JPS6382064A - Noise reducer - Google Patents

Abstract

PURPOSE:To greatly reduce the scale of a hardware by deciding which of a current line signal and a line signal which is one field above or below, for example, has larger correlation without any arithmetic between frames, and performing noise reducing operation by using the line signal having the larger correlation. CONSTITUTION:A television signal from an input terminal 11 is inputted to a field memory 14 through an adder 13. The output of the field memory 14 is inputted to an adder 17, a line memory 16, and a selector 15. The adder 17 outputs the difference between the signal of the current field and a signal which is 262H (horizontal scanning period) delayed to an absolute value circuit 19. The output of the line memory 16 is inputted to the selector 15 and also inputted to an adder 18. The adder 18 outputs the difference between the current field signal and a 263H delayed signal to an absolute value circuit 20. A decision circuit 21 decides which of the outputs of the absolute value circuit 19 and 20 have larger correlation and the selector 15 is so controlled as to select the signal having the larger correlation. The output of the selector 15 is inputted to an adder 12 and a nonlinear circuit 23 to perform operation as a noise reducer.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial Application Field) The present invention relates to a noise reducer that reduces noise by inter-field processing.

(Prior Art) Much research and development has been conducted on methods for reducing noise contained in television signals. Among them,
The method using a nonlinear low-pass p-wave filter in the time direction of a television signal is called a noise reducer, and can achieve significant noise reduction.

The principle of noise reduction by the noise reducer will be explained with reference to FIG. An image signal is input through input terminal 1. The input image signal is multiplied by (1-K) by the amplifier 2 and input to the adder 3.

The output of the adder 3 is simultaneously K when output from the output terminal 7.
, are input to the frame memory 4. The 7-frame memory 4 outputs the signal of one frame before, and inputs it to the adder 3 via the amplifier 5. The above operations reduce noise.

Now, let the input signal be ``(n), and the output signal be y(n), 1
Assuming that the output signal before the frame is 3'(n-Nr), the above operation is expressed by the following equation.

)'(n)=(I K)X(n)+KY(n -NF
) −(1) Converting the above equation into 2 yields Y(1)=(1-K)X(1)+KY(1)Z″″N′
Ru.

Therefore, noise can be reduced. In actual application to television signals, K is set to a value close to 1 in the still part, and K is set to a value close to 0 in the moving picture part, since blurring occurs when calculations are performed between frames in the moving picture part. this is,
It is sufficient to change the value of j5 based on the difference signal between one frame. A noise reducer using such a method is shown in FIG. A television signal is input to input terminal 31, and then input to adders 32 and 34. The output of adder 32 is input to frame memory 33. For example, in the case of an NTSC signal, this is a delay of 525H (H: 1 horizontal period). The output of the frame memory 33 is sent to the adder 3
4 is input. The adder 34 supplies a difference signal between the current frame and the previous frame signal to the nonlinear circuit 35.

Here, the characteristics of the nonlinear circuit 35 are as shown in FIG. 4, when the input signal is small, the input signal is multiplied by and outputted.
When the input signal is large, the output is 0. Therefore, when the inter-frame difference signal is large and it is determined that it is a moving image, the operation as a noise resolver is stopped, and when the inter-frame difference signal is small and it is determined that it is a still image, it is operated as a noise reducer. Perform noise reduction. Nonlinear circuit 35
The output of is input to the adder 32 and subtracted from the signal of the current frame.

The above operation is '(n) = "(n) -K("(n) -'(n-
Np) ) ”'(2) and when expanded, (1)
It will be the same as the formula.

Therefore, the noise reducer shown in FIG. 3 stops the noise reduction function for moving images to prevent image blurring, and performs the noise reduction operation according to (2) 2 for still images. Therefore, significant noise reduction is possible in still images.

(Problems to be Solved by the Invention) As described above, the noise reducer can reduce noise in still images, but requires a frame memory that delays the image by one frame.

For example, if the frame memory performs 8-pit quantization on an NTSC signal at a sampling frequency of 4fIle, the frame memory will produce 910 [samples] x 525 [lines] x 8 (btt).
A capacity of 3/3 Mbt is required, and the hardware becomes large-scale.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a noise reducer that can reduce noise in the same way as conventional noise reducers and can significantly reduce the hardware scale.

[Structure of the Invention] (Means for Solving the Problems) The noise reducer according to the present invention does not perform calculations between frames, and determines whether the current line signal and the previous line signal, for example, the upper or lower line signal, are correlated. The noise reduction operation is performed using a line signal with a large correlation.

(Function) Since the above method uses calculations between fields, the capacity is half that of the conventional noise reducer, and the hardware scale is significantly smaller. Since calculation is performed, it is possible to significantly reduce noise in still images.

