JP3513161B2 - Impulse noise eliminator - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [0001] BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a satellite broadcast receiving apparatus and the like.
Regarding a suitable impulse noise removal device, especially for moving images
To reduce false detection of impulse noise
The present invention relates to a configuration of pulse noise detection means. [0002] 2. Description of the Related Art Satellite broadcasting radio waves are in the microwave band.
Reception level is low due to rainfall or snow on the antenna.
When lowered, the C / N (carrier / noise) of the received signal is low.
Down. As C / N decreases, after FM demodulation
S / N (signal / noise) of the video signal of
Furthermore, C / N decreases and exceeds the threshold value.
When this happens, an impulse noise called medaka noise is generated.
Live. FIG. 6 shows medaka noise generated in a video signal.
This shows a waveform, in which the luminance changes stepwise as shown in FIG.
As shown in (b), C / N is
As a result, medaka noise B occurs. This medaka noise
Occurs, the brightness becomes extremely high for several hundred nanoseconds.
It becomes low and low. Watch this on the monitor screen
It looks like a white or black medaka, and the image quality is large
Be impaired. To remove such medaka noise
For example, August 1990, The Television Society of Japan
Technical Report, Vol. 14, No. 42, pp. 16-18
A known noise removing circuit as shown in FIG.
In FIG. 7, reference numeral 51 denotes a digital noise filter;
Is a difference circuit, 53 is a comparison circuit, and 54 is an erroneous detection prevention circuit.
is there. First, the demodulated signal is A / D converted and the digital notch
Filter 51 separates color subcarrier and digital audio subcarrier
Remove. Next, the high frequency component of the video signal is
Start. The signal is compared with the threshold value by the level comparison circuit 53
Pixels that exceed the threshold are considered as impulse noise candidates.
judge. In order to prevent erroneous detection, the circuit 54
Only if the element is not a candidate for impulse noise,
Judge as pulse noise. [0004] Using this noise elimination circuit, a video signal
Medaka noise is detected from the demodulated signal of
Pixels containing noise are reduced to one field or one scan line
By replacing pixels with no medaka noise
Removes medaka noise. Specifically, replace
When selecting a pixel, the frequency multiplex
The color subcarrier phase is continuous after replacing pixels.
As shown in FIG. 3, one field or one run
It is necessary to replace pixels with oblique pixels before and after the scanning. [0005] SUMMARY OF THE INVENTION
Conventional noise detection based on horizontal and vertical correlation
In the impulse noise detection circuit, false detection of medaka noise
There was a drawback that easily occurred. In addition, in conventional circuits,
Since the image signal is interpolated as it is, the phase of the color subcarrier
Need to be interpolated, so it is necessary to interpolate,
Since interpolation is performed with adjacent diagonal pixels, the accuracy of interpolation is
There was a problem of bad. Accordingly, the present applicant has previously received the
The C / N of the communication signal decreased and impulse noise occurred
Detection and elimination of impulse noise
To ensure accurate and accurate compensation for impaired image quality
Japanese Patent Application No. Hei 3-16 / 1990
No. 3626. FIG. 8 shows the above noise detection circuit.
A television receiver including a channel, which will be described in detail below.
You. In the figure, a signal received by an antenna 1
The received signal is frequency-converted to the first IF signal 3 in the LNB 2.
The number is converted and sent to the tuner 4. In tuner 4, hope
Channel is selected and FM demodulated. This tuner 4
Outputs video signal 5, detection signal 7, and audio signal 11.
I do. The detection signal 7 is a signal immediately after FM demodulation.
After de-emphasizing the detection signal 7, a 15Hz triangular wave
And pass through the LPF of the video signal band
The video signal 5 is obtained. In the noise reduction circuit 6, the
Detects and removes dhaka noise.
Will be described later. In the noise reduction circuit 6, the
When detecting noise, the part that is not medaka noise is detected.
When erroneously detected as erroneous noise (erroneous detection)
The noise removal operation is performed for this erroneously detected part.
