JPH04101579A - Television signal processor - Google Patents

Abstract

PURPOSE:To prevent unnatural moving image due to malfunction of motion detection by dividing an input television signal into a high band component and a low band component so as to pick them up, judging whether the input television signal is a moving picture or a still picture by means of the low band component and inter-frame processing the high band component for noise removal. CONSTITUTION:The low band component is picked up from a Y signal containing noise by a low-pass filter 22, and a reduction signal from the filter 22 is subtracted by an adder 23 to pick up the high band component. Then, an output signal from an adder 25 is supplied to a one-frame delay circuit 27, the high band signal of a signal in one frame before is picked up from an adder 29, and this high band component signal is supplied to a coefficient multiplier 30. In the meantime, output signals from the filter 22 and a filter 28 are supplied to an adder 31 to find the differential value, which is then supplied to a motion detection circuit 32 through an absolute circuit, and the condition of the motion of a picture is discriminated according to an input absolute value.

Description

DETAILED DESCRIPTION OF THE INVENTION (Object of the Invention) [Industrial Field of Application] The present invention relates to a television signal processing device that performs processing to remove noise contained in a television signal.

[Prior art] Advances in digital signal processing technology, and digital LS
2. Description of the Related Art In the current situation where high integration of I and mass production of large-capacity memories are progressing, digital processing of television signals is frequently performed using frame or field memories in signal processing of television receivers. One of the signal processing systems for television signals using frame or field memories is a motion-adaptive frame or inter-field noise reduction system. Figure 7 shows an example of a conventional noise reduction system. A luminance signal (Y signal) containing a noise component as shown by Xl in FIG. 8 is input to the input terminal 11. This human input Y signal is input to the first coefficient unit 12, and the output signal from this coefficient unit 12 is supplied to the adder 14 together with the output signal from the second coefficient unit 13. The output signal is supplied to an output terminal 15 and also to a one-frame delay circuit 16. Then, the output signal X2 from this delay circuit 1B is input to the second coefficient multiplier 13.

The first coefficient multiplier 12 has its gain controlled by the coefficient ko given from the nonlinear circuit 17, and the second coefficient multiplier 1
3 is changed to (1-,
o) Its gain is controlled by a factor of two. Then, the output signals from each of the coefficient multipliers 12 and 13 are added by an adder 14, and the signal X1 and the signal X2 are mixed at a mixing ratio ko.

Here, in the one frame delay circuit 16, a signal X obtained by delaying the output television signal from the adder 14 by one frame
In this signal processing circuit, the input signal X1 and the signal X2 on the past frame at exactly the same position on the screen are processed through coefficient units 12 and 13, and an adder 14. The noise components are removed by repeatedly taking the weighted average of the mixture at a mixing ratio of ko.

However, if you simply take the average value using past signals with the same screen position, if there is no correlation between the signal X1, which is the current image, and the signal X2, which is the past image, that is, the image has moved. In this case, in addition to noise removal, the image is mixed with an image at a different time, resulting in side effects such as afterimages. For this reason, a so-called motion adaptive method is being considered, which switches the noise reduction operation depending on the movement of the image.

In order to detect the movement of the image, first, the adder 17 calculates the difference value X3 between the current input signal X1 and the signal X2 one frame before.
seek. Then, the difference value between the levers of the absolute value circuit 19 is determined, and this absolute value is given to the nonlinear circuit 17.

Here, the characteristics of the nonlinear circuit 17 are set as shown in FIG. When the state exceeds βO, it outputs “1.0′ (-, k o), and conversely, when the absolute value is smaller than αO, it outputs “0”.And,
The coefficient multiplier 12 uses the output kO from this nonlinear circuit 17.
and 13, and controls the mixing ratio of the mixing circuit section including the adder 14.

That is, in this device, when the image is moving, the inter-frame difference value X3 is large as in the period shown by TAI in FIG. As shown by □, the difference value X3 between frames is small, and the magnitude of this difference value X3 is used to determine whether it is a moving image or a still image. Then, when it is determined that the image is a still image, the value of the mixing ratio ko is reduced to increase the effect of noise reduction processing by inter-frame calculations and strongly suppress noise components. Furthermore, when it is determined that it is a moving image, the value of the mixing ratio ko is increased to reduce the frame calculation effect, and the current television signal input to the input terminal 11 is output as is to prevent side effects such as afterimages. I have to.

