JPS624907B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a noise removal circuit for video signals such as television signals, and the present invention relates to a noise removal circuit for video signals such as television signals, and can be used to remove noise from black and white video signals, color video signals, or NTSC television system.
The objective is to propose a system that can faithfully remove noise from video signals, regardless of whether the system is PAL television system or SECAM television system.

The applicant first proposed a prediction error detection circuit that detects a prediction error between a video signal at a certain point in time and another signal at a different point in time, and a prediction error signal supplied from this prediction error detection circuit. By calculating the signal at a certain point and the output of the non-linear circuit, when the level of the prediction error signal is less than a predetermined value, the prediction error signal is removed from the signal at a certain point to eliminate noise in the video signal. We proposed a noise removal circuit that eliminates noise. The noise removal circuit will be explained below.

In the following description, it is assumed that the video signal from which noise is to be removed is a sampled signal. First, referring to FIG. 1, this is the case of a noise canceller circuit. _t1 is an input terminal, to which a video signal (monochrome video signal or color video signal, television signal such as NTSC, PAL, SECAM, etc.) from which noise is to be removed is supplied. _t2 is the output terminal, from which a signal with noise removed is obtained.

The video signal from the input terminal _t1 is supplied to an emphasis filter 1 as a prediction error detection circuit, the output of which is a prediction error signal is supplied to a nonlinear circuit 2, and the output thereof is supplied to an attenuator 3 with an attenuation ratio a. be done. On the other hand, the video signal from input terminal _t1 has an attenuation ratio b
is supplied to the attenuator 4. and attenuators 3 and 4
are supplied to the combiner 5, the output of the attenuator 3 is subtracted from the output of the attenuator 4, and the output is obtained at the output terminal _t2 .

The transfer function of emphasis filter 1 is H(Z)
Then, this can be expressed as the following equation.

H(Z)=1−F(Z) Here, F(Z) represents the transfer function of the prediction filter. Further, Z can be expressed as follows, where ω is the signal angular frequency and T is the period of the sampling pulse.

Z=e ^j 〓 If ^{the T} video signal is X(Z), the predicted signal is F
(Z)・X(Z), therefore, the prediction error signal that is the output of emphasis filter 1 is
(Z)}・X(Z).

In this specification, a nonlinear circuit is generally defined as having an input level-output level characteristic that is linear in one range and non-linear (e.g. This is a circuit in which the output level is constant regardless of the input level value. In the case of FIG. 1, due to the nature of the noise canceller, the input level-output level characteristic of the nonlinear circuit 2 as shown in FIG. 2 is linear when the absolute value of the input level is below a predetermined level. A stripping circuit whose output level is zero when it is above a predetermined level, or a stripping circuit where the input level-output level characteristic is linear when the absolute value of the input level is below a predetermined level, as shown in FIG.
A limiter circuit is used in which the absolute value of the output level is a constant value when the output level is above a predetermined level.

In general, a continuous signal at a certain point in time can be predicted using statistical properties, but it is difficult to statistically predict signals of a small number of types. By the way, in the case of a video signal, if the noise is white noise, the linear combination signal of the video signal and its predicted signal is a signal with reduced noise. Therefore, in a video signal where there is a large prediction error, such as the outline of an image, the signal is assumed to have been predicted incorrectly and the video signal is mainly output, and noise is not removed and flat parts of the image are In the case of a statistically large number of signal types, noise in the video signal can be reduced or removed by mainly outputting a predicted signal or a linear combination signal of a video signal and its predicted signal.

Although the above-described noise removal circuit in FIG. 1 is a noise canceller circuit, a circuit in the case of a noise eliminator circuit will be explained with reference to FIG.
In addition, in FIG. 4, parts corresponding to those in FIG. 1 will be described with the same reference numerals. The video signal from input terminal t ₁ is supplied to prediction filter 6 . Then, the prediction signal output from the prediction filter 6 and the video signal from the input terminal _t1 are supplied to the synthesizer 7, where the prediction signal is subtracted from the video signal to obtain a prediction error signal. Therefore, this prediction filter 6 and synthesizer 7
The emphasis filter 1 is constructed in a manner similar to that shown in FIG. The output of the synthesizer 7, that is, the prediction error signal which is the output from the emphasis filter 1, is supplied to a nonlinear circuit 2', and its output is supplied to an attenuator 8 having an attenuation ratio c. On the other hand, the prediction signal output from the prediction filter 6 is supplied to an attenuator 9 having an attenuation ratio d. Then, the respective outputs of the attenuator 8 and the attenuator 9 are supplied to the combiner 10 and added, and the added output is obtained at the output terminal _t2 .

The characteristics of the prediction filter 6 and the emphasis filter 1 are the same as those described above, but because the nonlinear circuit 2' is a noise eliminator, the input level-output level characteristics as shown in FIG. 5 or 6 are different from the input level. A coring circuit is used in which the output level is zero when the absolute value of is below a predetermined level, and linear when it is above a predetermined level.

