JPS6037885A - Noise reducing circuit - Google Patents

Abstract

PURPOSE:To prevent blur of the edge of steplike signal part of a video signal by using delaying means and logical means, detecting a pulselike signal from an input signal by utilizing correlation between adjoining picture elements, and removing a projecting part as noise. CONSTITUTION:Delay lines 18, 19 with a delay time T1 and delay lines 20, 21 with a delay time T2(T1<T2) are connected to an input terminal 17. The output terminals of delay lines are connected to the input terminals of switching circuits 22, 23. The output terminals of an AGC circuit 24 for detecting input electric field intensity is connected to the switching circuit 22, 23, and the switch is changed from a terminal (a) to a terminal (b) according to the input electric field intensity. Accordingly, in case of strong electric field having small noise component and ordinary electric field, noise is removed by delay time T1, and the noise is reduced without deteriorating picture. In case of weak electric filed having much noise component, the noise is removed by the delay time T2 and switching is made automatically to have larger noise reducing effect.

Description

[Detailed Description of the Invention] [Technical Field of the Invention] The present invention relates to a noise reduction circuit, for example, in a video processing circuit of a color television receiver, which minimizes image deterioration and reduces pulse-like noise components mixed into video signals. The present invention relates to a noise reduction circuit configured to detect and remove noise.

[Technical background of the invention and its problems]

Generally 1. In a video processing circuit in a color television receiver, noise components mixed into a video signal greatly impair reproduced images, resulting in deterioration in image quality and information loss.

For this reason, attempts have been made to reproduce high-quality images by removing noise components from video signals mixed with noise. For example, the following methods can be cited.

(2) Improved by changing the transmission band of the video processing circuit according to the electric field strength and lowering the band with many noise components.

■Since the power spectrum of the video signal of television broadcasting has peaks in each horizontal period, a filter that passes only the band in which the video signal power spectrum is present, and an open-shape filter can be used to eliminate noise. components can be reduced.

(2) A process of replacing the gray level of a pixel with the average value of gray levels in the vicinity of that pixel, that is, a smoothing process is performed.

In addition to the above methods, there are various methods that can be used to repair images containing noise by performing digital processing on images using a computer. However, digital processing using a computer tends to require a large system, and In terms of time, it is generally the mounting part that is applied to home television receivers.

The above-mentioned noise reduction methods (1) and (2) also have the following problems.

Regarding method This results in blurring of the edges and deterioration of the image.

The noise reduction using the comb filter described in item (2) above is basically the same principle as the smoothing process in item (2) above, and both have similar drawbacks.

Here, noise reduction by the smoothing process described in (2) above will be briefly explained.

As shown in FIG. 1, if we take the case of 9-point smoothing in the vicinity of a pixel as an example, 9 pixels y (i, j) (where 1≦i≦1.-1≦j≦1) sample mean y (i
, Dke, and this average y (1, j) becomes the noise-removed image. Now, assuming that the true image X (iIJ) can be approximated by a plane at the above nine points, x (i, j) = ai + bj + x (b, o) where -1≦i≦1. −1≦j≦1)・・・・・・・・・・・・
It can be written as ・・・・・・・・・・・・・・・(2). Here, a and b are gradients in each direction. Assuming that the shunted noise component is n(i, j) and the noise is additive, y (i r D −x (i, j) + n'(i
, D −−−−−−−−<3). formula(
Substituting equations (2) and (3) into 1) yields. Now, assuming that the noise n (1, j) is uncorrelated and has a uniform variance width, the new noise will have the variance reduced to 1. E is the expected value.

Therefore, noise is reduced.

However, in the above-mentioned method using the smoothing process, a problem arises in that the image X (ITJ) as assumed above differs from the true value in a portion that cannot be approximated by a plane. For example, since the image cannot be approximated by a plane image, the edges tend to be blurred by the smoothing process.

Further, the noise reduction by the comb filter described in (2) above is a special case of the smoothing process described in (2) above, and can be considered as a case where the number of samples of a pixel is two. Therefore, the above-mentioned method (2) also has the problem of fraying of the edges.

