JPH09284604A - Video camera and contour emphasis device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a stable video output with reduced noise by extracting a minimum signal among signals corresponding to a plurality of pixels before and after a video signal and providing an output of the extracted signal to a compression means so as to avoid a level of a contour emphasis signal from being largely changed between pixels even when a level of the video signal is varied between the picture elements. SOLUTION: A low frequency component Si from a low pass filter 171 is given to a minimum value extract circuit 173, which selects a minimum signal among three signals, that is, a signal before being latched by a D flip-flop and signals latched by D flip-flop circuits 174, 175. Thus, a signal with a higher level is not in existence in an output signal Sm of the minimum value extract circuit 173. The output signal Sm of the minimum value extract circuit 173 is compressed for a level by a level depend circuit 178, resulting that a detail signal is not deflected large when the output signal is multiplied with the detail signal and no large noise is produced in a camera output.

Description

Detailed Description of the Invention

[0001]

The present invention relates to a video camera,
And a contour enhancement device used in a video camera or other video equipment.

[0002]

2. Description of the Related Art In a video camera, particularly a commercial or broadcast video camera which is required to output a high-quality video, a process for emphasizing the contour of an image is performed on a video signal obtained from an image sensor. It This contour enhancement processing is called image enhancement, aperture, detail, or the like.

Specifically, in a video camera in which video signals are digitally processed, for example, R (red), G (green), obtained by photographing a subject image with a CCD solid-state image sensor,
The B (blue) channel video signal is digitized, a digitized signal for contour enhancement is generated from the digitized main line video signal, and the generated contour enhancement signal is converted to a main line video signal. Is added.

In this case, when the level of the video signal of the main line is low and the contour emphasis signal is added to this, the S / N (signal-to-noise ratio) of the video signal after the addition of the contour emphasis signal
Gets worse. Therefore, it is possible to suppress the contour emphasis signal when the video signal of the main line is at a low level.

Specifically, a system for controlling the gain of the contour emphasis signal is provided with a compression circuit whose characteristics of the gain of the contour emphasis signal with respect to the input level of the video signal of the main line are as shown in FIG. The main line video signal is supplied to this compression circuit through a low-pass filter. The input level La at which the gain becomes zero and the input level Le at which the compression starts to work
Between and is within the level range of, for example, -7% to 7% of the video signal of the main line.

According to this, when the video signal of the main line that has passed through the low-pass filter is supplied to the compression circuit as the input signal Si, at the low level of the video signal of the main line, the contour emphasis signal of the contour emphasis signal is obtained according to the characteristic of the broken line portion 5a. The gain is reduced and the contour emphasis signal is suppressed.

Regarding the contour enhancement processing of a video camera, Japanese Patent Laid-Open No. 6-46444 discloses that the video signal after the addition of the contour enhancement signal exceeds the dynamic range when the level change of the video signal of the main line is large. It has been shown to limit the amplitude of the contour enhancement signal so that it will not be clipped and compromise the quality of the image.

Further, in Japanese Patent Application Laid-Open No. 3-277089, an image portion of a specific color such as the color of human skin is detected from the video signal of the main line, and the contour emphasis signal is suppressed in the image portion of the specific color. Has been shown to avoid the enhancement of the outline of a part such as a human face.

Further, in Japanese Patent Application Laid-Open No. 6-54232, a contour enhancement signal having a low clock rate is recorded before and after gamma knee correction for recording in a VTR which may be a video signal having a relatively low resolution. In addition to the video signal of, for the camera output that requires a high-resolution video signal,
It is shown that a contour enhancement signal having a high clock rate is added to a luminance signal obtained from a video signal after gamma knee correction by a matrix circuit and having a clock rate increased by an up converter.

With respect to the video camera, the CCD solid-state image pickup device for the R and B channels is horizontally shifted with respect to the CCD solid-state image pickup device for the G channel by 1/2 pixel pitch, which is a so-called spatial pixel shift method. It is known to obtain a video output with a resolution, and it is disclosed in Japanese Patent Laid-Open No.
No. 217326 and JP-A-6-153217.

Further, in JP-A-6-217326 and JP-A-6-153217, a CCD solid-state image pickup device is driven in order to perform shading correction by digital processing and prevent beat interference from occurring. It is shown that the clock rate to be used and the clock rate for digitizing the video signal from the CCD solid-state image sensor are made to match.

Further, in Japanese Patent Laid-Open No. 6-217326 and Japanese Patent Laid-Open No. 6-153217, when the above-mentioned spatial pixel shift method is used, digital signals are generated at a clock rate equal to the clock rate for driving the CCD solid-state image sensor. It is shown that the converted video signal is then processed at a clock rate twice the clock rate for driving the CCD solid-state image sensor.

[0013]

In the compression characteristic of the above-mentioned compression circuit, as shown in FIG. 20, the gain changes sharply with respect to the input level in the compression region. for that reason,
Even if the level of the video signal varies slightly among the pixels, the level of the contour enhancement signal changes significantly between the pixels, and the change appears as noise in the camera output.