It is necessary to minimize the square root of the mean square of the above-mentioned prediction error of the video signal, and for this purpose, the transfer function H(Z) of the emphasis filter as the prediction error detection circuit is It is known that the amplitude characteristics may be selected so as to be opposite to the frequency spectrum of the video signal, which is the input signal.

Next, a noise removal circuit when the video signal from which noise is to be removed is an NTSC monochrome television signal will be described with reference to FIG. Since the frequency spectrum of such a black and white television signal has the characteristic that energy is concentrated in the low range and gradually decreases as it goes to the high range, the transfer function of the emphasis filter 1 as a prediction error detection circuit is H
(Z) is expressed as the following equation when using the correlation between adjacent pixels.

H(Z)=1-αZ ^-m However, α is a coefficient and 0<α≦1, and m is m>
It is 0. Then, a transfer ^function F( A prediction filter 6 with Z)=αZ ^-m is constructed, and a transfer function H(Z)=1−F(Z)=1−α is generated in the synthesizer 7 of this prediction filter 6.
An emphasis filter 1 of Z ^-m is constructed. 2
is a stripping circuit or a limiter circuit whose characteristics are shown in FIG. 3 or 4, and its transfer function is indicated by A. The attenuation ratio of attenuator 3 is a(=
1/2). Therefore, the transfer function W(Z) of the entire noise removal circuit is expressed as follows.

W(Z)=1-1/2(1-Z ^-m )・A FIG. 8 shows the arrangement of the pixel of interest on the horizontal scanning line of a field with a black and white television signal, where x is a, b are the pixels of the signal at points one sampling period before and after the signal of pixel x, and c, d are the pixels of the signal at points one horizontal period before and after the signal of pixel x. pixels e, f
denote pixels with a signal one sampling period before and after the signal of pixel c, and g and h denote pixels with a signal one sampling period before and after the signal of pixel d, respectively. Note that the sampling points on each horizontal scanning line are in phase.

Transfer function H of emphasis filter 1 when using correlation between adjacent pixels on the same scanning line
For example, m in (Z) is 1. In this case, the noise removal circuit obtains a prediction error signal between the signal of pixel x and the signal of pixel a in order to remove noise from the signal of pixel x in FIG. At this time, this prediction error signal is subtracted from the signal of pixel x. In this case, there is a possibility that the resolution of a pattern of fine horizontal stripes with little change in brightness in the horizontal direction may be reduced.

Furthermore, m of the transfer function H(Z) of emphasis filter 1 when using the correlation between vertically adjacent pixels on adjacent scanning lines of the same field is h
(=H/T, where H is the horizontal period). In this case, in order to remove noise in the signal of pixel x, the noise removal circuit obtains a prediction error signal between the signal of pixel x and the signal of pixel c, and when the level of this prediction error signal is below a predetermined value, the prediction error The signal will be subtracted from the signal of pixel x. In this case, the resolution of a pattern of fine vertical stripes with little variation in brightness in the vertical direction may be reduced.

The noise removal circuit in the case where the noise removal circuit shown in FIG. 7 is configured to detect prediction errors also for signals at sampling points later than the sampling point of interest will be explained with reference to FIG. 9. do.

A circuit whose transfer function is Z ^-m , that is, a delay circuit 12 having a delay amount m times the period T of the sampling pulse.
a, 12b and an attenuator 13a with a damping ratio α (=1),
13b and combiners 7a and 7b constitute emphasis filters 1a and 1b, respectively. 2a and 2b are stripping circuits or limiter circuits whose characteristics are shown in FIG. 3 or 4, and their transfer function is
Indicated by Aa and Ab. Attenuation ratio a of attenuators 3a and 3b
is 1/3. Note that the input signal to the emphasis filter 1a is from the input terminal _t1 , and the input signal to the emphasis filter 1b is from the delay circuit 1.
This is the output signal of 2a. Furthermore, in the synthesizer 5, each output signal of the attenuators 3a and 3b is subtracted from the output signal of the delay circuit 12a. Therefore, the transfer function W(Z) of the entire noise removal circuit is expressed as follows.

W(Z)=1-1/3 {(1-Z ^-m )・Ab+(1-Z ^m )・Aa} Emphasis filter 1 when using correlation between adjacent pixels on the same scanning line The transfer function H
m in (Z) is 1. In this case, in the noise removal circuit, in order to remove noise from the signal of pixel x,
, and the signals of pixels a and b, and when the level of the prediction error signal is less than a predetermined value, the prediction error signal is subtracted from the signal of pixel x. In this case as well, there is a possibility that the resolution of a pattern of fine horizontal stripes with little change in brightness in the horizontal direction may be reduced.