In this case, the image quality of that part is impaired. Therefore, false detection
Probability is sufficiently smaller than the probability of medaka noise
Must. No medaka noise
When C / N (for example, about 6 dB or less) is obtained
Prohibits noise removal in the noise reduction circuit 6.
Need to be To this end, a noise detection circuit 8 is provided, and C
/ N is detected. This noise detection circuit 8 detects the detection signal.
No. 7 only the noise component contained in the signal is detected and the noise level is detected.
(For example, about 1-2 V).
And C / N at which medaka noise is generated from the voltage
The noise reduction circuit 6 determines whether or not the
Output as a noise reduction control signal 9
It is. This control signal 9 causes the occurrence of medaka noise.
Noise reduction is used to indicate that
Circuit 6 detects medaka noise in video signal
And interpolate the luminance of that part to reduce noise
It is supplied as a video signal 10 to a monitor 12. Also control
Signal 9 is C / N with no medaka noise
Indicates that the video signal 5 from the tuner 4 is not interpolated.
Then, it is supplied to the monitor 12 as it is. On the other hand, tuner 4
Is supplied to the monitor 12 as in the prior art. FIG. 9 is a schematic diagram of the noise reduction circuit.
The configuration is shown. In the figure, FM demodulation from tuner 4
The converted video signal 5 is converted to 4 fsc by the A / D converter 13.
A / D at a sampling frequency (approximately 14.318 MHz)
After being converted, the luminance signal (Y signal) is
It is separated into a roma signal (C signal). Y signal is impulse
The Y signal is applied to the noise detection circuit 6-1 and
Detects impulse noise (medaka noise) and detects it
Depending on the result, Y by the noise interpolation circuits 6-2 and 6-3
For the signal and the C signal, noise detected pixels are
Are interpolated with adjacent pixels not including. Interpolated Y and C signals
The signals are mixed by the YC mix circuit 39 and D / A converted.
The signal is D / A converted by the circuit 40 and output. FIG. 5 shows a specific example of the noise reduction circuit 6.
FIG. 13 is a block diagram showing a configuration example, where 13 is an A / D converter, 14
Is a YC separation circuit, 15 is a luminance signal, 16-1 to 16-4
Is a line memory, 17 is a 1H delayed luminance signal, 18 is 2H
Delayed luminance signal, 19 is a frame memory, 20 is a frame
The delayed luminance signals, 21 to 23 are difference signals, 24-1 to 24-
3 is an absolute value circuit, 25-1 to 25-3 are comparison circuits, 26
Is a threshold, 27 is a 3-input AND circuit, 28 is a medakano
Noise detection signal, 29 is an interpolation luminance signal, and 30 and 31 are switches.
Switch, 32 is a chroma signal, 33 is a 1H delayed chroma signal,
34 is a 2H delay chroma signal, 35 is a phase inversion circuit, 36
Is an interpolation chroma signal, 37 and 38 are switches, 39 is YC
Mix circuit, 40 is a DA converter, 41-1 to 41-3
Is a subtractor, 42-1 to 42-2 are adders, 43-1 to 4-4.
3-2 indicates a 1/2 multiplier. In FIG. 5, the video signal from the tuner 4
5 passes through an A / D converter 13 and then a YC separation circuit 14
To separate into a luminance signal 15 and a chroma signal 32,
It is supplied to the line memories 16-1 to 16-4. And
1H from the line memories 16-1 and 16-2 respectively
And 2H delayed luminance signals 17 and 18 are obtained. Meanwhile, the line
1H and 2H clocks from the memories 16-3 to 16-4, respectively.
Roma signals 33 and 34 are obtained. Also, a 1H delayed luminance signal
17 through a frame memory 19 to output a frame delay luminance signal.
No. 20 is obtained. Next, the 1H delayed luminance signal n (0) 17
2H delayed luminance signal n (A) 18, luminance signal n (B)
15 and the frame delayed luminance signal n (C) 20
Subtracters 41-1 to 41-3 perform subtraction, and a difference signal between them is given by:
That is, n (0) -n (A) signal 21, n (0) -n
(B) Calculate signal 22, n (0) -n (C) signal 23
Then, those difference signals are sent to absolute value circuits 24-1 to 24-3.