Therefore, according to the present invention, when the compression means for suppressing the contour emphasis signal when the video signal is at a low level is provided, even if the level of the video signal varies from pixel to pixel, the level of the contour emphasis signal varies from pixel to pixel. It does not change significantly, and stable video output with reduced noise can be obtained.

[0015]

According to the present invention, there is provided a signal generating section for generating a contour emphasis signal from a main line video signal from an image pickup device, and a signal for adding the contour emphasis signal from the signal generating section to the main line video signal. A video camera provided with a compression means for suppressing the contour emphasis signal when the main line image signal is at a low level in the signal generation part, A minimum value extracting means for extracting the minimum value of the signal values corresponding to the front and rear pixels and outputting the minimum value to the compression means is provided.

Further, according to the present invention, there is provided a signal generating section for generating a contour emphasizing signal from the input video signal, and a signal adding section for adding the contour emphasizing signal from the signal generating section to the input video signal. In a contour enhancement device having a compression unit that suppresses the contour enhancement signal when the input video signal is at a low level in the generation unit, a signal corresponding to a plurality of pixels before and after the input video signal is provided in the preceding stage of the compression unit. A minimum value extracting means for extracting the minimum value of the values and outputting it to the compression means is provided.

In the video camera or the contour emphasizing device of the present invention configured as described above, even if the level of the video signal varies among pixels, the output of the minimum value extracting means which is the input of the compressing means varies between the pixels. There is no variation, and the level of the edge enhancement signal does not change significantly between pixels.

[0018]

1 and 2 show the overall structure of an example of a video camera of the present invention, which is shown in two separate figures for convenience.

The video camera of this example is composed of an image pickup section 10 and a camera body 20 as a whole, and the image pickup section 10 is constructed so as to be attachable to and detachable from the camera body 20. It is mechanically coupled to the camera body 20 and electrically connected to the camera body 20 through a cable such as a connector or a flexible cable.

The image pickup section 10 comprises an image pickup lens 11 and an image pickup block 12, and the image pickup block 12 divides the light incident through the image pickup lens 11 into R, G, and B color lights, and this color separation. CCD solid-state image pickup devices 14R, 14G, 14B on which red, green, and blue images of an object are formed by incidence of R, G, and B color lights separated by the prism 13, and the CCD solid-state image pickup device 1
CCD driver 15 for driving 4R, 14G, 14B,
An EEPROM (electrically erasable and rewritable ROM) 16 in which data described later is written is provided.

As the image pickup section 10, CCD solid-state image pickup elements 14R, 14G and 14B having 400,000 pixels and those having 500,000 pixels can be exchanged and used.
Reference numeral 16 denotes the CCD solid-state image sensor 14R, 14G, 1
4B pixel count data and other setup data are written in advance.

The camera body 20 is provided with a control unit 30 such as a microcomputer which will be described later.
By mounting and connecting the image pickup unit 10 to the camera body 20 and turning on the power of the camera body 20, the control unit 30 sets up the data such as the number of pixels of the CCD solid-state image pickup devices 14R, 14G, 14B from the EEPROM 16 by the control unit 30. The data is read out, and the imaging unit 10 and the camera body 20 are automatically set up based on the data.

The CCD solid-state image pickup devices 14R, 14
When G and 14B have 400,000 pixels, the CCD driver 15 uses the data from the control unit 30 to fs1 = 14.3.
When the CCD solid-state imaging devices 14R, 14G, 14B are driven at a clock rate of MHz (accurately 14.31818 MHz), and the CCD solid-state imaging devices 14R, 14G, 14B have 500,000 pixels, data from the control unit 30. Therefore, the CCD driver 15 has fs1 = 18 MHz (exactly 1
The CCD solid-state image pickup devices 14R, 14G, and 14B are driven at a clock rate of 8.00 MHz).

In the image pickup section 10, the B-channel CCD solid-state image pickup device 14B is vertically displaced from the R- and G-channel CCD solid-state image pickup devices 14R and 14G by one pixel pitch. The B channel image signal Ba obtained from the solid-state image sensor 14B is C
It is delayed by 1H (1 horizontal period) with respect to the R and G channel image signals Ra and Ga obtained from the CD solid-state image pickup devices 14R and 14G.

As described above, the camera body 20 includes the control unit 3
0 is provided, and the control unit 30 has a CPU 31 and a C
It includes a ROM 32 in which a control program to be executed by the PU 31 is written, a RAM 33 functioning as a work area of the CPU 31, and a RAM 34 in which the data read from the EEPROM 16 is written.

The control unit 30 supplies not only the CCD driver 15 of the image pickup unit 10 but also each unit of the camera body 20 with a clock and a control signal as described later. An operation panel 40 is connected to the control unit 30 for the user to make necessary settings and instructions.

CCD solid state image pickup device 14R of the image pickup unit 10,
Video signals Ra, Ga, Ba obtained from 14G, 14B
Is taken into the camera body 20 and processed in the analog process circuits 51R, 51G, and 51B in the camera body 20 such that reset noise is reduced by CDS (correlated double sampling).

