JPS62626B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention relates to a television signal noise reduction device, and more particularly, to a noise reduction device for a television signal, in particular, a weight ratio between a current input signal and a signal predicted by a past input signal. The present invention relates to a television signal noise reduction device that performs control based on signal differences and reduces noise by using the correlation of each frame image signal.

[Background of the invention]

Conventionally, various noise reduction apparatuses for television signals of this type have been proposed (for example, Japanese Patent Publication No. 115018/1982 and Japanese Patent Publication No. 118724/1989).

First, problems with conventional devices will be explained using the example shown in FIG. 1, which is known to have a relatively simple configuration.

A part of the current input television signal input from the input terminal 1 of the device is subtracted by the subtracter 11 from the prediction signal (past signal) from the prediction circuit 4 (including the frame memory and chroma inverter). The output of 11 is converted into a corrected signal by a correction signal generation circuit 21, added to another part of the currently input television signal by an adder 31, and outputted from an output terminal 2 as a television signal with reduced noise. Ru. A portion of the output signal is applied to the prediction circuit 4, subjected to a predetermined delay and processing, and becomes a prediction signal.

In the circuit shown in FIG. 1, the input is x, and the prediction circuit 4
Letting the output of the subtracter 11 be w and the output of the adder 31 be y, the output of the subtracter 11 is expressed as (w-x). The correction signal generation circuit 21 generates K(w-
x) as a correction signal, and the adder 31 adds this correction signal to its input and outputs y=x+K(w-x)=(1-K)x+Kw (0K<1).

Systematically, in the part where |w−x| is small,
The subject is often stationary, and large parts are often moving. Therefore, by bringing K closer to 1 when |w−x| is small and closer to 0 when it is large, w(〓x), which has less noise than x in the stationary part, can be
By mainly outputting x in moving parts to avoid afterimages, it is possible to obtain y with improved image quality of input x.

FIG. 2 shows the configuration of the noise reduction device according to the application filed by the applicant, in which 12 is a subtracter;
2 represents a correction signal generation circuit, 32 an adder, and 4 a prediction circuit.

Here, the input is x, the output of the prediction circuit 4 is w, the output of the adder 32 is y, and the correction signal generation circuit 2
2 outputs (1-K) (x-w), which is the input (x-w) multiplied by the variable coefficient (1-K), then y = (1-K) (x-w) + w =(1-K)x+Kw, and it can be seen that an output y similar to that of the device shown in FIG. 1 can be obtained.

In the conventional device as described above, the prediction error, that is, the difference (x-w) between the input x and the output w of the prediction circuit, is
The determination as to whether the problem is mainly due to noise or movement becomes rough, and the image quality deteriorates in areas where the determination is incorrect. Conventionally, the following measures have been taken to address this problem.

(1) The variable coefficient K is made to depend not only on the prediction error but also on the signal smoothed by this, and when the signal smoothed by the prediction error is close to zero, it is assumed that there is no movement, and the overall K Increase the value of One example of this is a method of measuring the S/N of a video signal and controlling the value of K based on this. (Journal of the Television Society, Vol. 33, No. 4, pp. 296-300
page).

(2) Apply orthogonal transform such as Hadamard transform to the prediction error, and if power is concentrated in a specific component, it is assumed that there is movement and the value of K is reduced. (June 1978 Television Society Seminar ``Digital Processing Technology in Television''
Text, supplementary pages 7-13).

However, all of these and other conventional methods utilize the statistical properties of video signals;
There is a problem in that image quality deteriorates significantly in areas that have statistically exceptional values.

[Purpose of the invention]

An object of the present invention is to improve the above-mentioned television signal noise reduction device, and to provide a noise reduction device that obtains an appropriate correction signal depending on the magnitude of the noise and performs the noise reduction effect without deteriorating the image quality. That's true.

[Summary of the invention]

In order to achieve the above object, the present invention is configured to focus on the synchronization signal portion of the television signal, predict the magnitude of noise in the input television signal, and change the variable coefficient K according to this value. It is characterized by:

[Embodiments of the invention]

Hereinafter, the present invention will be explained in detail with reference to Examples.