Absolute values are obtained through the comparators 25-1 to 25-25
-3 is used to determine the magnitude of the threshold value 26, and the result is entered as 3
What is obtained by inputting to the force AND circuit 27 is Medakano
Noise detection signal 28. Here, the threshold 26 is
To set the optimal value. When medaka noise is detected, it is detected.
The operation of replacing the output part with some signal,
Time is needed. Therefore,FIG.Now, the luminance signals 15 and 2
The H-delayed luminance signal 18 is added to the adder 42-1, the 1/2 multiplier 4
Through 3-1 to perform interpolation by the average value of those signals.
No, this is input to the switch 30 as the interpolation luminance signal 29.
Power. In the switch 30, the 1H delayed luminance signal 17
And the interpolation luminance signal 29 are output from the AND circuit 27.
It is selected by the medaka noise detection signal 28 and output. S
In switch 31, C / N with no medaka noise
The 1H delayed luminance signal 17 is selected,
Select the output of switch 30 in C / N where
Control by the noise reduction control signal 9.
You. As for the chroma signal 32, medaka noise
Is detected, interpolation must be performed. There
The adder adds the chroma signal 32 and the 2H delayed chroma signal 34
42-2.
The average value of the signal is obtained, and the
By inverting the phase, an interpolation chroma signal 36 is obtained.
H delay chroma signal 33 and input to switches 37 and 38
Switch 37 that operates in synchronization with the switch 30.
To select and output the signal based on the medaka noise detection signal 28.
You. The switch 38 is synchronized with the switch 31 and
Is exactly the same. Here, the interpolation chroma signal 36 is inverted in phase.
The reason why the phase is reversed by the path 35 is that the NTSC method is used.
Means that the chroma signal has its phase inverted every line
Therefore, the interpolation chroma signal 36 is the same as the 1H delayed chroma signal 33.
This is to make the phase. Output of switches 31 and 38
Are combined in a YC mix circuit 39, and then a D / A converter
Analog signal by 40, noise reduction image
Output as signal 10. Now, a television image is viewed in pixel units.
Is highly correlated with pixels that are adjacent in the horizontal and vertical directions.
Also, high correlation is shown between adjacent frames. to this
In contrast, impulse noise must be random in time
From the correlation between lines (scanning lines) and between frames
Very low. The impulse noise detector in FIG.
Uses these characteristics to detect impulse noise
The outline is described below. First, smell the Y signal
The target pixel is located above and below in the same field.
Perform a level comparison with each pixel and determine the absolute value of the level difference.
When both the upper and lower pixels exceed the set threshold
Is determined as the first candidate of the impulse noise. Figure this
10 will be described. In FIG. 1, the target pixel is S₁₁,Up
The pixel at the same horizontal position of the line is S_TenThe same water in the lower line
The pixel at the flat position is S₁₂And the threshold is R₁Then ｜ S₁₁-S_Ten|> R₁ , And | S₁₁-S₁₂|> R₁ When the condition of is satisfied, S₁₁Is the impulse noise
It is assumed that the candidate is 1. So vertical correlation?
Is detected when the impulse noise is constant.
Noise is horizontal when viewed in pixel units due to the spacing
This is because there is a correlation in the direction. However, the correlation between the upper and lower lines is
The first candidate of the detected impulse noise is an image signal.
Contains vertical high-frequency components in the issue, for example, a telop sentence
Erroneous detection of vertical edges such as characters as noise
There is. To prevent such erroneous detection, one frame
Pixel S before_{twenty one}Level comparison with
The absolute value of the difference is the threshold R₁Impulse
Let it be a noise candidate. That is, the first impulse noise
Among the candidates, ｜ S₁₁-S_{twenty one}|> R₁ Is satisfied, the pixel S₁₁The impulse noise
As a candidate. According to the above impulse noise detection method,
For example, false detection can be almost completely prevented for still images.
Can be. By the way, the above-mentioned noise detection method is a kind of
This is noise extraction by a filter.