Analog process circuits 51R, 51G, 5
The video signal from 1B is the A / D converter 52R, 52
G, 52B, CCD solid-state image sensor 14R, 14
Sampling is performed at a clock rate of fs1 = 14.3 MHz or fs1 = 18 MHz according to the number of pixels of G and 14B, and A / D conversion is performed.

The digitized R channel video signal R0 from the A / D converter 52R is delayed by the delay circuit 53.
The video signal R1 from the delay circuit 53R is delayed by 1H by the R, and further delayed by 1H by the delay circuit 54R.
Similarly, the digitized G channel video signal G0 from the A / D converter 52G is delayed by 1H by the delay circuit 53G, and the video signal G1 from the delay circuit 53G is delayed by 1H by the delay circuit 54G.

As described above, the CCD solid-state image pickup device 14B
The B-channel video signal Ba from A to D is delayed by 1H with respect to the R and G channel video signals Ra and Ga from the CCD solid-state image pickup devices 14R and 14G, so the digitized B channel from the A / D converter 52B. The video signal B1 of 1 is temporally matched with the digitalized R and G channel video signals R1 and G1 from the delay circuits 53R and 53G.

Then, the A / D converters 52R, 52
The video signals R0, G0, B1 from the G and 52B, the video signals R1 and G1 from the delay circuits 53R and 53G, and the video signals R2 and G2 from the delay circuits 54R and 54G are supplied to the image enhancer 100, and an image is obtained. In the enhancer 100, the detail signal BGD after gamma knee correction and the detail signal after gamma knee correction that are to be added to the video signal before gamma knee correction with the clock rate of fs2 = 2fs1 are used as the contour emphasis signals by the configuration and operation described later. The detail signal AGD to be added to the video signal,
And the detail signal VFD to be added to the video signal for the viewfinder is generated.

In this case, the video signals R1, G1 and B1 are supplied to the specific color detecting circuit 61, and the specific color detecting circuit 61 outputs a specific color detecting signal STA indicating an image portion of a specific color such as human skin color. The specific color detection signal STA obtained is supplied to the image enhancer 100. Specific color detection signal S
Although TA has a plurality of bits, its value is the minimum value in the image portion of the specific color, and its value is the maximum value in the image portion other than the specific color.

Further, an interface circuit 65 is connected to the control unit 30, various control port data are obtained from the interface circuit 65, and the control port data is supplied to each unit of the image enhancer 100 described later.

Fs1 = 14.3 MHz or fs1 = 1
The digitized video signals R1, G1, B1 having a clock rate of 8 MHz are converted into up converters 71R, 71.
G, 71B, the clock rate fs1 to fs
2 = 2fs1, that is, the CCD solid-state image sensor 14R,
Fs2 = 28.6 MH depending on the number of pixels of 14G and 14B
z or fs2 = 36 MHz.

The up converters 71R, 71G, 7
The video signal having the clock rate of fs2 from 1B is subjected to processing such as color correction in the masking circuit 72.

Then, adder circuits 73R, 73G, 73B
, The detail signal BGD from the image enhancer 100 is added to the R, G, and B channel video signals from the masking circuit 72, and the R, G after addition is added.
The G and B channel image signals are gamma and knee corrected in gamma and knee correcting circuits 74R, 74G and 74B.

Further, adder circuits 75R, 75G, 75B
In the gamma knee correction circuit 74R, 74G, 74
The detail signal AGD from the image enhancer 100 is added to the R, G, and B channel video signals after the gamma knee correction from B.

Then, in the matrix circuit 76, the luminance signal Y, the red color difference signal Cr, the blue color difference signal Cb, and the luminance for the viewfinder are calculated from the R, G, B channel video signals from the adding circuits 75R, 75G, 75B. Signal Y
vf is created.

The luminance signal Y, the red color difference signal Cr, and the blue color difference signal Cb from the matrix circuit 76 are supplied to the blanking circuit 7.
Blanking processing is performed by 7R, 77G, and 77B.

Although not shown in the figure, after the blanking processing, depending on the application, for example, for camera output, the luminance signal Y, the red color difference signal Cr, and the blue color difference signal Cb are used by an encoder to generate a composite video signal. Is formed, for example, for recording with a VTR, a luminance signal Y, a red color difference signal Cr,
The blue color difference signal Cb is converted to fs2 by the rate converter.
It is converted into a signal with a lower clock rate and then extracted.

Further, in the adder circuit 81, the detail signal VFD from the image enhancer 100 is added to the viewfinder luminance signal Yvf from the matrix circuit 76.
Is added, and the luminance signal after addition of the detail signal VFD from the adding circuit 81 is added to the blanking circuit 77.
Blanking processing is performed by R, 77G, and 77B, D / A conversion is performed by the D / A converter 83, and the result is supplied to the viewfinder 90.

FIG. 3 shows the image enhancer 100 described above.
The overall structure of one example is shown, and the image enhancer 100 of this example has a comb filter 110, a high-frequency aperture filter 120, a horizontal detail filter 130, a vertical detail filter 140, an edge detection unit 150, a gain setting unit 170, and a gain adjusting unit 180. , And a detail mixing section 190.