Furthermore, m of the transfer function H(Z) of emphasis filter 1 when using the correlation between vertically adjacent pixels on adjacent scanning lines of the same field is h
(=H/T, where H is the horizontal period). In this case, in order to eliminate noise in the signal of pixel x, the noise removal circuit obtains prediction error signals between the signal of pixel x and the signals of pixels c and d, and detects when the level of the prediction error signal is below a predetermined value. Then, the prediction error signal is subtracted from the signal of pixel x. In this case as well, there is a possibility that the resolution of a pattern of fine vertical stripes with little variation in brightness in the vertical direction will be reduced.

Furthermore, noise as shown in FIG. 9 is generated using an emphasis filter 1 that utilizes the correlation between adjacent pixels on the same scanning line and the correlation between vertically adjacent pixels on adjacent scanning lines of the same field. A removal circuit can also be used together. In that case, the transfer function W(Z) of the entire noise removal circuit is expressed as the following equation.

W (Z) = 1-a {(1-Z ^-1 )・Ab ₁ + (1-Z ¹ )・Aa ₁ + (1-Z ^-h )・Ab ₂ + (1-Z ^h )・Aa ₂ } In this case, a=1/5. Note that Aa ₁ , Ab ₁ , Aa ₂ , and Ab ₂ are transfer functions of the nonlinear circuit, respectively.
In this case, there is a possibility that the resolution of a fine checkered pattern with little variation in brightness in the horizontal and vertical directions may be reduced.

In this case, if a is fixed at 1/5, the noise in parts of the pattern with small brightness changes can be removed well, but in parts with large brightness changes, such as contour patterns, the noise may not be removed so well. If the number of nonlinear circuits whose outputs are linear is N, this drawback can be eliminated by varying a so that a becomes 1/N+1.

Next, a noise removal circuit when the video signal from which noise is to be removed is an NHSC color television signal will be described with reference to FIG. In the frequency spectrum of such a color television signal, the energy of the luminance signal is concentrated in the low range, and as it goes to the high range, the energy gradually decreases, and the energy is concentrated near the color subcarrier frequency in the higher range. Since the energy of the carrier color signal is concentrated, the transfer function H(Z) of the emphasis filter 1 as a prediction error detection circuit is It is expressed as the following equation.

H(Z)=(1-αZ ^-m )(1-βZ ^-n ) However, α and β are coefficients, and 0<α≦1, 0<
β≦1, and n is a value obtained by dividing the frequency of the sampling pulse by the color subcarrier frequency f _sc , and this n is selected to a value that satisfies the sampling theorem. For example, n=
3 or n=4. Further, m is m≧0. The emphasis filter 1 is a circuit whose transfer functions are Z ^-m , Z ^- 〓〓〓, Z ^- 〓〓〓, Z ^-m , respectively, that is, the sampling pulse (in this case, the sampling pulse is n times the color subcarrier frequency). pulses of frequency) m,
Delay circuits 12A to 12D having delay amounts of nm/2, nm/2, and m cycles, and damping ratios α (=1) and β, respectively.
(=1), αβ (=1) attenuators 13A, 13B,
13C, and synthesizers 7A, 7B, and 7C. 2 is a stripping circuit or a limiter circuit whose characteristics are shown in FIG. 2 or 3, and its transfer function is indicated by A. The attenuation ratio a of the attenuator 3 is
It is 1/2. Note that the delay circuits 12B and 12C can be combined into one delay circuit.

Thus, the transfer function W of the entire noise removal circuit
(Z) is as shown in the following equation.

W(Z)=1-a(1-αZ ^-m )(1-βZ ^-m )A FIG. 11 shows the arrangement of pixels of interest on the same horizontal scanning line of a color television signal. In this case, the sampling frequency is 3 times the color subcarrier frequency.
Double it. x′ is the pixel of the signal at a certain point, a′,
b' is the pixel of the signal one sampling period before and after the signal of pixel x', c', d' is the pixel of the signal three sampling periods before and after the signal of pixel x',
e' and f' respectively indicate pixels of signals before and after four sampling periods with respect to the signal of pixel x'.

If m=1 and n=3 in the transfer function W(Z) of the noise removal circuit in FIG. 10, this noise removal circuit will Obtain a prediction error signal between the signal of x′ and the signal of pixel a′, obtain a prediction error signal between the signal of pixel c′ and the signal of pixel e′, add these prediction error signals, and add this added signal. When the level of pixel x is below a predetermined value, this added signal is subtracted from the signal of pixel x. In this case, the resolution of a pattern of fine horizontal stripes with little change in brightness or chromaticity in the horizontal direction may be reduced.

