN (color difference gain adjustment circuit) 24 BFMIX (burst signal mixing circuit) 25 CMOD (color modulation circuit)

Claims (1)

(57) [Claim 1] A lens unit including an optical filter and a CCD
The color filters of the image sensor (green, red)
(Blue, green) Horizontal direction 2 arranged in the vertical direction
Pixel, having a color arrangement having a periodicity of two pixels in the vertical direction,
In a video camera composed of a CCD image pickup device having a function of independently outputting a signal converted by a photoelectric conversion element of a CD for each pixel, an AD converter, a digital signal processing circuit, and a DA converter, a horizontal spatial frequency band Means for creating a luminance signal having a wide horizontal spatial frequency band, means for creating a luminance signal having a narrow horizontal spatial frequency band, and standardization of the respective signal values of the color difference signals RG and BG. A video camera system comprising: means for creating; and a multiplier and an adder for changing an addition ratio between a horizontal broadband luminance signal and a horizontal narrowband luminance signal according to the value of the sum of squares.