JPS6222312B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an apparatus that uses a frame memory to configure a recursive filter to reduce random noise in a video signal.

Generally, in a television signal, in a still part of the screen, the correlation between frames is large for signal components, but not for noise components. Therefore, the operation of calculating the correlation between the signal of the current frame and the signal of the previous frame, mixing the two signals at a ratio according to the magnitude of the correlation, and storing the mixed signal in the frame memory as a new television signal is performed. By repeating the process, it is possible to reduce the level of noise components with no frame correlation, without reducing the level of signal components only for image signals with little movement, and improve the S/N ratio.

A conventional example of a noise reduction device using frame correlation as described above will be described with reference to FIG.

The basic operation of this noise reduction device is to pass the television signal through a recursive filter.
The purpose of this is to pass the signal through a filter called "filter" and output a new television signal. A recursive filter has the characteristic that its output signal is determined not only by the input signal but also by the signal already output by the recursive filter.For example, the output section has a delay circuit, and the output of this delay circuit is In general, the signal is returned to the input side and mixed with the input signal, and the mixed signal is the input.

In Figure 1, 1 is an A/D converter that generates an encoded output, 2 is an arithmetic circuit that receives the output of A/D converter 1, and 3 inputs the output of A/D converter 1. 4 is a frame memory into which the output of the arithmetic circuit 2 is input, and the output of the frame memory 4 is input to the motion detection circuit 3 via the chroma inversion circuit 5. Reference numeral 6 denotes a mixing ratio generating circuit which generates a constant K to the arithmetic circuit 2 in accordance with the output of the motion detecting circuit 3. To explain the drive section of the frame memory 4, 7 is a sync separation circuit that separates a sync signal from an input television signal, 8 is a clock generation circuit that generates various clocks from the output of the sync separation circuit 7, and 9 is an address counter. , 10 is an R/W control circuit that controls reading and writing of the frame memory 4 based on the output of the address counter 9. Further, 11 is a D/A converter that converts the output of the arithmetic circuit 2 into an output television signal.

Next, the operation of this device will be explained. The input television signal is encoded by an A/D converter 1, and this encoded signal A is applied to an arithmetic circuit 2 and a motion detection circuit 3, respectively. The motion detection circuit 3 receives the signal A and the signal B of one frame before, which has the subcarrier phase inverted by the chroma inversion circuit and has the same phase as the signal A. The signal difference is determined for each picture element. The difference signal output for each picture element of the motion detection circuit 3 is applied to the mixing ratio generation circuit 6, and the value of the constant K is determined according to the difference signal level. The arithmetic circuit 2 receives the signal A, the signal B one frame before the signal A, and a constant K which is the output of the mixing ratio generating circuit 6.
is input, and the following calculations are performed.

C=(1-K)A+KB (0≦K≦1) As the value of K, for example, K=0, 1/4, 1/2, 3/
By using simple values such as 4 and 1, the configuration of the arithmetic circuit 2 itself becomes simple. The output signal of the arithmetic circuit 2 (signal C above) is sent to the frame memory 4 and the D/A converter 1.
It has been added to 1. The output of the D/A converter 11 is the output television signal.

Here, the output of frame memory 4 needs to be a signal just one frame before signal A.
Since the phase of the subcarrier of the television signal is inverted every frame, the chroma inversion circuit 5 inverts the phase to create signal B. In this way, the two input signals A of the motion detection circuit 3,
B becomes a signal with a time difference of one frame as long as the phases of the subcarriers match, and it becomes possible to obtain a difference signal between frames.

FIG. 3 shows the degree of improvement in the S/N ratio with respect to the value of the constant K when the screen does not move at all, that is, when there is no difference in signals between frames. For example, if there is no difference in signals between frames, setting K=0.6 will improve the S/N ratio by about 6 dB.

In the conventional device shown in Fig. 1, the degree of improvement in the S/N ratio is determined in accordance with the magnitude of movement in moving parts of the screen, and the S/N ratio in moving parts is higher than that in static parts. It has a characteristic that the degree of improvement in the N ratio gradually decreases. However, if the value of K for the signal difference between frames is set so that the degree of improvement is unnecessarily large, afterimages will increase, that is, resolution will deteriorate in moving parts due to these afterimages.

The present invention takes advantage of the characteristics of a recursive filter,
Especially when zooming up (enlarging a part of the screen) or panning (horizontal movement) with a TV camera,
This is an apparatus that detects this and obtains a good image with a good S/N ratio while reducing deterioration in resolution. FIG. 2 shows a block diagram of one embodiment of the apparatus and will be described. Note that in FIG. 2, circuits and devices having the same functions as those in FIG. 1 are given the same numbers.

The feature of this embodiment is that the absolute value of the output of the motion detection circuit 3 is taken, and this absolute value is transferred to the mixing ratio generation circuit 13.
The mixing ratio is controlled by inputting the D and comparing it with the set value D in the comparator circuit 14, and counting the number of absolute values that are equal to or higher than the set value.

This example will be explained in detail below.