Show the extracted frequency domain on the vertical-time frequency axis.
For example, the hatched portion in FIG. However, this noise detection method is suitable for moving images.
If used, the frequency domain shown by the hatched portion in FIG.
An erroneous detection is performed on the image signal. This false positive
When interpolation is performed with adjacent pixels,
Distortion results in deterioration of image quality. An object of the present invention is to specify a specific frequency in a moving image.
Of image distortion due to erroneous detection of impulse noise in the region
Impulse noise eliminator that prevents image generation and improves image quality
To propose a location. [0026] [MEANS FOR SOLVING THE PROBLEMS] To achieve the above object
The impulse noise elimination device of the present invention
Demodulates the received signal from the
Demodulation means for obtaining a video signal including
Noise component, and based on that noise component amount,
Determines the presence or absence of impulse noise components contained in the video signal.
Noise detection and determination means that outputs a noise determination signal
And the noise detection determination means,
When the noise component is detected, the noise component
Output interpolated video signal that has been subjected to luminance component interpolation processing
Signal interpolating means, a video signal containing the luminance information,
VideoDelayed the signal by a predetermined frameDelayed videoSignal andDifference
FromFirst and secondExtract frame difference component signalAnd the second
The first and second frame difference component signals are respectively
Within the same frame as the video signal corresponding to the noise
Each of the existing video signals after one scanning line and one scanning line before
Frame difference component signalFrame difference component extraction means,
First and second frame difference component signalsAnd predetermined reference signalWhen
And the level of the two frame difference component signal is
When the level is higher than the level of the reference signal, the impulse noise
Exists in the same frame as the video signal corresponding to the part
Motion in video signal after one scanning line and one scanning line before
1H after motion detection signal and 1H before motion detection signal indicatingGet
The motion in the video signal based on these two detection signals.
Is detected, and a motion detection signal is output according to the presence or absence of motion.
The motion is detected by the motion detecting means for applying force and the motion detecting means.
If it is detected, make sure that the interpolation process is not performed.
Control means for controlling the signal interpolation means.
Features. [0027] In the apparatus of the present invention, the signal obtained from the received signal is used.
The noise component contained in the detected signal is detected, and the noise component is detected.
Determine the presence or absence of an impulse noise component based on the component amount
A noise determination signal is output. This noise judgment signal
When the impulse noise component is present,
Detects the impulse part in the image signal and
The image with the impulse noise part removed by interpolating the degree
A signal is output. When there is no impulse noise component,
The video signal is output. A video signal containing the luminance information
And a signal obtained by delaying this signal for a predetermined frame period.
Between the frame difference component signal extracted from
The levels are compared. Based on the comparison result,
The motion that detected the presence or absence of motion in the upper and lower adjacent pixels in the field
A detection signal is output. This motion detection signal
Only when a detection is made, the impulse noise component
Only the video signal is output, assuming that there is no video signal. [0029] BRIEF DESCRIPTION OF THE DRAWINGS FIG.
Prevention of erroneous detection of impulse noise in the moving image
False detection occursFIG.In the time-frequency domain
Pixels are detected, and the detected pixels are
It becomes possible by excluding from the noise noise candidates. But,
By the frame correlation of one pixel as is usually done,
In the motion detection, the method for detecting the impulse noise and the total
Same and noiseWhenUnsuitable because movement cannot be determined
That's right. Therefore, the present invention provides a motion as described below.
The detection method was adopted. FIG. 3 shows the above-described motion detection method.
Elephant pixel S₁₁Upper and lower pixels S_TenAnd S₁₂Together
Think of the three pixels as one block,_TenAnd S₁₂of
If a motion is detected in any one of the pixels S₁₁To
It is determined that there has been movement. Pixel S_TenAnd S₁₂Motion detection
Is the pixel S one frame before each₂₀, S_{twenty two}Level with
Perform a comparison, and perform a comparison based on the level difference and the threshold value.