The comb filter 110 uses the digitized R channel image signals R0, R1, R2 and the G channel image signals G0, G1, G2 to perform a line-to-line operation to obtain a high-range aperture (horizontal detail in the highest region). Signal GR for horizontal detail, signal YCOMB for horizontal detail from high range to low range, and signal C for vertical detail
It outputs COMB.

The signal GR for the high frequency aperture is given by the equation (1).
As shown by, it is a signal obtained by multiplexing the G-channel comb filter output and the R-channel comb filter output with the clock of fs1.

GR = Kgr × MPX [{(G1 / 2) + (G0 + G2) / 4} × (1 + Kg). {(R1 / 2) + (R0 + R2) / 4} × (1 + Kr) + B1 × Kb] (1) Coefficients Kgr, Kg, Kr, Kb are selected according to the values of the above control port data, and Kb
Accordingly, the B channel video signal B1 can be added to the R channel comb filter output.

However, this is when the data APCOMB, which is one bit of the control port data, is 0, and when the data APCOMB is 1, the signal GR is the G and R channel video signals G1, which do not pass through the comb filter. It outputs a signal obtained by directly multiplexing R1.

The signal YCOMB for horizontal details from the high frequency band to the low frequency band is multiplexed with the comb filter output of the G channel and the comb filter output of the R channel with the clock of fs1 as represented by the equation (2). Signal, that is, the signal GR when the data APCOMB is 1
Is the same signal as.

YCOMB = Ky × MPX [{(G1 / 2) + (G0 + G2) / 4} × (1 + Kg). {(R1 / 2) + (R0 + R2) / 4} × (1 + Kr) + B1 × Kb] (2) The coefficient Ky is selected according to the value of the control port data.

The vertical detail signal CCOMB is generated by the line-to-line calculation of the G and R channel video signals as represented by the equation (3). However, the G channel signal is extracted through the low-pass filter having the (101) configuration.

CCOMB = {(G1 / 2) − (G0 + G2) / 4} × LPF (101) × (1/2) + (1 + Kc) × {(R1 / 2) − (R0 + R2) / 4} (3) ) The coefficient Kc is selected according to the value of the control port data.

The high-pass aperture filter 120 is a high-pass filter that generates a high-pass aperture signal APT, which is the highest horizontal detail signal, from the signal GR from the comb filter 110, according to the value of the control port data, equation (4). Four filter configurations shown by (7) to (7) can be selected.

APT = GR × (-1020-1) × (-1020-1) (4) APT = GR × (-1020-1) × (-12-1) (5) APT = GR × ( −1020-1) × (−47-4) (6) APT = GR × (−1020-1) (7) The filter characteristics in this case are shown by the solid line 1a in FIG.
It becomes like a long broken line 1b, a dashed-dotted line 1c, and a short broken line 1d.

The horizontal detail filter 130 uses the signal YCOMB from the comb filter 110 to output horizontal detail signals H, M in the high band, middle high band, middle band, and low band, respectively.
A high-pass filter that generates H, M, and L, and is expressed by equation (8)-
It has a filter configuration shown in (11).

H = YCOMB × (-1020-1) × (585) × (565) × (-12-1) (8) MH = YCOMB × (-1020-1) × (585) × (565) (9) M = YCOMB * (-1020-1) * (585) * (111) * (111) (10) L = YCOMB * (-1020-1) * (585) * (111) * ( 101) × (101) (11) The respective filter characteristics are as shown by the one-dot chain line 2a, short broken line 2b, long broken line 2c, and solid line 2d in FIG.

The vertical detail filter 140 generates a vertical detail signal VDTL and a cross color compression signal CCS from the signal CCOMB from the comb filter 110, and is one of the characteristic parts of the video camera of this example.

FIG. 6 shows this vertical detail filter 14
0, showing the signal CCO from the comb filter 110.
After MB has passed through horizontal low-pass filters 141 and 142 on the one hand, vertical detail compression circuit 143
By the gain adjusting circuits 144a, 14
The vertical detail signal VD is adjusted by gain adjustment by 4b.
A TL is generated.

As an example, the low-pass filter 141 has a (10201) configuration and the low-pass filter 142 has a (100020001) configuration. Therefore, the vertical detail signal VDTL is expressed by equation (12).

VDTL = CCOMB × [(10201) × (100020001) × Vs] × Va × Vb (12) where Vs is the compression ratio in the vertical detail compression circuit 143, and Va and Vb are the gain adjustment circuits 144a and 144b. Is the gain at.

The gain Va is 256 between 0 and 1 depending on the value of the control port data VG1.
The gain Vb is adjusted in stages, and is set to, for example, 0, 1/2, 1, or 2 depending on the value of the control port data VG2.

As described above, the clock rate fs1 is C
Fs1 = 14.3 MHz or fs1 = 18 MHz depending on the number of pixels of the CD solid-state imaging devices 14R, 14G, and 14B.
Can be switched to On the other hand, the low-pass filter 1
The properties of 41 and 142 are unchanged. Therefore, the composite characteristic of the low-pass filters 141 and 142 is CC
D solid-state image sensor 14R, 14G, 14B has 400,000 pixels, and even when fs1 = 14.3 MHz, CCD solid-state image sensor 14R, 14G, 14B has 500,000 pixel, fs1
= 18 MHz, the solid line 3 in FIG. 7 is obtained.