In the motion detection circuit 3, the signal A of the current frame,
Find the difference signal M from the signal B one frame before,
The absolute value circuit 12 forms the absolute value |M| of the difference signal. And the absolute value |M| is the mixing ratio generation circuit 13
is input to generate a constant K. Here, FIG. 4 shows an example of the relationship between the absolute value |M| of the difference signal and the above-mentioned mixing constant K. The relationship shown in Figure 4 was set using PROM, and when the absolute value of the difference signal between frames is zero, that is, for a stationary part of the screen, |M| = 0 and K = 0.6, and the next 3rd to
From the figure, it can be seen that the degree of improvement in the S/N ratio is approximately 6 dB. As mentioned above, if the S/N ratio is improved too much with the above settings, the resolution will deteriorate.

On the other hand, the absolute value |M| is input to the comparison circuit 14,
It is compared with the set value D applied to the other input terminal of the comparator circuit 14. This comparison is performed between the absolute value |M| of each pixel and the set value D, and the set value D
The larger one is identified.

The output of the comparison circuit 14 is added to the counting circuit 15, where the absolute value greater than the set value D |
Count the number of M|. The counting period is one field or one frame of the input television signal, and a register is provided to hold the counted value during this period, and is reset every field or frame. The output of this counting circuit 15 is
The output of the mixing ratio generating circuit 13 is input to a mixing ratio control circuit 16 which has another input.

The mixing ratio control circuit 16 controls the value of the constant K generated according to the absolute value signal of the difference signal using the output of the calculation circuit 15. For example, if this circuit 16 is configured with a multiplier circuit, and when the output of the counting circuit 15 is small, the multiplier applied to the output of the mixing ratio generating circuit 13 is set to a value close to 1, and when the output of the counting circuit 15 is large, the multiplier is set to a value close to 1. Let be a value close to 0. In other words, when the entire screen moves significantly due to zooming up or camera panning, and most of the absolute values |M| of the difference signal signals of each picture element exceed the set value D, the output of the counting circuit 15 becomes large, so the mixing The constant K is forcibly set to a small value, thereby reducing the degree of improvement in the S/N ratio. This effect reduces afterimages in scenes with large movements, making it possible to reduce deterioration in resolution. On the other hand, in a normal screen with many static parts, there are few picture elements whose absolute value |M| exceeds the set value D, and therefore the output of the counting circuit 15 is small, so the value of the constant K is shown in FIG. It is almost the same as the previous image, but the degree of improvement in the S/N ratio is greater and an image with less noise can be obtained.

The mixing ratio control circuit 16 described above is not limited to a multiplication circuit; for example, a switching circuit that sets the mixing constant K to 0 when the output of the counting circuit 15 exceeds a certain value may be used to obtain the same effect as above. is obtained.

In the above embodiment, the absolute value |M| of the difference signal is compared with the set value D, and the number of pixels that is greater than or equal to the set value is counted to detect that a zooming or panning operation has been performed. By reducing the mixing constant K, the degree of improvement in the S/N ratio will be reduced, but since the movement of the screen is large, noise will not be noticeable, and by reducing K, the afterimage will be reduced and resolution deterioration will be suppressed. is possible. On the other hand, it goes without saying that noise is sufficiently reduced on screens with many static parts. Also, the mixing ratio control circuit 1
If a multiplier circuit is used in 6, the above effect will be noticeable depending on the speed of zooming up or panning.

As described above, the present invention reduces noise when there are few changes in the screen, and reduces afterimages when there are large changes in many parts of the screen, such as when zooming up or panning the camera, resulting in resolution degradation. By reducing the size of the image, a noise reduction device with very good image quality can be provided.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a conventional noise reduction device, FIG. 2 is a block diagram of a noise reduction device according to an embodiment of the present invention, FIG. 3 is a relationship diagram between constant K and S/N ratio improvement degree, and FIG. FIG. 4 is a diagram showing the relationship between the absolute value of the difference signal and the constant K. 2... Arithmetic circuit, 3... Motion detection circuit, 4...
Frame memory, 8...Clock generation circuit, 12
... Absolute value circuit, 13 ... Mixing ratio generation circuit, 1
4... Comparison circuit, 15... Counting circuit, 16... Mixing ratio control circuit.

Claims (1)

[Claims] 1 A movement that detects the movement between frames of an image by detecting the difference between the signal of one frame of a television signal (hereinafter referred to as the current frame) and the signal of the current frame one frame before. a detection circuit; an absolute value circuit that determines the absolute value of the output signal of the motion detection circuit; and a mixer that takes the output of the absolute value circuit as input and determines a mixing ratio between the signal of the current frame and the signal of one frame before the current frame. a ratio generating circuit; a comparison circuit that takes the output of the absolute value circuit as an input and compares and detects things that are above a certain set value; and a counter that takes the output of the comparison circuit as an input and counts the number of things that are above the certain set value. a mixing ratio control circuit that controls the output of the mixing ratio generating circuit according to the output of the counting circuit; Noise reduction characterized by comprising: an arithmetic circuit that mixes the signals, and a frame memory that stores the output signal of the arithmetic circuit as an input and adds the output to the motion detection circuit as a signal from one frame before. Device.