U. That is, the threshold for motion detection is R_TwoThen ｜ S_Ten-S₂₀|> R_Two  Or | S₁₂-S_{twenty two}|> R_Two S if the condition of₁₁Is determined to have moved,
Exclude from impulse noise candidates. FIG. 4 illustrates the motion detection method of the present invention described above.
Fig. 2 shows a basic circuit configuration for performing the operation. In the figure, 2
00 is the pixel S₁₂, 201 is S₁₁, 202 is S_Ten, 2
03-1 to 203-4 are line memory, 204 is frame
Memory, 205 is S_{twenty two}, 206 is S_{twenty one}, 207
S₂₀, 208 is the motion detection signal before 1H, 209 is after 1H
, 210 is a motion detection signal, 211 is an OR
Circuits, 212-1 and 212-2 are frame correlation determination circuits
It is. S₁₂(200) is the line memory 203-1
And the line memory 2
S from 03-1₁₁(201) is obtained and the line memory
S through 203-2_TenIs output. Frame memory
S from 204_{twenty two}Is output to the line memory 203-3.
By S_{twenty one}(206) and the line memory 2
S through 03-4₂₀Is output. The frame correlation judgment circuit 212-1 calculates | S_Ten
-S₂₀|> R_TwoAnd S₁₁S 1H before (201)
_TenPerform motion detection for (202) and perform 1H previous motion detection
An output signal 208 is output. Also, a frame correlation determination circuit 2
12-2 is | S₁₂-S_{twenty two}|> R_TwoAnd S₁₁(20
S after 1H of 1)₁₂Perform motion detection for (200)
A motion detection signal 209 after 1H. The signals 208 and 209 are output to the OR circuit 211.
Is ORed, and a motion detection signal 210 is obtained. Figure 1 is above
In an embodiment based on the principle of the present invention described above, a portion indicated by a broken line
The impulse noise detector is the same as in FIG.otherPart
5 is a motion detection unit added according to the present invention. FIG. 2 shows another embodiment of the present invention, which is different from FIG.
Points are indicated by dashed lines to save frame memory.
The configuration of the pulse noise detector is partially changed. In FIG. 1, the motion detecting section includes a frame.
Memory 101, adder 115, absolute value circuit 104, ratio
Comparison circuit 106, line memories 109-1 and 109-2,
The circuit 102 includes a NOR circuit 111 and an AND circuit 113.
Frame delayed luminance signal, 105 is a frame difference signal, 10
7 is a threshold signal for motion detection, 108 is motion detection after 1H
Signal, 110 is 1H previous motion detection signal, 112 is motion detection
Signal. In the motion detecting section, the frame memory
101, the adder 115 and the absolute value detection circuit 104
It constitutes a frame difference component extracting means. Further, the comparison circuit 1
06, line memories 109-1 and 109-2 and Noah times
The path 111 constitutes the motion detecting means, and
113 constitutes the impulse noise determination means. The impulse noise detecting section is
In-memory 116-1, 116-2, 116-3, addition
Units 117-1 and 117-2, comparison circuits 118-1 and 11
8-2, 118-3, absolute value circuits 119-1, 119-
2 and an AND circuit 121, the noise detection determination
Means, and the noise detection and removal means comprises
Since the configuration is the same as that of the interpolation units (Y) and (C), illustration is omitted.
Was. In the embodiment of FIG. 2, the same reference numerals as those in FIG.
1 or a similar circuit, and the configuration of the motion detection unit is the same.
However, the impulse noise detection unit is
2. Addition circuit 123, absolute value circuit 124, and comparison circuit 1
25, which is almost the same as that of FIG.
Note that the impulse noise detection unit in FIG.
Etc. are shared with the motion detection unit, and the substantial functions are the same as those in FIG.
The same. Next, the operation of the motion detecting section of the embodiment of FIG. 1 will be described.
I will tell. Luminance signal (Y) from the Y / C separation circuit 14
100 is input to the frame memory 101 and the adder 115
This frame memory delays one frame period.
The difference from the calculated frame delay luminance signal 102 is calculated, and
The frame difference signal 105 is extracted via the logarithmic circuit 104.