However, if the image movement is directly detected from the inter-frame difference value using the fixedly set non-linear circuit 17 in this way, the period shown by To in FIG. Even in a still image portion, if a noise component with an amplitude exceeding the level blurred by αO of the characteristics of the nonlinear circuit 17 shown in FIG. 9 is input, this portion is determined to be a moving image. Therefore, noise reduction processing between frames is no longer performed, and noise components are output as they are. Furthermore, even in a moving image part such as the period indicated by TB in FIG. 8, the inter-frame difference value
3, and if this is less than αO, this portion is determined to be a still image and an interframe noise reduction effect is applied. As a result, an afterimage occurs and an unnatural screen is output.

Particularly, if such a malfunction occurs in a relatively flat part of the image, such as during the period T8, an afterimage will occur over a wide area, and its side effects will be very noticeable.

In order to prevent such side effects due to motion detection errors, it is conceivable to reduce the value of α0 so that, for example, the characteristics of the nonlinear circuit 17 can be determined to be a moving image even with a small inter-frame difference value. However, if we reduce the value of α0,
This time, for relatively large noises like the period shown by Tc, even if it is a still image, it will be easier to judge it as a moving image.
Therefore, the opposite problem arises.

In other words, in such a conventional motion adaptive noise reduction system, a motion detection signal for controlling the noise reduction operation is detected directly from the inter-frame difference value, so a signal with a relatively large amplitude is detected. For noise components, this is determined to be a moving image, and no noise removal operation is performed.

Also, in a moving image portion, a portion where the inter-frame difference value is relatively small may be determined to be a still image. In the past, interframe processing was performed in this area, resulting in unnecessary afterimages. In order to prevent the occurrence of such afterimages, the threshold value of the nonlinear circuit in the motion detection section is lowered to increase the motion detection sensitivity. come to life.

[Issue to be solved by the inventionffi] This invention has been made in view of the above points, and is capable of reliably capturing still images and moving images of television signals, regardless of the amplitude of noise components, etc. Noise reduction using inter-frame or inter-field processing can be used to prevent unnaturalness in moving images due to motion detection malfunctions while maintaining the noise reduction effect during inter-frame or inter-field processing. The present invention aims to provide a television signal processing device that performs the following steps.

(Structure of the Invention) [Means for Solving the Problems] A television signal processing device according to the present invention extracts, for example, a first signal component that is a high frequency component and a first signal component that is a low frequency component from an input television signal. The second signal component is used to determine whether it is a moving image or a still image, and the first signal component is subjected to interframe processing to perform noise removal processing. ing.

[Function] In a television signal processing device configured in this way, the input television signal generally contains many noise components in the high frequency side and not much in the low frequency side. . Therefore, the interframe processing that performs noise removal processing is performed on the first signal component, which is a high frequency component of the input signal, and the image movement is performed on the second signal component, which is a low frequency component with less noise component. I try to do this depending on the ingredients. Therefore, it is possible to reduce afterimage interference in a sufficiently large-area image without losing the noise reduction effect, and to be able to stably reproduce a high-quality image.

[Example] Hereinafter, one embodiment of the present invention will be described with reference to the drawings.

FIG. 1 shows the circuit configuration thereof, and a Y signal containing noise is inputted to an input terminal 21 as shown at 21 in FIG. This Y signal is supplied to a low-pass filter 22, and by passing through this filter 22, a low frequency component Z4 (see FIG. 2) in the Y signal is extracted. Further, the human input Y signal Z1 is supplied to an adder 23, and this adder 23
The low-frequency signal Z4 from the low-pass filter 22 is subtracted, and the high-frequency component in the Y signal is extracted. The high frequency component signal obtained from this adder is supplied to the first coefficient multiplier 24, and the low frequency component signal from the low pass filter 22 is supplied to the adder 25.

The output signal from the adder 25 is supplied to the output terminal 26, and this signal is further supplied to a one-frame delay circuit 27. The output signal shown as 22 in FIG.
Subtract the signal from.

That is, the adder 29 extracts the high-frequency component signal of the signal one frame before, and this high-frequency component signal is extracted from the second frame.
is supplied to the coefficient unit 30. Also, the low-pass filter 28
The low frequency component of the signal from one frame before is supplied to the adder 31. This adder 31 is supplied with the low-frequency signal component Z1 of the current frame from the robust filter 22, and subtracts the low-frequency signal component of the signal Z2 of the previous frame from this signal Z4. , this adder 31 outputs a signal Z3 of the difference value.

