JP2000022991A - Noise reducing device and recording medium where noise reducing program is recorded - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a noise reducing device which can precisely detect a quickly moving component even when an image has an abrupt change and obtain excellent reduction effect on noise based upon the moving component. SOLUTION: A subtracter 103 calculates the difference between an input video signal from a frame memory 102a and an output video signal of one frame befor a frame memory 102b. An absolute value calculator 104 calculates the absolute value of the output of the subtracter 103. A low-pass filter 105 reduces high-frequency noise components for the calculated absolute value. Consequently, a moving component of a moving picture is calculated with high precision by a field-by-field smoothing process and excellent noise reduction based upon this moving component is actualized.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital signal noise reduction apparatus, and more particularly to a digital video signal noise reduction apparatus using a frame memory and a recursive filter, and a recording medium on which a noise reduction program is recorded.

[0002]

2. Description of the Related Art In general, a noise component contained in a video signal of a television or the like is so-called triangular noise having a high frequency mixed in at the time of imaging or the like, and exists near a color subcarrier frequency.
If this noise component is not removed before A / D conversion,
It remains as a digital video signal as it is, causing image degradation.

[0003] Further, since a video signal is a signal in which information is repeated at a frame period, the correlation between frames is strong.
Noise components included in one video signal tend to have little correlation. For this reason, when the video signal is temporally averaged for each frame period, only the noise component can be reduced. On the other hand, if a video signal is temporally averaged for each frame period in a moving image, a noise component is reduced, but a so-called afterimage phenomenon occurs. Therefore, it is necessary to balance the opposite phenomena of reducing the noise component and avoiding the afterimage phenomenon according to the degree of the image movement, that is, the motion component of the video signal.

[0004] Many systems have been proposed for such a noise reduction device utilizing the frame correlation of a video signal.
l.33 No4 (1979) PP40-44 ". Further, as an example of a device that suppresses erroneous detection of a motion component due to noise and reduces a noise component according to a more accurate motion component,
"Motion detection circuit" described in JP-A-6-225178
There is.

FIG. 10 is a schematic block diagram of a noise reduction device 1000 provided with the “motion detection circuit”. This conventional apparatus 1000 smoothes an input video signal 1001 by using a two-dimensional low-pass filter 1008 for the input video signal 1001, and compares the input video signal 1001 with a video signal of a previous frame stored in a frame memory 1004. The difference is calculated by the subtractor 1006. As a result, the motion component of the video signal between frames is extracted, and the coefficient K is calculated in the motion determination circuit 1007 based on the motion component. This coefficient K is calculated by the adder 1
003 is a coefficient that determines the ratio of two signals composing the output video signal 1009, which is stored in the input video signal 1001 input to the variable attenuator 1002 and the frame memory 1004 input to the variable attenuator 1005. The video signal of the previous frame is
K) is a coefficient for combining at a ratio of K. The motion determining circuit 1007 determines the coefficient K so that the ratio of the video signal of the previous frame increases when the image is a still image or an image having a small motion component, and when the motion component is large, the ratio of the input video signal 1001 is reduced. Decide to be more. When the motion component is extremely large, the motion determination circuit 1007 determines the value of the coefficient K to be zero so that the input video signal 1001 is output as the output video signal 1009 as it is.

FIG. 11 shows an example of a characteristic curve used for calculating the coefficient K in the movement / discrimination circuit 1007. According to FIG. 11, the value of the coefficient K can be uniquely determined for the output of the subtractor 1006. Further, in the figure, Δ1 is a threshold value at which the coefficient K becomes zero. Generally, the value of K is used in the range (0 <K <1).

[0007]

However, the noise reduction apparatus 1000 using the conventional "motion detection circuit"
Since the two-dimensional low-pass filter 1008 is applied only to the input video signal 1001 in the current frame, if the output video signal in the previous frame includes a noise component, an appropriate motion component for the noise component is used. Cannot detect. Further, the input video signal and the output video signal of the previous frame are synthesized at a ratio determined based on the inappropriate motion component, and the output video signal 10 of the current frame is synthesized.
Since it is output as 09, there is a problem that noise components of an output video signal finally output cannot be appropriately reduced. This is because, for example, when the image changes abruptly from the frame before the previous frame to the previous frame, since the motion component is large, the input video signal of the previous frame is output as it is as the output video signal and is included in this output video signal. This is because the noise component is output without being reduced.