[Purpose of the invention]

In view of the above-mentioned points, the present invention provides a noise reduction circuit that can prevent blurring at the edges of an image, minimize image deterioration, and reduce noise components in image information processing. It is intended to.

[Summary of the invention]

The noise reduction circuit of the present invention is configured to detect only a pulse-like signal from an input signal by utilizing the correlation between instantaneously connected pixels, and to determine and remove the protruding portion as noise. Two delay means for sampling input signals at different delay time intervals, a signal supply means for outputting these delay means by switching, and this ("
k % k+D This will be discussed when the signal level detection means for switching and controlling the means according to the input electric field level and the level of the AC component of the video signal and the signal sampled by the delay means are pulsed signals. It consists of logical means for removal 1 and output.

[Embodiments of the invention]

Embodiments of the present invention will be described below based on the drawings.

FIG. 2 is a waveform diagram showing an example of a video signal waveform, FIG. 3 is a waveform diagram showing an example of a step signal waveform, and FIG. 4 is an operation explanatory diagram for realizing the noise reduction circuit according to the present invention.

In a television transmitting/receiving system, by scanning the image on the image sending side, the gray dust of the pixels is converted into a temporally continuous electrical signal, and on the receiver side, it is synchronized with the resolution scanning of the image on the sending side. It is reconstructed into a two-dimensional image by assembling and scanning.

An example of a video signal converted into an electrical signal is shown in FIG. 2, and the video signal is expressed as a function of time t. In other words, the video signal X is written as x=i(t).

Now, let X at time t be x(i), and for a certain time τ, X at time L−τ and t+τ be x (1−
1), x (1+1). Here, the following function p is defined.

p (i −1) , l) (i) , p (i+
1) Let P be the set of P = (p (i-1), p (i), p
(i+t))...(6). In this way, the pattern of the video signal shown in FIG.
) can be represented by P. For example, the step signal shown in FIG. 3 is P=(0, 1, 11...)
...(7). The relationship between P and the turn of the video signal is shown in Thorn 1. (Margin below) Table 1 By the way, in general, noise components mixed into video signals are often in the form of pulses. To reduce noise components without deteriorating the image by removing the pulse signals shown in Py and Ps in Table 1 by adjusting the video signal P by an appropriate time τ. Can be done.

For example, a signal in which the noise signal shown in FIG. 4(b) is added to the video signal shown in FIG. 4(a) is shown in FIG. 4(e).
If the signal waveform is as shown in
When it is passed through a filter that blocks 1 pH, the output has a waveform as shown in FIG. 4(d), and the noise is reduced. At this time, the rising edge and falling edge of the step signal are
3~P6ilt is not blocked, so the edges will not fray. FIG. 4(6) shows the output waveform when the video signal of FIG. 4(c) is passed through a filter that blocks high-frequency components as described in the conventional method (2). f) Vi This is an output waveform when the signal in FIG. 4(e) is subjected to smoothing processing (in the figure, averaging of 3 samples) as described in the conventional method (2). In both cases of FIGS. 4(e) and 4(f), edge blurring occurs in the step signal portion, resulting in image deterioration.

Next, a specific circuit configuration of the noise reduction circuit that blocks P7 and P8 will be described.

To block P7 and P8, p (
You can replace i) with 0 and make P = (0, 0, 0) = P1, and for P8, if you replace p(i) with 1 and make P = (1, 1, 1) = Pz. good.

In this method, p'(i-1) IP(1)
.

The power P7 + PI to take the majority vote of p (ii1)
A noise filter circuit is constructed to prevent this. In other words, if p(t) after passing through the filter circuit is Q(i), 'li> ip (t -1)
xr(iiz) ii(p (i −1) xpc1+
J)・・・・・・・・・・・・・・・・・・・・・(8
) can express the characteristics of the filter circuit.

However, X and + in equation (8) represent logical product and logical sum. Further, the filter characteristic shown in equation (8) is equivalent to the characteristic shown in the following equation.

Q (1) = (p (i-1) ten tones)) × (recompression p (
ii1) ) x (p (t-υtenp (ii1) )・
・・・・・・・・・・・・・・・・・・・・・(9) Therefore, using equations (8) and (9), a noise filter circuit as shown in FIGS. 5 and 6 is created. can be configured.