Refer to FIG. 12 for a noise removal circuit in the case where the noise removal circuit shown in FIG. 10 is configured to detect prediction errors also for signals at sampling points earlier than the sampling point of interest. explain. Emphasis filters 1a and 1b are provided. The emphasis filter 1a is
A circuit whose transfer function is Z ^-1 , that is, a delay circuit 12 having a delay amount of one period of the sampling pulse (in this case, a pulse with a frequency three times the color subcarrier frequency).
Aa ~ 12Da, damping ratio α (= 1) and β (=
1), αβ (=1) attenuator 13Aa, 13Ba, 1
3Ca, and synthesizers 7Aa, 7Ba, and 7Ca. 2a is a stripping circuit or a limiter circuit whose characteristics are shown in FIG. 2 or 3;
Let the transfer function be _A1 . The attenuation ratio a of the attenuator 3a is 1/3.

The emphasis filter 1b includes a circuit whose transfer function is Z ^-1 , that is, delay circuits 12Ab to 12Db each having a delay amount corresponding to one period of the sampling pulse described above;
Damping ratio α (=1), β (=1), αβ (=1) respectively
attenuators 13Ab, 13Bb, 13Cb, and combiner 7
It is composed of Ab, 7Bb, and 7Cb. 2b is a stripping circuit or a limiter circuit whose characteristics are shown in FIG. 2 or 3, and its transfer function is
Let's say A ₂ . The attenuation ratio a of the attenuator 3b is 1/3.

Thus, the transfer function W of the entire noise removal circuit
(Z) is as shown in the following equation.

W(Z)=1-a(1-αZ- ¹ )(1-βZ ^-3 )A2 _- a(1-αZ)(1- ^βZ3 ) _A1In this noise removal circuit, the pixel x in Fig. 11 In order to remove noise in the signal of pixel x', a prediction error signal is obtained between the signal of pixel x' and the signal of pixel a', and a prediction error signal is obtained between the signal of pixel c' and the signal of pixel e'. Equal prediction error signals are added, and when the level of this added signal is below a predetermined value, this added signal is subtracted from the signal of pixel x', and a prediction error signal of the signal of pixel x' and the signal of pixel b' is obtained. , on the other hand, obtain a prediction error signal between the signal of pixel d′ and the signal of pixel f′, add these prediction error signals, and when the level of this added signal is below a predetermined value, add this added signal to the signal of pixel x′ It will be deducted more. In this case as well, there is a possibility that the resolution of fine horizontal striped patterns with little change in luminance or chromaticity in the horizontal direction may be reduced.

The noise removal circuits shown in FIGS. 10 and 12 above use an emphasis filter that utilizes the correlation between adjacent pixels on the same horizontal scanning line of a video signal. With reference to the figure, a noise canceling circuit for the above-mentioned NTSC color television signal, which is a noise canceller type using an emphasis filter that also utilizes the correlation between adjacent horizontal scanning lines of a video signal, will be explained. . In the noise removal circuit of FIG. 13, the color television signal is sampled with a sampling pulse having a frequency three times the color subcarrier frequency. In Fig. 13, 12Ea and 12Eb are delay circuits with a delay amount of T (sampling period);
Fa, 12Fb is a delay circuit with a delay amount of 2T, 12Ga,
12Gb is a delay circuit with a delay amount of T, 12Ha, 12
Hb is a delay circuit with a delay amount of H-5.5T (H is a horizontal period), and 12Ia and 12Ib are delay circuits with a delay amount of 3T.

Also, 7Da, 7Db, 7Ea, 7Eb, 7Fa, 7
Fb, 7Ga, 7Gb, 7Ha, 7Hb, 7Ia, 7Ib,
7J, 7K, 7L and 5 are combiners, respectively. 2
Cb, 2Bb, 2Ab, 2Aa, 2Ba, and 2Ca are stripping circuits or limiter circuits having transfer functions A ₁ to A ₆ and whose characteristics are shown in FIG. 2 or 3, respectively. 3b and 3c are attenuators with damping ratios a ₁ and a ₂ . Further, 20 is a delay circuit for compensating the amount of delay.

Thus, the transfer function W of the entire noise removal circuit
(Z) is as shown in the following equation.

W(Z)=1−a ₁ (1−Z ^−h−1 . ⁵ )A ₁ −a ₁ (1−Z ^−h+1 . ⁵ )A ₂ −a ₂ (1−Z ⁻¹ )(1 -Z ^-3 )A ₃ -a ₂ (1-Z) (1-Z ³ )A ₄ -a ₃ (1-Z ^h-1 . ⁵ )A ₅ -a ₃ (1-Z ^h+1 . ⁵ ) A ₆ However, h is H/T and a ₃ is a ₃ = a ₁ .