As shown in FIG. 8, the compression characteristic of the vertical detail compression circuit 143 has three points Va on the positive side of the input level.
p, Vbp, Vcp and the three negative points Vam, Vbm,
It is assumed that the slope of the output level with respect to the input level changes before and after Vcm.

Points Vap, Vbp, Vcp, Vam, Vb
m and Vcm are 25% in the range of 0% to 110% independently of each other depending on the value of the control port data.
Can be changed in 6 levels. In addition, depending on the value of another control port data, it is possible to select either 1/8 or 1 as the inclination in the region above the point Vcp and below the point Vcm.

In the vertical detail filter 140, the signal CCOMB from the comb filter 110, on the other hand, passes through the above low pass filter 141 and then through the horizontal high pass filters 145 and 146 to obtain the subcarriers in the vertical detail signal. The frequency component is extracted. As an example, the high-pass filter 145 is (-
10002000-1) and the high-pass filter 146 has a (-1020-1) configuration.

However, the selector 147 selects either the output signal of the high-pass filter 146 or the output signal of the high-pass filter 145 according to whether the value of the 1-bit control port data VL is 0 or 1. , The selected signal is gain adjustment circuits 148a, 1
The gain is adjusted by 48b to generate the cross color compressed signal CCS.

Then, the CCD solid state image pickup devices 14R, 14
G, 14B has 400,000 pixels and clock rate fs1 is 1
When it is set to 4.3 MHz, the selector 147 selects the output signal of the high-pass filter 146, and when the CCD solid-state image pickup elements 14R, 14G, and 14B have 500,000 pixels and the clock rate fs1 is set to 18 MHz, The selector 147 selects the output signal of the high pass filter 145. Therefore, the cross color compressed signal CCS is CC
When the D solid-state image sensor 14R, 14G, 14B has 400,000 pixels and fs1 = 14.3 MHz, the CCD solid-state image sensor 14R, 14G, 14B has 500,000 pixels and fs1 = 18 MHz. Sometimes the formula (1
4).

CCS = CCOMB × [(10201) × (−10002000-1) × (−1020-1)] × Ca × Cb (13) CCS = CCOMB × [(10201) × (−10002000-1)] × Ca × Cb (14) where Ca and Cb are gain adjustment circuits 148a and 148, respectively.
It is the gain at b.

The gain Ca is adjusted in 16 steps from 0 to 1 according to the value of the control port data CG1, and the gain Cb is adjusted to 0, 1/2, 1, according to the value of the control port data CG2. It is set to either of two.

As described above, the characteristics of the filter unit for extracting the sub-carrier frequency component in the vertical detail filter 140 are switched between fs1 = 14.3 MHz and fs1 = 18 MHz. The characteristics are as shown by the long dashed line 4a in FIG. 7 when fs1 = 14.3 MHz and as shown by the short dashed line 4b in FIG. 7 when fs1 = 18 MHz, regardless of the clock rate. Will be taken out at maximum.

Then, the image enhancer 100 shown in FIG.
, The cross color compression signal CCS from the vertical detail filter 140 is subtracted from the detail signal in the detail mixing section 190 to extract the sub-carrier frequency component in the vertical detail filter 140. Eventually acts as a filter that suppresses the subcarrier frequency component in the detail signal.

Therefore, regardless of the clock rate, the sub-carrier frequency component in the vertical detail signal is sufficiently suppressed, and cross color interference is sufficiently reduced in the video output of the video camera.

In the image enhancer 100 shown in FIG. 3, the edge detection section 150 identifies and detects the image edge portion of the video signal and the portion where the level changes in a burst shape, and at the image edge portion, the gain for the horizontal detail signal is detected. Is a characteristic part of the video camera of this example.

FIG. 9 shows an example of the edge detecting section 150. In the horizontal detail filter 130 shown in FIG. 3, the signal YCOMB from the comb filter 110 passes through the high-pass filter 131, so that the signal YC
The high frequency component FILH of OMB is taken out. That is,
The high frequency component FILH is extracted from the video signal VDS as shown in FIG.

Here, the partial BU in the video signal VDS is
The portion ED is an edge portion of the image, which is a portion where the level changes in a burst form (hereinafter referred to as a burst portion, but of course, different from the burst as the color synchronization signal). In FIG. 11 and FIGS. 14 to 16 which will be described later, the video signal VDS and the like are shown as analog waveforms, but of course, they are actually processed in a digitized state.

In the edge detection section 150 of FIG. 9, the absolute value extraction circuit 151 obtains the absolute value signal ABSO of the high frequency component FILH from the high pass filter 131, and the absolute value signal ABSO is obtained by the D flip-flop 152. Latched.

Further, by the maximum value selection circuit 153, D
The larger of the signal before being latched by the flip-flop 152 and the signal after being latched is selected. That is, as shown in FIG. 11, the absolute value signal ABSO from the absolute value extraction circuit 151 is obtained every clock.
The larger of the two adjacent ones is selected.