FIG. 3 shows the configuration of one embodiment of the noise reduction device according to the present invention. In the figure, 11, 31 and 4 represent the same circuits as 11, 31 and 4 in FIG. 1, respectively. Also, 231 and 232
53 is a circuit for detecting the average size of the prediction error in the synchronization portion of the television signal; 63 is the output of the detection circuit 53; Therefore, it is a circuit that selects the outputs of the dedicated memories 231 and 232, and together with the dedicated memories, constitutes a circuit that creates a correction signal. FIG. 4 shows the configuration of one embodiment of the above detection circuit. where 531 and 53
2 is an input terminal for an input television signal and a prediction error, respectively, 533 is an input terminal for a predetermined threshold value, 534 is an output terminal for a control signal to the selection circuit 63, 535 is a circuit that takes an absolute value, and 536 is a reduction 537 is a pass filter, 537 is a subtracter, 538 is a circuit that takes a logic value "1" in the synchronization part of the television signal or a part thereof, and creates an output that becomes a logic value "0" in the remaining part, 539 is a circuit 538 It is set when the output of the subtracter 537 takes a logical value "1" and the output of the subtracter 537 is positive, and the circuit 53
8 takes the logical value "1", and the subtracter 53
This is a flip-flop that is reset when the output of 7 is negative.

As is clear from the above configuration, when the noise superimposed on the synchronizing signal portion is large, the flip-flop 539 is set, and when it is small, the flip-flop 539 is reset. In the synchronization signal part, unlike the video signal part, the predicted signal is almost zero if there is no noise, regardless of the statistical properties of the subject. This allows you to know the level of noise present.

The read-only memories 231 and 232 respectively store a value of the variable coefficient K suitable when there is little noise, and a value of the variable coefficient K suitable when there is a lot of noise. (the value of the read-only memory 232 is generally larger than the value of the read-only memory 232), and the selection circuit 63 selects the read-only memory 2 according to the output of the control circuit 53.
Either the output of 31 or 232 is selected and input to the adder 31.

The method of changing the variable coefficient K according to the magnitude of noise is not limited to the method described in the above embodiment.

Another embodiment of the invention is shown in FIG. The embodiment shown in FIG. 3 implements the present invention in a circuit of the type shown in FIG. 1, while the embodiment shown in FIG. 5 implements the present invention in a circuit of the type shown in FIG. . In this figure, the same components and functions as those in FIG. 2 are given the same numbers as corresponding parts in FIG. 2. Also, 54 and 64 are respectively 5 in FIG.
This circuit has the same configuration and function as those of No. 3 and No. 63. And 741 and 742 are coefficient multipliers that multiply by a constant coefficient. Of course, it may be constructed using a memory as in FIG. 3. In this embodiment, the magnitude of the noise is identified by the detection circuit 54, and if there is a lot of noise, the coefficient multiplier 741
and 742, the selection circuit 64 operates to select the one with the smaller coefficient value. As a special case of this embodiment, the coefficient value of the coefficient multiplier is set to 1 and
There is a configuration that makes it 1/2.

In the above embodiment, the level of noise in the synchronized portion of the input television signal is classified into two types using a single threshold value, and the control circuit outputs a binary signal, but it is possible to obtain better image quality. In order to do this, a multivalued control signal is created using multiple threshold values, and a selection circuit selects the output of a read-only memory or coefficient multiplier provided corresponding to each value of the control signal. Just configure it.

〔Effect of the invention〕

As is clear from the principle of operation, the present invention is capable of eliminating noise in the synchronization signal portion, although a lot of noise is superimposed on the video signal, such as the output of a television camera captured in a place with poor lighting conditions. It is not effective if there are few. However, like video tape recorder signals that have deteriorated due to repeated recording and playback, and television signals that have deteriorated due to analog multi-stage relay, television signals have noise superimposed on both the video signal and the synchronization signal. For signals, it does not depend on the statistical properties of the image, so
Image quality can be improved extremely effectively.

[Brief explanation of the drawing]

Figures 1 and 2 are diagrams showing the configuration of a conventional device;
3 and 5 are both diagrams showing the configuration of an embodiment of the noise reduction device according to the present invention, and FIG.
5 is a block diagram showing the configuration of a detection circuit 53 shown in the figure. FIG. 1...Input terminal, 2...Output terminal, 11,1
2,537...subtractor, 31,32...adder,
4... Prediction circuit, 22... Correction signal creation circuit, 5
3, 54...detection circuit, 63, 64...selection circuit, 231, 232...read-only memory, 74
1,742...Coefficient multiplier.

Claims (1)

[Claims] 1. A prediction circuit that predicts an input television signal, and a correction signal generation circuit that generates a correction signal according to a prediction error that is a difference between the input television signal and the output of the prediction circuit, and In a television signal noise reduction device that reduces noise by adding or subtracting noise from an input television signal or the output of the prediction circuit, the correction signal generating circuit calculates the average of the prediction error in the synchronized portion of the television signal. 1. A television signal noise reduction device comprising: a detection circuit for detecting the magnitude of a signal; and a circuit for producing the correction signal based on the output of the detection circuit.