FIG. 14 shows the arrangement of pixels of interest on a horizontal scanning line of a field of a color television signal. In this case, the sampling frequency is three times the color subcarrier frequency, and since the carrier color signal has an opposite phase on the adjacent horizontal scanning line, the sampling position of the color video signal on the adjacent horizontal scanning line is also 180 degrees.
It's off by ゜.

x″ is the pixel of the signal at a certain point, a″, b″ are the pixels
The pixels of the signal one sampling period before and after the signal of pixel x'' are
The pixels of the signal before and after the sampling period are expressed as e″,
f'' is the pixel of the signal 4 sampling periods before and after the signal of pixel x'', and h'' is the pixel of the signal of pixel x''.
i″, j″ are the pixels of the signal 1.5 sampling periods before and after the point in time before the horizontal period, and i″, j″ are the pixels in the signal 1.5 sampling periods before and after the point in time after 1 horizontal period of the signal of pixel x″. Show each.

In the noise removal circuit shown in Fig. 13, the pixel shown in Fig. 14
To remove noise in the signal of x″, pixels x″ and
g″ signal set, pixel x″ and h″ signal set, pixel
Set of signals x″, a″, c″ and e″, pixels x″, b″,
Separate prediction error signals are created for the set of d″ and f″, the set of signals for pixels x″ and i″, and the set of signals for pixels x″ and j″, and the level of each prediction error signal is set to a predetermined level. When the level is below the predetermined value, the prediction error signal whose level is below the predetermined value is subtracted from the signal of pixel x''.

In this case, the resolution of fine striped patterns with little change in brightness or chromaticity in the horizontal or diagonal directions may be reduced.

In view of the above, the present invention provides an error detection circuit that detects an error between a signal at a certain point in time of a video signal and a signal at a point in time with a different time difference, and a nonlinear circuit to which the error signal from the error detection circuit is supplied. circuit, and by calculating the signal at a certain point in time and the output of the nonlinear circuit, when the level of the error signal is less than a predetermined value, the error signal is removed from the signal at a certain point in time to remove noise in the video signal. Regarding such a noise removal circuit, it is an object of the present invention to propose a noise removal circuit that can effectively remove noise with little reduction in resolution even in images with a fine striped pattern as described above.

The noise removal circuit according to the present invention includes an error detection circuit that detects an error between a signal at a certain point in time of a video signal and an average signal at a plurality of points in time with different time differences, and an error signal output from the error detection circuit. A nonlinear circuit is supplied, a variable attenuator is supplied with an error signal from the nonlinear circuit, and the attenuation ratio of the variable attenuator is connected to the nonlinear circuit and detects that the level of the error signal is below a predetermined value. The control circuit is configured to remove the error signal from the variable attenuator from the signal at a certain point in time, thereby removing noise from the video signal.

Next, an embodiment of the present invention will be described with reference to the drawings from FIG. 15 onwards, and the same reference numerals are given to the parts corresponding to those in FIGS. 1 to 14 described above.
First, a first embodiment of the present invention will be described with reference to FIG. This is a circuit that removes noise from NTSC black and white television signals. In the figure 12-
1 to 12-4 have transfer functions Z ^-h+1 , Z ^-1 ,
This is a delay circuit with Z ^-1 and Z ^-h+1 . 1-1~1
-4 is an emphasis filter as a prediction error detection circuit; 2-1 to 2-4 are stripping circuits or limiter circuits whose characteristics are shown in FIG. 2 or 3 as nonlinear circuits; 5, 7-1 ~7-11
is a combiner, and 13-1 to 13-4 are attenuators with a damping ratio of 1/2. 14 is a nonlinear circuit 2-1 to 2-4
This is a control circuit that detects the number of those outputting linear output signals, and when the number is N, controls the attenuation ratio a of the variable attenuator 15 to 1/N+1, which will be described later. 15 is a variable attenuator to which the outputs of the nonlinear circuits 2-1 to 2-4 added by the combiners 7-9 to 7-11 are supplied; is subtracted from the output of

The emphasis filter 1-1 is a delay circuit 12.
-1, 12-2, 12-3, combiner 7-2, 7-
5 and an attenuator 13-1, where the eighth
A prediction error signal between the average signal of the signals of pixels a and d in the figure and the signal of pixel x is obtained, and is supplied to the nonlinear circuit 2-1. equal pixels a, d and x
is located at the vertex of a triangle as a polygon on the screen.

The emphasis filter 1-2 is a delay circuit 12.
-1, 12-2, combiners 7-1, 7-6, and attenuator 13-2, where the average signal of the signals of pixels b and d in FIG. 8 and the signal of pixel x A prediction error signal is obtained, which is the nonlinear circuit 2-2.
is supplied to These equal pixels b, d, and x are located at the vertices of a triangle on the screen.

The emphasis filter 1-3 is a delay circuit 12.
-3, 12-4, combiners 7-3, 7-8, and attenuator 13-3, and here the average signal of the signals of pixels a and c in FIG. 8 and the signal of pixel x A prediction error signal is obtained, which is supplied to the nonlinear circuit 2-3. These equal pixels a, c, and x are located at the vertices of a triangle on the screen.