FIG. 5 shows a filter circuit configuration corresponding to equation (8),
FIG. 6 shows a filter circuit configuration corresponding to equation (9).

In FIG. 5, the filter circuit connects a delay line 2 having a delay time τ to an input terminal 1 to which an input signal Vin is applied, and connects a delay line 3 having a delay time τ in series with this input terminal 2. , the AND circuit 4 is connected between the input and output terminals of the delay line 2, and the delay line 3
An AND circuit 5 is connected between the input and output terminals of the delay line 2, and an AND circuit 6 is connected between the input terminal of the delay line 2 and the output terminal of the delay line 3. And AND circuit 4゜5.
A logical sum circuit 7.8 is arranged to calculate the logical sum of the respective outputs of 6. That is, the outputs of the AND circuits 4 and 5 are input to the OR circuit 7, and the OR output and the output of the AND circuit 6 are input to the OR circuit 8. The output vOut of the OR circuit 8 is taken out from the output terminal 9.

Also, in FIG. 6, OR circuits io, ii, 12 are replaced in place of the AND circuits 4, 5゜6 shown in FIG.
The circuit configuration is such that AND circuits 13 and 14 are substituted for OR circuits 7 and 8. Number 15 in Figures 5 and 6
and a logic circuit section for removing Hals-like components.

In such a configuration, the circuits shown in 1 and 5 are
I Kaju person No. 1 plane --I V Doro 11 life 1
4 ii 7, i gigantic nrl -1), the signal p (i) output from delay line 2 and the signal p (1+1) output from delay line 3 are P7=i0,1+O1 shown in Table 1. In this case, the output of the AND circuit 4 is "0#" and the output of the AND circuit 5 is "0", so
The output of the OR circuit 7 is "0". Also, since the output of the AND circuit 6 is "0", the output of the OR circuit 8 is "0".
0", and the signal P(i) obtained at the output terminal 9 after passing through the filter circuit becomes "0". Therefore, p7f'
After passing through the i circuit, (0, 0, 0)=Pt, and the prominent portion of the pulsed signal is removed as a noise component. Also,
In Figure 5, p (i -1), p (i),
p (iit) is shown in Table 1 Pa=(1,0,1)
When , the output of the AND circuit 4 is 0'', and the output of the AND circuit 5 is
Since the output of is 0", the output of the OR circuit 7 is "θ"
becomes.

Further, since the output of the AND circuit 6 is 1'', the output of the OR circuit 8 is ``1'', and after passing through the filter circuit, the signal p(i) #-1' obtained at the output terminal 9Vc becomes 1''. Therefore, P8 is (1, 1, 1) = P after passing through the circuit
The protruding portion of the chivalry signal such as x is removed as a noise component. In addition, in FIG. 5, regarding the rise and fall points of the step signal P3 to P6, even after passing through the circuit,
Since (i) becomes a signal of the same level and outputs P3 to P6, the edge portion does not become frayed.

For the circuit shown in FIG. 6, p (i −1),
p(i).

P7 = (0, 1, 0) where p (10i) is expressed by clothes 1
When , the output of the OR circuit 10 is "1" and the output of the OR circuit 11 is @1#, so the output of the AND circuit 13 is ``1''. The output of the logical sum circuit 12 is
0', the output of the AND circuit 14 becomes "1", and the signal p(i)rri obtained at the VC output/terminal 9 after passing through the filter circuit becomes "0". Therefore, P7
After passing through the circuit, (0, 0, 0)=Px, and the protruding portion of the pulsed signal is removed as a noise component. Also, the 6th
In the figure, p (1-1), p (i), p (i+
When i) is Pg = (1, 0, 1) shown in 1, the output of the OR circuit 10 is @1# and the output Vi'''' of the OR circuit 11 is 12, so the AND circuit 13 output V
i″1″.

Also, since the output of the OR circuit 12 is "1", the output Vi"12 of the AND circuit 14 becomes "1", and the signal p(i) obtained at the output terminal 9 after passing through the filter circuit becomes "1". Therefore, , ps is (1, 1, 1) = P after passing through the circuit.The protruding part of the chivalry signal is removed as a noise component.In addition, in Fig. 6, the P3 of the rising and falling edges of the step-like signal is Regarding ~Pa, even after passing through the circuit, p (i) becomes a signal of the same level and P3 to P6 are outputted, so that the edge portion does not fray.