The first and second coefficient multipliers 24 and 30 are controlled by a control signal of 1 from the motion detection circuit 32,
The coefficient multipliers 24 and 30 limit the gain of the signal output from the adder 23 and the high-frequency components of the signal Z2, which is a past signal in units of frames, to 1 and (1-kl), respectively. .

Thereafter, the output signals from each of the coefficient multipliers 24 and 30 are added in an adder 33, and this added signal is added in an adder 25 as a high frequency signal component of the output signal. That is, the first and second coefficient multipliers 24.
30 and an adder 33 constitute a mixing circuit that can be controlled by the coefficient kl.

In this case, the output signals from the coefficient multipliers 24 and 30 include:
Since the input television signal does not contain low-frequency components, the output signal from the low-pass filter 22 and the adder 25
By adding them together, the low frequency components are compensated for.

In other words, in a signal processing device configured in this way, the inter-frame noise reduction processing is limited to noise components or high-frequency component signals that contain a relatively large amount, and only low-frequency components that contain few noise components are processed. Interframe processing is not performed on the area component signals.

The output signals from the low-pass filters 22 and 28, that is, the low frequency components of the current signal Z1 and the signal Z2 one frame before this, are supplied to an adder 31, and the difference value thereof is determined. This inter-frame difference value signal is supplied to an absolute value circuit 34, and its absolute value is taken. This absolute value is then supplied to the motion detection circuit 32, and the state of the motion of the image is determined in accordance with the input absolute value.

FIG. 3A shows a specific example of the configuration of this motion detection circuit 32. An output signal Z5 from the absolute value circuit 34 is input to a terminal 41 and first given to a nonlinear circuit 42. The characteristics of this nonlinear circuit 42 are set as shown in (B) in the figure, and when the output signal Z5 from the absolute value circuit 34 is less than the threshold value α1, it outputs "0" and the signal When Z5 is equal to or greater than the threshold value β1, “1.0
” is output.

The output signal Z6 from this non-linear circuit 42 is given to the shift register 43.
3 is configured as shown in FIG. 3(C), for example. That is, it is constructed by connecting n stages of D-type flip-flops 431 to 43n in series, and the supplied clock signal C
Input data is shifted one step in units of K and output in parallel.

The output signal from this shift register 43 is input to a maximum value circuit 44, and the maximum value of the output data from each of the n stages is outputted from a terminal 45.

FIG. 4 shows the state of signals in the motion detection circuit 32.

That is, in the conventional type explained using FIG.
While motion detection is performed based on signal components of all bands, in the device shown in this embodiment, the adder 31
The signal from which frame differences are taken is passed through a low-pass filter 2.
Motion detection is performed only on the low-frequency component signals extracted in steps 2 and 28.

Therefore, the inter-frame difference signal Z3 rises at the edge portion as shown by the period TD shown in FIG.
The decline is slowing down. As a result, this portion may not be detected as a video and may be missing.

Therefore, a shift register 43 configuring the motion detection circuit 32
, and a maximum value circuit 44 detect the maximum value of the motion signals of surrounding pixels, thereby spatially expanding the moving image portion to prevent omissions.

When the pixel is moving, the coefficient kl is set to be close to "1.0" by applying 1 to the motion detection signal detected in this manner. value to reduce the noise reduction effect caused by frame calculation. Conversely, when the frame difference value Z3 is small, it is determined that the image is a still image and the frame correlation is high, and the noise reduction effect by frame calculation processing is increased.

As explained above, in this embodiment, the signal source for detecting motion is limited to low-frequency component signals with relatively few noise components. Therefore, the inter-frame difference value signal Z3 output from the adder 31 contains a small amount of noise component, which is T shown in FIG. As shown by the period , even noise components with relatively large amplitudes can be prevented from being erroneously determined to be moving images. Furthermore, by reducing the influence of noise, it becomes possible to set the threshold values α1 and β1 of the nonlinear circuit 42 to smaller values. By expanding spatially, it is possible to reliably detect a moving portion as a moving image even for a moving portion where the inter-frame difference level is relatively small, such as the period TB shown in FIG.

In addition, by limiting the signal components to which frame calculations are applied to high-frequency parts, which generally have a lot of noise components, it is possible to maintain the noise reduction effect without losing the noise reduction effect. Since no calculation is performed, afterimage disturbance can be reduced over a wide range.

FIG. 5 shows another embodiment, in which the same components as those in the embodiment of FIG. 1 are given the same reference numerals and their explanations will be omitted.