The emphasis filter 1-4 is a delay circuit 12.
-2, 12-3, 12-4, combiner 7-4, 7-
7 and an attenuator 13-4, where the eighth
A prediction error signal between the average signal of the signals of pixels b and c in the figure and the pixel signal is obtained, and is supplied to the nonlinear circuit 2-4. These equal pixels b, c, and x are located at the vertices of a triangle on the screen.

Thus, the transfer function W(Z) of the entire noise removal circuit shown in FIG. 15 becomes as shown in the following equation.

W (Z) = 1-a {(1-Z ^-1 +Z ^h /2)・A ₁ + (1-Z ¹ +Z ^h /2)・A ₂ + (1-Z ^-1 +Z ^{- h} /2)・
A ₃ +(1-Z ¹ +Z ^-h /2)·A ₃ } However, A ₁ to _{A 4} indicate the transfer functions of the nonlinear circuits 2-1 to 2-4.

Next, a second embodiment of the present invention will be described with reference to FIG. This is a circuit that removes noise from NTSC color television signals. Note that the sampling frequency is three times the color subcarrier frequency.
In the figure, 12-5 to 12-10 have transfer functions, respectively.
The delay circuits are Z ^-3 , Z ^{-h+4 .5} , Z ^-3 , Z ^-3 , Z ^{-h+4 .5} , and Z ^-3 . 1-5 to 1-10 are emphasis filters as prediction error detection circuits; 2-5 to 2;
-10 is a stripping circuit or limiter circuit whose characteristics are shown in FIG. 2 or 3 as a nonlinear circuit; 5, 7-12 to 7-28 are combiners; 13
-5 to 13-10 are attenuators with a damping ratio of 1/2. 14 detects the number of nonlinear circuits 2-5 to 2-10 that output linear output signals, and when the number is N, the attenuation ratio a of the variable attenuator 15 to be described later is set to 1/N+1. This is a control circuit that controls the 15 is a variable attenuator to which the outputs of the nonlinear circuits 2-5 to 2-10 added by the combiners 7-24 to 7-28 are supplied, and the output is sent to the delay circuit 12- by the combiner 5. is subtracted from the output of 7.

The emphasis filter 1-5 is a delay circuit 12.
-5, 12-6, 12-7, combiner 7-14, 7
-18 and an attenuator 13-5, where a prediction error signal between the average signal of the signals of pixels d'' and j'' in FIG. 14 and the signal of pixel x'' is obtained, and this is a nonlinear The equal pixels are supplied to the circuit 2-5.
d'', j'' and x'' are located at the vertices of a polygon, ie, a triangle, on the screen.

The emphasis filter 1-6 is a delay circuit 12.
-5, 12-6, 12-7, combiner 7-12, 7
-19 and an attenuator 13-6, here a prediction error signal between the average signal of the signals of pixels j'' and j'' in FIG. 14 and the signal of pixel x'' is obtained, and this is a nonlinear Pixel i″, which is supplied to circuit 2-6.
j″ and x″ are located at the vertices of the triangle on the screen.

The emphasis filter 1-7 is a delay circuit 12.
-6, 12-7, 12-8, combiner 7-13, 7
-20 and an attenuator 1-7, and here a prediction error signal between the average signal of the signals of pixels c'' and i'' in Fig. 14 and the signal of pixel x'' is obtained, and this is a nonlinear These pixels are supplied to the circuit 2-7.
i″ and x″ are located at the vertices of the triangle on the screen.

The emphasis filter 1-8 is a delay circuit 12.
-7, 12-8, 12-9, combiner 7-15, 7
-21 and an attenuator 13-8, where a prediction error signal between the average signal of the signals of pixels d'' and h'' in FIG. 14 and the signal of pixel x'' is obtained,
This is supplied to the nonlinear circuit 2-8. equal pixels
d'', h'' and x'' are located at the vertices of the triangle on the screen.

The emphasis filter 1-9 is a delay circuit 12.
-8, 12-9, 12-10, combiner 7-16,
7-22 and an attenuator 13-9, where a prediction error signal between the average signal of the signals of pixels g'' and h'' in FIG. 14 and the signal of pixel x'' is obtained, and These pixels g'', h'' and x'' are located at the vertices of a triangle on the screen.

Emphasis filter 1-10 is delay circuit 1
2-8, 12-9, 12-10, combiner 7-1
7, 7-23 and an attenuator 13-10, where a prediction error signal between the average signal of the signals of pixels c'' and g'' in FIG. 14 and the signal of pixel x'' is obtained. , are supplied to the nonlinear circuit 12-10.The equal pixels c'', g'' and x'' are located at the vertices of a triangle on the screen.

Thus, the transfer function W(Z) of the entire noise removal circuit shown in FIG. 16 becomes as shown in the following equation.