According to such a circuit configuration, it is possible to reduce noise components without causing image deterioration such as blurring of edge portions of a step-like signal.

By the way, as mentioned above, the method of blocking the pulse-like signals indicating Py and Ps at an appropriate time τ of the video signal (in other words, fL) is a method of blocking the pulse-like signals indicating Py and Ps at an appropriate time τ of the video signal (in other words, fL is a method that uses the correlation with adjacent pixels on the television screen to block the prominent pulse signals. Although the noise removal method (method in which only the signal in the form of a signal is determined to be noise and removed) has a high noise removal effect, improvements in the following points are particularly desired.

■In the circuits of Figures 5 and 6, the sampling time τ is 1/2τ()IZ) as shown in Figure 7(a).
Since higher frequency noise cannot be removed, this is improved.

(2) As shown in FIGS. 7(b) and 7(e), it is impossible to remove noise at a frequency lower than 1/8 τ (Hz) with respect to the sampling time τ, so this is improved.

■Since it does not only judge and remove noise components,
Fine pulse-like patterns are determined to be noise and removed, degrading the image.

In particular, this tendency becomes more pronounced as the noise reduction effect is increased, so this is improved.

Therefore, in the present invention, the circuits shown in FIGS. 5 and 6 are improved so that the delay time τ due to the delay lines 2 and 3 is switched according to the state of the input signal level, for example, the strength of the input electric field. This structure enables effective noise removal even in a weak electric field with many noise components.

FIG. 8 is a block diagram showing the noise reduction circuit of the present invention.

In FIG. 8, the noise reduction circuit connects a delay line 8 with a delay time τ1 to the input terminal 17 to which the input signal Vin is applied, and further connects a delay line 19 with a delay time τl in series with the delay line 18. In addition, the input terminal 171c is connected to a delay line 2 having a delay time τ2 different from the delay time τl (for example, τ1 minus τ2).
0 is connected, and further a delay line 21 with a delay time τ2 is connected in series to the delay line 21. The output end of the delay line 18 and the output end of the delay line 20 are connected to input terminals a and b of the switch circuit 22, respectively, and
The output end of the delay line 19 and the output end of the delay line 21 are connected to input terminals a and b of a switch circuit 23, respectively. Switch circuits 22, 23 if In addition to input terminals a and b, control terminal C1 output terminal d is connected. When a signal of, for example, L" level (low level) is input to control terminal C, input terminal a and output terminal d are electrically connected. For example, when a high level signal is input to the control terminal C, the input terminal V is connected to the output terminal d.

Each control terminal C of the switch circuits 22 and 23 is connected to the output end of a signal level detection means, such as an AGC (automatic gain braking circuit) for detecting the input electric field strength, and when the input electric field strength is weak, the KAGC circuit 24 The 'H' level AGC voltage is applied to each control terminal C of the switch circuits 22 and 23, and the switch is switched from terminal A to terminal A. Note that in the case of a strong electric field or a normal electric field, the AGC circuit is The input terminal 17 and the output terminals of the switch circuits 22 and 23 are connected to the logic circuit 15 or 16 shown in FIG. 5 or 6. As mentioned above, 6 is a filter that removes the pulsed signals shown in P7 and Ps in Table 1.

The output Vout of the logic circuits 15 and 16 is taken out from the output terminal 25.

In such a configuration, when removing pulse noise, the delay time τ1 of the delay lines 18 and 19 should be a delay time set to remove high frequency noise that does not damage the picture, and , delay line 2
The delay time τ2 of 0 and 21 is set to be longer than the delay time τ1, and when noise is removed, it is set to remove noise of a frequency lower than the frequency of the noise removed by the delay time τIK. Therefore, in the case of a strong electric field with few noise components or a normal electric field, the noise is removed by the delay time τ to reduce noise without degrading the image, and in the case of a weak electric field with many noise components, the delay time τ2 is used to remove the noise. It is possible to automatically replace VC with VC+A in order to have a greater noise reduction effect.Furthermore, there will be no scratches at the edges at the rise or fall of the step signal.