The absolute value signal MAXO selected by the maximum value selection circuit 153 is supplied to the edge detection circuit 160, and the edge detection circuit 160 causes the image edge portion ED of the video signal VDS to change to the burst portion as described later. B
It is detected separately from U.

Detection signal DE from the edge detection circuit 160
The time width of TO is expanded by the time width expansion circuit 154, and the detection signal EXPO having the expanded time width is
It is supplied to the compression circuit 155 and level-compressed with the characteristics described later. Then, the detection signal from the compression circuit 155 is inverted by the inversion circuit 156, and the inverted detection signal is taken out as the edge detection signal EDET of the output of the edge detection unit 150 through the low pass filter 157.

FIG. 10 shows a specific example of the edge detection circuit 160, in which the absolute value signal MAX from the maximum value selection circuit 153 is used.
O is sequentially latched by the D flip-flops 161a to 161h, and the maximum value selection circuit 162 selects the maximum value signal among the absolute value signals latched by the D flip-flops 161a to 161c and the maximum value. The selection circuit 163 causes the D flip-flop 161f.
The maximum value signal among the absolute value signals latched in ˜161 h is selected.

Further, the minimum value selection circuit 164 selects either the absolute value signal from the maximum value selection circuit 162 or the absolute value signal from the maximum value selection circuit 163, whichever is smaller, and the maximum value selection circuit 165, the absolute value signal from the maximum value selection circuit 162 and the maximum value selection circuit 1
The larger of the absolute value signals from 63 is selected.

Then, depending on whether the value of the selection signal SEL which is one of the control port data is 0 or 1, the selector 166 causes the absolute value signal from the minimum value selection circuit 164 and the maximum value selection circuit One of the absolute value signals from 165 is selected, and the selected absolute value signal is latched by the D flip-flop 167.

Further, the subtraction circuit 168 subtracts the absolute value signal from the D flip-flop 167 from the absolute value signal from the D flip-flop 161e, and the signal DETO after the subtraction detects the output of the edge detection circuit 160. It is taken out as a signal.

In the edge detection circuit 160, in the minimum value selection mode in which the absolute value signal of the minimum value from the minimum value selection circuit 164 is subtracted from the absolute value signal of the main line in the subtraction circuit 168, the burst part is detected as the detection signal DETO. In the maximum value selection mode in which absolute value signals are obtained at both ends of BU and the image edge portion ED and the subtraction circuit 168 subtracts the absolute value signal of the maximum value from the maximum value selection circuit 165 from the absolute value signal of the main line, As the detection signal DETO, an absolute value signal is obtained only in the image edge portion ED as shown in FIG.

Then, in the edge detecting section 150 of FIG.
As described above, the detection signal DETO from the edge detection circuit 160 is expanded in time width, level-compressed, inverted, and taken out as the edge detection signal EDET through the low-pass filter 157. Therefore, in the maximum value selection mode, the output signal EXPO of the time width expansion circuit 154 is output.
As the edge detection signal EDET, the one shown in FIG. 11 is obtained. However, in extending the time width, relative phase adjustment is performed with respect to the video signal VDS or the high frequency component FILH.

The compression characteristic of the compression circuit 155 is, for example, as shown by the solid line in FIG.
At a, Eb, and Ec, the slope of the output level with respect to the input level is assumed to change. Points Ea, Eb, Ec
Respectively, depending on the value of the control port data,
Can be changed independently of each other.

By the compression characteristic of the compression circuit 155 being set in this way, the gain characteristic of the edge detection signal EDET becomes as shown by the broken line in FIG. Therefore,
The video signal VDS has a burst portion B as shown in FIG.
When the detail signal DTLS including U and also having different level differences as the image edge portion is as shown in the same figure, the edge detection signal EDET is set to the above-mentioned maximum value selection. In the mode, as shown in the figure, in the burst portion BU of the video signal VDS, the gain is constant without lowering, and in the image edge portion, the gain becomes smaller as the level of the detail signal DSLS is higher.

Then, the image enhancer 100 shown in FIG.
In this case, as will be described later, this edge detection signal ED
The gain adjustment unit 180 multiplies the detail signal by ET. Therefore, the detail signal is compressed in the image edge portion with the characteristic shown in FIG.

When the detail signal is not controlled by the edge detection signal as described above, the detail signal becomes large when the level change of the video signal is large. Therefore, the video signal after addition of the detail signals exceeds the dynamic range. As a result, the image is clipped as shown in FIG. 15 (A), and an edge-like contour having a width of a frame is generated on the image, which may make the image unsightly.

On the other hand, when the detail signal is controlled by the edge detection signal EDET as described above, the detail signal is compressed as it is as shown in FIG. The video signal after addition of the signals does not clip beyond the dynamic range, and the image does not become unsightly.

Moreover, in this case, the image edge portion of the video signal is detected separately from the portion where the level of the video signal changes in a burst shape, and the detail signal is controlled substantially only at the image edge portion. It does not substantially affect the portion of the signal whose level changes in a burst.