The motion component calculated from the output video signal of the previous frame and the input video signal of the current frame in which the noise component is not reduced contains an error due to the noise component. Therefore, when the motion component is determined to be too small due to this error, an afterimage phenomenon occurs due to the output video signal in which the ratio of the output video signal of the previous frame is large despite the relatively moving image. . Conversely, if the motion component is determined to be excessive, a large proportion of the input video signal is synthesized in spite of an image with almost no motion, and the noise component contained in this is output without being sufficiently reduced.

Further, when a video signal is captured in an interlaced system, one frame is composed of two fields. Therefore, particularly in the case of a moving image in which the movement is intense, the above-described conventional apparatus uses the video signal. , The reduction of the noise component included in. The reason is as follows. That is, in general, in a moving image that changes rapidly,
Although the video signals in the first field and the second field to be sampled are significantly different, in the conventional apparatus, the input video signals in both the first field and the second field constituting the current frame are smoothed by a two-dimensional low-pass filter. It will be converted. For this reason, it is impossible to accurately detect a motion component, and it is not possible to appropriately reduce a noise component with respect to an output video signal to be finally synthesized.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problem, and detects a motion component of an image with high accuracy even when the image changes abruptly between frames or fields. An object of the present invention is to provide a noise reduction device capable of suitably reducing a noise component according to a component.

[0011]

According to the present invention, there is provided an apparatus for reducing noise included in video data of a moving image represented by continuous frames, the apparatus comprising: First to store video data
A storage unit, a second storage unit that stores video data of a frame immediately before the frame to be processed, and one pixel of interest from all the pixels that constitute the frame to be processed, Difference data calculating means for calculating difference data that is the absolute value of the difference between the video data relating to a plurality of pixels around the pixel including the pixel of interest and the video data relating to the plurality of pixels in the immediately preceding frame corresponding thereto. Smoothing means for smoothing the difference data, video data relating to the pixel of interest in the frame to be processed, and video data relating to a pixel corresponding to the pixel of interest in the immediately preceding frame are obtained by the smoothing means. The synthesized video data is synthesized at a ratio determined by the size of the synthesized data, and the synthesized video data is subjected to noise related to the pixel of interest in the frame to be processed. And output means for outputting the image data after reduction.

Thus, since the difference between the video data of the frame to be processed and the video data of the immediately preceding frame is smoothed, even if the image changes abruptly, the It is possible to accurately detect a motion component.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a noise reduction device according to the present invention will be described with reference to the drawings. (Embodiment 1) FIG. 1 is a block diagram showing a configuration of a noise reduction device 100 according to Embodiment 1 of the present invention.

The present apparatus 100 is an apparatus for reducing a noise component contained in an input video signal 101, and includes frame memories 102a and 102b, a subtractor 103, an absolute value calculator 104, a low-pass filter 105, a motion decision unit 10
6, multipliers 107 and 108, adder 109, selector 1
It consists of ten. In the present apparatus 100, “signal” means “digital signal”.

The input video signal 101 is, for example, an 8-bit luminance signal for about 350,000 pixels per frame.
Digital data to be input continuously. The frame memory 102a captures the input video signal 101 related to about 350,000 pixels per frame every 1/30 second, holds the input video signal 101 for one frame while sequentially updating it in field units, and stores A video signal of nine pixels, which is the sum of the target pixel and the eight neighboring pixels, is output to the subtractor 103, and the input video signal 101 of each pixel is delayed by one frame time and output to the multiplier 107 and the selector 110. . Here, each pixel corresponding to the video signal output after being delayed by one frame time is referred to as a target pixel.

The frame memory 102b includes a selector 11
An output video signal 111 of about 350,000 pixels per frame output from 0 is captured every 1/30 second, and the output image signal 111 for one frame is held while being sequentially updated in field units. Also, 1 corresponding to the input video signal 101
The output video signal 111 before the frame is output to the subtractor 103 and the multiplier 108.

The subtractor 103 includes, for each pixel of interest, an input video signal 101 for nine pixels obtained by adding the pixel of interest and the eight neighboring pixels, which are output from the frame memory 102a;
Nine difference values with the output video signal 111 of one frame before output from the frame memory 102b corresponding to this are calculated and output to the absolute value calculator 104. In this case, the “difference value” represents a “motion component” in the input video signal 101 and the corresponding output video signal 111 one frame before.