Note that the above-mentioned noise reduction circuit describes the case where sampling is mainly performed in the horizontal direction in the configuration of a television circuit, but the same noise reduction circuit can be configured even if applied in the vertical direction or in both horizontal and vertical directions. can.

In the above embodiment, the case where the AGC voltage is used as a control signal for switching the switch circuits 22 and 23 has been described. That is, the amount of noise components is detected, and the switch circuit 22
, 23 may be switched.

〔Effect of the invention〕

As described above, according to the present invention, a pulse-like signal is detected from an input signal by using a delay means and a logic means by utilizing the correlation between adjacent pixels, and the protruding portion thereof is removed as noise. As a result, it is possible to prevent the blurring that occurs in the edge portion of the step signal portion of the video signal in the conventional noise reduction circuit, and also to provide two delay means, a signal supply means, and a signal level detection means. , the signal supply means is driven according to the level of the input electric field strength and the level of the alternating current component of the video signal, and the two delay means are switched to manually input the logic means, so that it can be used in periods with a lot of noise and periods with a little noise. Sampling can be performed with a corresponding delay time, and pulse-like noise components can be detected and removed while minimizing image deterioration.

[Brief explanation of drawings]

Fig. 1 is an explanatory diagram showing a sample of image processing by smoothing, Fig. 2 is a waveform diagram showing an example of a video signal waveform, and Fig. 3 is an enlarged part of the waveform in Fig. 2, showing a step signal waveform and An explanatory diagram showing a sample thereof, FIG. 4 is an explanatory diagram of operation for realizing the noise reduction circuit according to the present invention, FIGS. 5 and 6 are block diagrams showing an example of noise reduction means applied to the present invention, FIG. 7 is an explanatory diagram for explaining the operation of the noise reduction circuit of the present invention, and FIG. 8 is a block diagram showing the noise reduction circuit of the present invention. ], , 17...Input terminal, 2, 3.18
,19,20.21... Delay line, 4,
5, 6, 13. 14......AND circuit, 7, 8
, 10 , 11 , 12...OR circuit, 9
．． 25...Output terminal, 15, 16...
・Logic means, 18°19...first delay means,
20, 21... second delay means, 22, 2
3... Signal supply means (switch circuit), 24.
...Signal level detection means (AGC circuit). Figure 1 Figure 2 Figure 3 Figure 4

Claims (3)

[Claims] (1) Signal level detection means that detects the signal level in the process of being input and processed in an image processing device and outputs two Buddhist monk names according to the detected level, and a signal level detection means that inputs a video signal and converts this input signal into a first signal. and a second signal having delay times different from each other by a predetermined time with respect to the first signal. Third
a first delay means for outputting three signals, and a fourth delay means having a delay time different from the delay time of the first delay means for the first signal. When the second delay means for outputting the fifth signal and the output of the two Buddhist monks from the signal level detection means have one value, the first. Second. A third signal is outputted respectively, and when the output level of the two Buddhist priesthoods is the other value, the first signal. 4th. a signal supply means for outputting a fifth signal; and a signal supply means for outputting a fifth signal; Second.
Third signal or the skin 1. Food 4. Asahi 5 signal with 1 human power”
When the first signal level and the third signal level are equal and only the second signal level is different, or when the first signal level and the fifth signal level are When the second or fourth signal is equal and differs only in the fourth signal level, the second or fourth signal is straddled between the first and third signals. A noise reduction circuit comprising logic means for outputting a signal at the same level as the fifth signal. (2) The signal level detection means is constituted by a field strength detection circuit that detects the input electric field strength and outputs the two Buddha monk names depending on whether the strength is higher or lower than a predetermined electric field strength. A noise reduction circuit according to claim 1, characterized in that: (3) The signal level detection means is a circuit that extracts the video signal during the synchronization signal period and outputs the two Buddhist monk names depending on whether the wave components included in this period are higher or lower than a predetermined level. Is it possible to configure it? : The noise reduction circuit according to claim 1, wherein the noise reduction circuit is expressed as a percentage.