W(Z)=1-a{(1-Z ³ +Z ^{h+1.5 /} 2)・A ₁ +(1-Z ^{h- 1.5} +Z ^{h+1.5 /} 2)・A ₂ +(1-
Z ⁻³ +Z ^h−1 . ⁵ /2)・A ₃ + (1-Z ³ +Z ^{-h+1.5 /} 2)・A ₄ + (1-Z ^{-h+ 1.5} +Z ^{-h-1.5 /} 2)・A ₅ + (1- Z ⁻³ +
Z ^-h-1 . ⁵ /2)・A ₆ } However, A ₁ to A ₆ are nonlinear circuits 2-5 to 2-10
is the transfer function of

Next, a third embodiment of the present invention will be described with reference to FIG. This is a circuit that removes noise from NTSC black and white television signals. In the figure, 1
2-11 to 12-18 each have a transfer function of Z ^-1 ,
The delay circuits are Z ^-1 , Z ^-h+2 , Z ^-1 , Z ^-1 , Z ^-h+2 , Z ^-1 , and Z ^-1 . 1-11 to 1-14 are emphasis filters as prediction error detection circuits; 2-1
1 to 2-14 are stripping circuits or limit circuits whose characteristics are shown in FIG. 2 or 3 as nonlinear circuits; 5, 7-29 to 7-39 are combiners;
13-11 to 13-14 are attenuators with a damping ratio of 1/2. 14 detects the number of nonlinear circuits 2-11 to 2-14 that output linear output signals, and when the number is N, a variable attenuator 1, which will be described later;
This is a control circuit that controls the damping ratio a of No. 5 to 1/N+1. 15 is a variable attenuator to which the outputs of the nonlinear circuits 2-11 to 2-14 added by the combiners 7-37 to 7-39 are supplied; is subtracted from the output of

Emphasis filter 1-11 is delay circuit 1
2-11 to 12-18, combiner 7-32, 7-3
6 and an attenuator 13-1, where the eighth
A prediction error signal between the average signal of the signals of pixels e and h in the figure and the signal of pixel x is obtained, and is supplied to the nonlinear circuit 2-11. These equal pixels e, h, and x are located at three points on a straight line on the screen.

Emphasis filter 1-13 is delay circuit 1
2-12 to 12-17, combiner 7-31, 7-3
5 and an attenuator 13-12, where a prediction error signal between the average signal of the signals of pixels c and d in FIG. 8 and the signal of pixel x is obtained, and this is the nonlinear circuit 2-12. supplied to These equal pixels c, d, and x are located at three points on a straight line on the screen.

Emphasis filter 1-13 is delay circuit 1
2-13 to 12-16, combiner 7-30, 7-3
4 and an attenuator 13-13, where a prediction error signal between the average signal of the signals of pixels f and g in FIG. 8 and the signal of pixel x is obtained, and this is the nonlinear circuit 2-13. supplied to These equal pixels f, g, and x are located at three points on a straight line on the screen.

Emphasis filter 1-14 is delay circuit 1
2-14, 12-15, combiner 7-29, 7-3
3 and an attenuator 13-14, where a prediction error signal between the average signal of the signals of pixels a and b in FIG. 8 and the signal of pixel x is obtained, and this is the nonlinear circuit 2-14. supplied to These equal pixels a, b, and x are located at three points on a straight line on the screen.

Thus, the transfer function W(Z) of the entire noise removal circuit shown in FIG. 17 is expressed as the following equation.

W (Z) = 1-a {(1-Z ^-h-1 +Z ^h+1 /2)・A ₁ + (1-Z ^-h +Z ^h /2)・A ₂ + (1-Z ^-h+1 +Z ^{h- 1} /2)・A ₃ + (1-Z ⁻¹ +Z ¹ /2)・A ₄ } However, A ₁ to _{A 4} are nonlinear circuits 2-11 to 2-1
4 transfer function.

Next, a fourth embodiment of the present invention will be described with reference to FIG. This is a circuit that removes noise from NTSC color television signals. Note that the sampling frequency is three times the color subcarrier frequency. In the figure, 12-19 to 19-24 are delays whose transfer functions are Z ^-3 , Z ^{-h+4 .5} , Z ^-3 , Z ^-3 , Z ^{-h+4 .5} , and Z ^-3 , respectively. It is a circuit.

In the figure, 1-15 to 1-17 are emphasis filters as prediction error detection circuits, and 2-15 to 2
-17 is a stripping circuit or limiter circuit whose characteristics are shown in FIG. 2 or 3 as a nonlinear circuit; 5, 7-40, and 7-47 are combiners;
-15 to 13-17 are attenuators with a damping ratio of 1/2. 14 detects the number of nonlinear circuits 2-15 to 2-17 that output linear output signals, and when the number is N, the attenuation ratio a of the variable attenuator 15 to be described later is set to 1/N+1. This is a control circuit that controls the 15 is a variable attenuator to which the outputs of the nonlinear circuits 2-15 to 2-17 added by the combiners 7-46 and 7-47 are supplied; is subtracted from the output of 21.