It is. The frame difference signal 105 is supplied to the comparison circuit 10
6 and the level is compared with the threshold value signal 107 for motion detection.
And after 1H the motion detection signal 108 is obtained according to the magnitude.
And via the line memories 109-1 and 109-2.
A 1H previous motion signal 110 is obtained, and both signals are converted to a NOR circuit 111.
And the motion detection signal 112 is output. The motion detection signal 112 indicates that motion has been detected.
L level when detected, and H level when not detected.
Become. In addition, the impulse noise candidate signal (the medaka noi
Signal is detected as a candidate for impulse noise.
Becomes high when the motion detection signal 112 and the
AND circuit to which impulse noise detection signal 120 is input
Impulse noise detection signal 114 output from 113
Is a candidate for impulse noise if motion is detected
Is assumed to have no impulse noise component
Is done. In this way, according to the motion detection,
Pixels are excluded from impulse noise candidates, and
False detection of impulse noise can be reduced. In addition, in FIG.
The motion detection of the embodiment is the same as that described above. [0044] As described above, the impulse of the present invention
According to the noise elimination device, the impulse noise candidate
Of the detected pixel on the upper and lower scanning lines in the same field
Pixel motion is detected, and if motion is detected,
Pixels are excluded from impulse noise candidates.
In a specific frequency region of the moving image.
Image distortion caused by erroneous detection of pulse noise is visually disturbed.
Can improve the image quality
It becomes possible.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an embodiment of the present invention. FIG. 2 is a block diagram showing another embodiment of the present invention. FIG. 3 is an explanatory diagram of a motion detection method according to the present invention. FIG. 4 is a block diagram showing a basic configuration for implementing a motion detection method according to the present invention. FIG. 5 is a block diagram showing a specific configuration of a noise reduction circuit. FIG. 6 is a waveform diagram showing medaka noise generated in a video signal. FIG. 7 is a block diagram showing a conventional noise removal circuit. FIG. 8 is a block diagram illustrating a television receiver including a noise detection circuit. FIG. 9 is a block diagram illustrating a schematic configuration of a noise reduction circuit. FIG. 10 is an explanatory diagram of a method for detecting impulse noise in a still image. FIG. 11 is a characteristic diagram illustrating a frequency region in which noise is detected. [Description of Signs] 101 Frame memory 104 Absolute value circuit 106 Comparison circuits 109-1 and 109-2 Line memory 111 NOR circuit 113 AND circuit

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) H04N 5/14-5/217 H04N 5/38-5/46 H04N 7/20

Claims (1)

(57) [Claims 1] Demodulating a received signal from a receiving antenna,
Demodulation means for obtaining a video signal including a detection signal and luminance information, detecting a noise component included in the detection signal, determining the presence or absence of an impulse noise component included in the video signal based on the amount of the noise component, A noise detection determination unit that outputs a noise determination signal; and a signal interpolation that outputs an interpolated video signal obtained by performing an interpolation process on a luminance component in the impulse noise portion when the impulse noise component is detected by the noise detection determination unit. extracting means, the first and second full <br/> frame difference component signal from the difference between the delayed video signal a video signal and said video signal by a predetermined frame delay including said luminance information, first and Second frame
The difference component signals respectively correspond to the impulse noise portions.
After one scanning line in the same frame as the video signal
And each frame difference component signal of the video signal one scanning line earlier.
A certain frame difference component extracting means, first and second frame difference component signals, and a predetermined reference signal ;
And the level of the two frame difference component signal is
When the level is higher than the level of the reference signal, the impulse noise
Exists in the same frame as the video signal corresponding to the part
Motion in video signal after one scanning line and one scanning line before
Motion detection means for obtaining a 1H post-motion detection signal and a 1H pre-motion detection signal indicating the presence or absence of motion, detecting the presence or absence of motion in the video signal based on these two detection signals, and outputting a motion detection signal according to the presence or absence of motion. An impulse noise eliminator, comprising: a control unit that controls the signal interpolation unit so that the interpolation process is not performed when a motion is detected by the motion detection unit.