FIG. 16 shows the case where the burst portion BU is included, and similarly, when the detail signal is not controlled by the edge detection signal EDET (FIG. 16A) and when it is performed (FIG. 16B). Are shown by comparison. However, FIG. 7B shows the above-mentioned minimum value selection mode,
This is a case where the detail signal is compressed at both ends of the burst portion BU.

In the image enhancer 100 of FIG. 3, the gain setting section 170 finally sets the gain for the detail signal and compresses the level of the detail signal when the level of the video signal of the main line is small.
This is one of the characteristic parts of the video camera of this example.

FIG. 17 shows an example of the gain setting section 170, in which the signal CCOMB from the comb filter 110 passes through the low-pass filter 171 to obtain the signal CC.
The low frequency component Si of OMB is taken out.

The low-pass component Si from the low-pass filter 171 is not directly input to the level depend circuit 178 for compressing the level of the detail signal when the level of the low-pass component Si is small, and the minimum value extraction circuit 17
3 is input and sequentially latched by the D flip-flops 174 and 175.

Further, in the minimum value extraction circuit 173, the minimum value selection circuit 176 outputs the signal before being latched by the D flip-flop 174 and the signal after being latched by the D flip-flops 174 and 175, respectively. The smallest of the three signals is selected.

That is, as shown in FIG. 18 as the input signal Si, the low frequency component Si changes to the signal Si every clock.
1, Si2, Si3, Si4, Si5, ..., From the minimum value extraction circuit 173, signals Si1, S
The signal Si2 having the minimum value of i2 and Si3 is the signal Sm1.
Of the signals Si2, Si3, Si4, the minimum value signal Si2 is extracted as the signal Sm2, and the minimum value signal Si3 of the signals Si3, Si4, Si5 is extracted as the signal Sm3. In this way, the minimum value signal Sm is extracted.

Therefore, in the output signal Sm of the minimum value extraction circuit 173, the signal Si4 existing in the input signal Si is included.
A signal with a large level such as does not exist. Then, the output signal Sm of the minimum value extraction circuit 173 is supplied to the level depend circuit 178.

As an example, the compression characteristic of the level depend circuit 178 shows the slope of the gain with respect to the input level before and after the three points Lb, Lc, Ld of the input level, as shown by the broken line portion 5a in the solid line 5 in FIG. K1, K2, K
3, K4 is supposed to change.

The input level La at which the gain is zero and the above-mentioned input levels Lb, Lc and Ld are set by the values of the control port data. Also,
Depending on the value of another control port data, the slope K1
~ K4 can be set. Further, an offset of ± 7% can be added to the input signal Sm to the level depend circuit 178 according to the value of another control port data, whereby the input level Le at which compression starts to be effective can be changed. Also, depending on the value of another control port data, the level depend circuit 17
It is possible to add an offset to the output signal So of No. 8 so that the gain of the detail signal in the portion where the level compression is effective can be adjusted.

As described above, the output signal Sm of the minimum value extraction circuit 173 does not include a signal having a large level, and the output signal Sm of the minimum value extraction circuit 173 is as shown in FIG.
Level dependent circuit 17 having compression characteristics such as 8
As a result of level compression by 8, the output signal So of the level depend circuit 178 comes to have a low level, as shown in FIG.

Therefore, this level depend circuit 1
As will be described later, when the output signal So of 78 is multiplied by the detail signal as a gain for the detail signal, the detail signal does not greatly fluctuate and a large noise is not generated in the camera output.

In the gain setting section 170 of FIG. 17,
The minimum value selection circuit 179 outputs the output signal So of the level depend circuit 178 and the edge detection unit 150 of FIG.
Of the edge detection signal EDET from the specific color detection circuit 61 and the specific color detection signal STA from the specific color detection circuit 61 shown in FIG. It is taken out as a signal GCTL.

However, as described above, the specific color detection signal STA is set to the minimum value of zero in the image portion of the specific color, so that the gain setting signal GCTL becomes zero in the image portion of the specific color, and the gain of the detail signal is increased. Is also zero,
No contour enhancement is done.

In the gain adjusting section 180 of FIG. 3, the high band aperture signal APT from the high band aperture filter 120 is displayed.
To the gain setting signal G from the gain setting unit 170 after the core processing having the characteristics shown in FIG.
The CTL is multiplied to generate the high-frequency aperture signal HAPT.

Further, the gain adjusting section 180 adds the horizontal detail signals H, MH, and L of the high frequency band, the middle high frequency band, and the low frequency band from the horizontal detail filter 130, and FIG. After the core processing having the characteristics shown in B) is performed, the gain setting signal GCTL from the gain setting unit 170 is multiplied to generate the low-frequency aperture signal LAPT.

Further, the gain adjusting section 180 adds the horizontal detail signals MH, M and L of the middle high band, the middle band and the low band from the horizontal detail filter 130 and the vertical detail signal VDTL from the vertical detail filter 140. , The core signal having the characteristics shown in FIG. 19C is applied to the signal after the addition, and then the gain setting signal GCTL from the gain setting unit 170 is multiplied to generate the detail signal DTL.