In this embodiment, in the mixing circuit composed of the first and second coefficient multipliers 24 and 30 and the adder 33, the signal component to which interframe noise reduction processing is applied is a high-frequency component in the television signal. In addition, the signal source from which the inter-frame difference value is obtained in the adder 31 to detect motion is limited to the low-frequency components that are the outputs from the low-pass filters 22 and 28.

A signal obtained by subtracting the output signals from the low-pass filters 22 and 28 from each of the current television signal input from the input terminal 21 and the past signal in frame units output from the one-frame delay circuit 27;
That is, the high-frequency signal components SO and Sl of the current signal and the past signal are taken into the motion detection circuit 50, respectively.

FIG. 6(A) shows a specific example of the configuration of this motion detection circuit 50, in which high-frequency signal components SO and Sl of the current signal and past signals are detected by the absolute value circuit 51 and 52. The outputs from each of the absolute value circuits 51 and 52 are supplied to a maximum value circuit 53,
The maximum absolute value is selected and supplied to the comparator 54. This comparator 54 compares the signal with a threshold value ref input from a terminal 55, and outputs the output signal from this comparator 54 as a selection signal to a selector 56. This selector 56 uses the threshold value (α2,
Select one of β2) and (α3, β3).

FIG. 6B shows the characteristics of the nonlinear circuit 57.

Here, the control of the selector 56 is set so that when the output signal of the maximum value circuit 53 exceeds the threshold value ref of the comparator 54, the threshold value (α2, β2) is selected;
Conversely, when the output signal of the maximum value circuit 53 is below the threshold value rev of the comparator 54, the Ll value (α3, β
3) is selected.

That is, in this embodiment, an edge component having a relatively high level is detected from the current signal and a frame-by-frame past signal, whichever has a thicker high frequency level, and when this edge component is detected, the nonlinear circuit 57 The threshold values α and β are set smaller than usual, so that moving images can be detected even with smaller inter-frame differences. Therefore, missing detection of moving images in steep edge portions with large level differences can be reduced.

That is, according to this embodiment, it is possible to obtain a signal processing circuit that can improve afterimage interference in moving images and can realize noise removal with a sufficient noise reduction effect.

[Effects of the Invention] As described above, according to the television signal processing device according to the present invention, in a motion adaptive noise reduction system using inter-frame or inter-field processing, image motion detection is performed using low-frequency signals with relatively few noise components. By using these components, it is possible to suppress the occurrence of malfunctions due to noise, and the noise removal effect is effectively improved.

It is also said that it is possible to increase the sensitivity of motion detection at the same time, and by limiting the signal for inter-frame or field arithmetic processing to high-frequency components with relatively high noise components, it is possible to prevent afterimage interference in moving images. It will be improved.

[Brief explanation of the drawing]

FIG. 1 is a block diagram explaining a signal processing device according to an embodiment of the present invention, FIG. 2 is a signal waveform diagram explaining this embodiment, and FIG. 3 is a diagram of a motion detection circuit used in this embodiment. 4 is a signal waveform diagram, FIG. 5 is a block diagram illustrating another embodiment of the present invention, and FIG. 6 is a diagram illustrating a motion detection circuit used in this embodiment. , FIG. 7 is a block diagram showing a conventional noise removal device, FIG. 8 is a signal waveform diagram explaining this conventional example, and FIG. 9 is a diagram showing nonlinear characteristics. 22.28...Low pass filter, 23.25.29
．． 31.33...Adder, 24.30...Coefficient unit,
27-1 frame delay circuit, 32.50...Motion detection circuit, 34...Absolute value circuit. Applicant's agent Patent attorney Takehiko Suzue Figure 1 (B) (C) Figure 3 Figure 2 Figure 4 Figure (B)

Claims (2)

[Claims] (1) a first signal extraction means for extracting a signal component in a first frequency range from an input signal; and a second signal extracting means for extracting a signal component in a first frequency range from the input television signal; a second signal extraction means for extracting a signal component in a frequency range of; a signal including at least the first signal component;
signal delay means for outputting a third signal component delayed by at least one field with respect to the first signal component; a motion detection means for detecting motion; and a noise process for controlling and synthesizing the first and third signals based on the motion detection signal from the motion detection means and performing noise removal processing on the first signal component. 2. A television signal processing apparatus, comprising: means for processing a television signal, wherein the output signal is composed of the signal subjected to the noise removal processing and the second signal component. (2) the first signal extraction means extracts signal components centered on high frequency components in the input television signal;
2. The television signal processing apparatus according to claim 1, wherein said second signal extracting means extracts signal components centered on low frequency components in the input television signal.