Emphasis filter 1-15 is delay circuit 1
2-19 to 12-24, combiner 7-42, 7-4
5 and attenuators 13-15, where a prediction error signal between the average signal of the signals of pixels g'' and j'' in FIG. 8 and the signal of pixel x'' is obtained, and this is a nonlinear circuit. 2-15, the equal pixel g″,
j″ and x″ are located at three points on a straight line on the screen.

Emphasis filter 1-16 is delay circuit 1
2-20 to 12-23, combiner 7-41, 7-4
4 and attenuators 13-16, and here a prediction error signal between the average signal of the signals of pixels h'' and i'' in FIG. 8 and the signal of pixel x'' is obtained, and this is a nonlinear circuit. 2-16, the equal pixels h″,
i″ and x″ are located at three points on a straight line on the screen.

Emphasis filter 1-17 is delay circuit 1
2-21, 12-22, combiner 7-40, 7-4
3 and attenuators 13-17, and here a prediction error signal between the average signal of the signals of pixels c'' and d'' in FIG. 8 and the signal of pixel x'' is obtained, and this is a nonlinear circuit. 2-17, the same pixels c″,
d″ and x″ are located at three points on a straight line on the screen.

Thus, the transfer function W(Z) of the entire noise removal circuit
is as shown in the following equation.

W(Z)=1-a{(1-Z- ^{h-1.5 +} Z ^{h+1.5 /} 2)・A ₁ +(1-Z ^{-h+ 1.5} +Z ^{h- 1.5} /2)
・A ₂ + (1-Z ⁻³ +Z ³ /2)・A ₃ } However, A ₁ to A ₃ are nonlinear circuits 2-15 to 2-1
7 transfer function.

Although various other embodiments are possible without being limited to the above-mentioned embodiments, other embodiments are shown only in terms of transmission relationships. That is
This is the case with an NTSC black and white television signal.

W(Z)=1-a{(1-Z ^-1 +Z ^-h-1 +Z ^-h /3)・A ₁ +(1-Z ^-h +Z ^-h+1 +Z ^1/3 )・A ₂ +(1 −Z ¹ +Z ^h+1 +Z ^h /3)・A ₃ +(1−Z ^h +Z ^h−1 +Z ⁻¹ /3)・A ₄ } These are the pixels a, e, c; c, f, b;
b, h, d; The signals of d, g, and a are averaged, respectively, and a prediction error between the average signal and the signal of pixel x is created. A ₁ to _{A 4} are the transfer functions of the nonlinear circuit,
a is 1/N+1 as described above. Similarly, various transmission relationships of the noise removal circuit are possible in the case of NTSC color television signals.

According to the present invention described above, there is provided an error detection circuit that detects an error between a video signal at a certain point in time and a signal at a point in time with a different time difference, and a nonlinear circuit to which the error signal from the error detection circuit is supplied. circuit, and by calculating the signal at a certain point in time and the output of the nonlinear circuit, when the level of the error signal is less than a predetermined value, the error signal is removed from the signal at a certain point in time, thereby removing noise in the video signal. With the noise removal circuit configured as described above, it is possible to obtain an image that can effectively remove noise with little reduction in resolution even in images with a fine striped pattern.

[Brief explanation of the drawing]

Figure 1 is a block diagram showing an example of the previously proposed noise removal circuit, Figures 2 and 3 are characteristic curve diagrams, Figure 4 is a block diagram showing other examples, and Figures 5 and 3 are characteristic curve diagrams. 6 is a characteristic curve diagram, FIG. 7 is a block diagram showing other examples, FIG. 8 is an explanatory diagram, FIGS. 9 and 10 are block diagrams each showing other examples, and FIG. 11 is an explanatory diagram. 12 and 13 are block diagrams of other examples, FIG. 14 is an explanatory diagram, and FIGS. 15, 16, 17, and 18 are block diagrams of other examples, respectively. FIG. 1-1 to 1-17 are emphasis filters as prediction error detection circuits, and 2-1 to 2-17 are nonlinear circuits.

Claims (1)

[Claims] 1. An error detection circuit that detects an error between a video signal at a certain point in time and an average signal at multiple points in time with different time differences, and a nonlinear circuit to which the error signal from the error detection circuit is supplied. a variable attenuator to which an error signal from the nonlinear circuit is supplied; and a control connected to the nonlinear circuit that detects that the level of the error signal is below a predetermined value and controls the attenuation ratio of the variable attenuator. has a circuit,
A noise removal circuit characterized in that noise in the video signal is removed by removing an error signal from the variable attenuator from the signal at a certain point in time.