Further, in the gain adjusting unit 180, the horizontal detail signal MH in the middle-high range or the horizontal detail signal L in the low range from the horizontal detail filter 130 and the vertical detail signal VDT from the vertical detail filter 140.
19 is added to the signal after the addition of L and
After performing the core processing having the characteristics as shown in (A), the gain setting signal GCTL from the gain setting unit 170 is multiplied,
The detail signal VFD for the viewfinder is generated.

Further, the detail mixing section 190 of FIG.
Is the low-frequency aperture signal LA from the gain adjusting unit 180.
The detail signal to be added to the video signal before gamma knee correction as described above by adding PT and the detail signal DTL and subtracting the cross color compression signal CCS from the vertical detail filter 140 from the added signal. BGD is generated and is added to the addition circuits 73R and 73R of FIG.
Output to G, 73B.

The cross color compression signal CCS is passed through a low pass filter of (121) configuration in the detail mixing section 190, and then (3/4) fs1 and (5/4).
In order to suppress the bulge at fs1, fs1 = 14.3 MH
For z, a low-pass filter of (343) configuration
When s1 = 18 MHz, the low-pass filters of (232) configuration are passed.

Further, the detail mixing section 190 adds the high-frequency aperture signal HAPT, the low-frequency aperture signal LAPT and the detail signal DTL from the gain adjusting section 180 to obtain the video signal after gamma knee correction as described above. The detail signal AGD to be added is generated and output to the adder circuits 75R, 75G and 75B shown in FIG. Low-pass aperture signal LAPT and detail signal DTL
Adjusts its gain.

The minimum value extraction circuit 173 of the gain setting section 170 is not limited to the illustrated specific configurations, such as the minimum value extraction circuit 173 for extracting the minimum value of the signals of four adjacent pixels. .

[0111]

As described above, according to the present invention, when the compression means for suppressing the contour emphasis signal when the video signal is at the low level is provided, even if the level of the video signal varies among the pixels, the contour The level of the enhancement signal does not change significantly between pixels, and stable video output with reduced noise can be obtained.

[Brief description of drawings]

FIG. 1 is a block diagram showing a part of the overall configuration of an example of a video camera of the present invention.

FIG. 2 is a block diagram showing the rest of the overall configuration of an example of a video camera of the present invention.

FIG. 3 is a block diagram showing an example of an image enhancer in FIG.

FIG. 4 is a diagram showing a characteristic example of a high-frequency aperture filter in FIG.

5 is a diagram showing a characteristic example of the horizontal detail filter in FIG.

FIG. 6 is a block diagram showing an example of a vertical detail filter in FIG.

FIG. 7 is a diagram showing characteristics of the vertical detail filter of FIG.

FIG. 8 is a diagram showing a characteristic example of the vertical detail compression circuit in FIG. 6;

9 is a block diagram showing an example of an edge detection unit in FIG.

10 is a block diagram showing an example of an edge detection circuit in FIG.

FIG. 11 is a diagram illustrating waveforms used for edge detection.

FIG. 12 is a diagram showing characteristics used to describe edge detection.

FIG. 13 is a diagram showing characteristics used to describe edge detection.

FIG. 14 is a diagram showing waveforms used for explaining edge detection.

FIG. 15 is a diagram showing waveforms used for explaining edge detection.

FIG. 16 is a diagram showing waveforms used for explaining edge detection.

17 is a block diagram showing an example of a gain setting unit in FIG.

FIG. 18 is a diagram for explaining the minimum value extraction circuit in FIG. 17;

19 is a diagram for explaining a detail mixing unit in FIG. 3. FIG.

FIG. 20 is a diagram showing an example of compression characteristics of a compression circuit.

[Explanation of symbols]

Reference numeral 10 ... Imaging unit, 14R, 14G, 14B ... CCD solid-state imaging device, 20 ... Camera body, 100 ... Image enhancer, 140 ... High-pass detail filter, 141 ... Low-pass filter, 145, 146 ... High-pass filter, 14
7 ... Selector, 150 ... Edge detecting section, 160 ... Edge detecting circuit, 170 ... Gain setting section, 173 ... Minimum value extracting circuit

Claims (2)

[Claims] 1. A signal generating section for generating a contour emphasis signal from a main line video signal from an image pickup device, and a signal adding section for adding the contour emphasis signal from the signal generating section to the main line video signal. In a video camera provided with a compression unit that suppresses the contour emphasis signal when the main line image signal is at a low level in the generation unit, a signal value corresponding to a plurality of pixels before and after the main line image signal is provided in a stage before the compression unit. A video camera characterized in that a minimum value extracting means for extracting the minimum value of the above and outputting it to the compression means is provided. 2. A signal generation unit for generating a contour enhancement signal from an input video signal, and a signal addition unit for adding the contour enhancement signal from the signal generation unit to the input video signal. In a contour enhancement device provided with a compression unit that suppresses the contour enhancement signal when the input video signal is at a low level, in the preceding stage of the compression unit, among the signal values corresponding to a plurality of pixels before and after the input video signal, A contour emphasizing apparatus comprising a minimum value extracting means for extracting a minimum value and outputting the minimum value to the compression means.