(Example) Hereinafter, an example of the present invention will be described in detail with reference to the drawings.

FIG. 1 shows an embodiment of the present invention. For example, the input television signal is an NTSC signal.

A television signal is input to input terminal 1.

The input signals are input to adders 13.12.17 and 18. The output of the adder 13 is input to the field memo I) 14. This provides a 262H delay. Adder 1
The output of No. 3 is also output to the output terminal 22.

The output of field memory 14 is sent to adder 17° line memory 16. It is input to the selector 15.

The adder 17 outputs the difference between the current field signal and the 262H delayed signal and inputs it to the absolute value circuit 19. The output of the line memory 16 is input to the selector 15 and also to the adder 18.

The adder 18 outputs the difference between the current field and the 263H delayed signal and inputs it to the absolute value circuit 20.

The outputs of the absolute value circuits 19 and 20 are input to the judgment circuit 21,
It is determined which signal has the greater correlation, and a control signal is output to the selector circuit 15 to control the signal having the greater correlation to be selected. This operation is as shown in Fig. 5, when viewing a TV signal on a time-vertical basis, if it is an NTSC system, 262H is the lower line 53 and 263H is the upper line for the current signal 51. It is 52. Judgment circuit 2
The first action is.

As shown in FIG.
This is to determine which line of 3 has a higher correlation.

The output signal of the selector 15 is input to the adder 12, subtracted from the current signal, and input to the nonlinear circuit 23.

The nonlinear 23 has characteristics similar to those shown in FIG. The output of the nonlinear circuit 23 is input to the adder 13, which operates as a noise reducer.

As described above, since the noise reducer according to the present invention performs inter-field correlation calculation, it is possible to perform noise reduction even for a moving image of about 1 field and 1 line, since the inter-field correlation is large.

Although the above explanation has been made regarding NTSC system signals, the noise reducer according to the present invention can be similarly applied to other television signals such as PAL and SECAM systems.

Further, in the embodiment described above, a correlation calculation is performed between the upper and lower line signals and the current TV signal, but the number of line signals to be used may be increased.

FIG. 6 shows another embodiment of the invention. In Figure 6,
Low-pass filter (LPFIJ, adders 62, 63
, except for the low-pass filter (LPF) 64, are the same as in the previous embodiment and operate in the same manner.

The second embodiment shown in FIG. 6 is applied when the input signal is a color signal.
It is a TSC system composite color signal. An input signal is applied to input terminal 11 . The input signal is passed through a low-pass filter 6
1 and added to adder 62.

Here, the cutoff frequency fc of the low-pass filter 61 is 2 MHz. Therefore, the low pass filter 6
The output signal of No. 1 is only a luminance signal. On the other hand, adder 6
2 subtracts the output signal from the input signal of the low-pass filter 610, and supplies the output to the adder 63. The output signal of the low-pass filter 61 is subjected to synchronous addition by the adder 13 and subjected to noise reduction.
Otherwise, the original composite signal is regenerated. The output of the adder 63 is input to the 262H delay device 14. The output of the 262H delay device 14 is input to a low pass filter 64.

The low-pass filter 64 has the same characteristics as the low-pass filter RO1, and outputs components of 2 MHz or less. The subsequent operation is similar to that of the first embodiment described above. In addition to the above operations, it can also be applied to composite signals.

2 Ml (can reduce noise of components below z.

(Effects of the Invention) As described above, the noise reducer of the present invention can significantly reduce noise in still images with a hardware scale that is about half that of conventional noise reducers. Furthermore, it is possible to reduce the noise of a moving image of about one line per field, which could not be reduced with conventional noise reducers. Furthermore, by using a low-pass filter, it is possible to effectively reduce the noise of the reduced component of the composite signal.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing an embodiment of the present invention, Fig. 2 is a diagram explaining the principle of a noise reducer, Fig. 3 is a configuration diagram of a conventional noise reducer, and Fig. 4 is a characteristic of a nonlinear circuit possessed by the noise reducer. 5 is an explanatory diagram of the operation of an embodiment of the present invention, and FIG. 6 is a block diagram showing another embodiment of the present invention. 12.13, 17, 111, 62.63...adder,
14...Field memory, 15...Selector, 1
6... Line memory, 19.20... Absolute value circuit,
21... Judgment circuit, 61.64... Low pass filter. Applicant's agent Patent attorney Takeshi Suzue Industrial Figure 3 Figure 4

Claims (1)

[Claims] A noise reducer for reducing noise in a television signal, comprising: means for obtaining the absolute value of each difference between a current signal of the television signal and a plurality of field-delayed scanning line signals; Selects a scanning line signal that is equal to and means for subtracting the output of the means from the television signal to derive a noise-reduced output.