The absolute value calculator 104 further calculates the absolute values of the nine difference values calculated by the subtractor 103. The low-pass filter 105 is, for example, a two-dimensional symmetric Gaussian filter, and calculates smoothing, that is, an average value, of the nine absolute values output from the absolute value calculator 104 in consideration of the weight coefficient.

Here, the contents of calculation in the low-pass filter 105 will be described. FIG. 2 is a diagram illustrating an example of calculation contents in the two-dimensional low-pass filter 105. In this filter, in order to reduce a noise component included in the video signal related to the target pixel 205, the target pixel 205
The average value is calculated using 3 × 3 nine pixels 201 to 209 including. The numbers shown at the lower right of each pixel are weighting factors assigned to each pixel when calculating the average value by applying this filter.

The motion determiner 106 has a low-pass filter 1
Based on the average value calculated in step 05, for each pixel of interest, it is determined whether the image in the vicinity of the pixel of interest is a “moving image” or a “static image”. Determine the value. Here, the “moving image” refers to an image having a large temporal change, and refers to an image that is not subjected to processing for reducing a noise component included in a video signal related to each pixel in the present apparatus 100. On the other hand, "static image"
The term “image” refers to an image having a small temporal change.
, An image to be subjected to processing for reducing noise components included in a video signal related to each pixel.

The "coefficient K" refers to the degree to which the output video signal one frame before in the output video signal is included in determining the output video signal 111 for each pixel of interest.
Generally, it is used in the range of 0 <K <1. Therefore, the smaller the change in the image over time, the larger the value of the coefficient K and the higher the ratio of the output video signal one frame before,
It is possible to reduce the ratio of noise components having no correlation between frames. On the other hand, if the value of the coefficient K is increased in a case where the change of the image is large in time, even if the noise component can be reduced, the output video signal one frame before has a strong influence, so-called afterimage. A phenomenon occurs.

FIG. 3 shows an example of a characteristic curve used when the motion determining unit 106 determines the coefficient K. FIG. 3 shows examples of three types of characteristic curves, and any one of the characteristic curves is selected and used. As shown in FIG. 3, the coefficient K is uniquely determined by the relationship with the output of the low-pass filter 105. Here, δ1, δ2, and δ3 are thresholds at which the value of the coefficient K becomes zero in each characteristic curve. Note that the characteristic curve 302 has a larger falling slope than the characteristic curve 301 and indicates that the characteristic curve 302 can be applied to a case where more accurate motion detection is possible. The characteristic curve 303 is
This is a characteristic curve when the value of the coefficient K is fixedly used (for example, 0.9 is used), and the value of the coefficient K is zero in a range where the output of the low-pass filter 105 is larger than the threshold value δ3.

The selector 110 controls the output video signal 111 in accordance with the result of the “moving image” or “still image” decision received from the motion decision unit 106. When the determination result is “moving image”, the input video signal 101 related to each pixel of interest output from the frame memory 102a is output as the output video signal 111 as it is. On the other hand, when the determination result is “still image”, the input video signal 101 multiplied by (1-K) by the multiplier 103 and the signal K multiplied by the multiplier 105 of the output video signal 111 one frame before are output. The combined signal is output as an output video signal 111.

Next, the operation of the present apparatus 100 will be described using specific luminance signal data. For convenience, in the present embodiment, an 8-bit luminance signal and a non-interlaced scanning method are assumed as the input video signal 101. 4A to 4C show data examples of luminance signals applied to the present embodiment and calculation results thereof.

FIG. 4A shows the luminance signals of the nine pixels 411 to 419 in the first frame and the nine pixels 421 to 419 in the second frame, which are input to the frame memory 102b and the subtractor 103 from the frame memory 102a. 4
29 shows a specific example of a luminance signal according to No. 29. In FIG. 4A, it is assumed that the input video signal is output as it is as the output video signal because the image of the first frame is an image having a large change with respect to the image of the previous frame. Furthermore, the case where the data of the luminance signal related to the pixel 415 includes the noise component corresponding to the luminance of about 6 points without being reduced is shown. FIG. 4A shows a case where the luminance change of 9 pixels in the second frame is reduced by 3 to 4 points from the first frame.

FIG. 4B shows the absolute value of the difference between the luminance signals of the nine pixels of the first and second frames shown in FIG. 4A, that is, the output value of the absolute value calculator 104. From this result, the absolute value 431 of the difference between the luminance signals at the target pixel 425 is significantly higher than the surrounding pixels, indicating that a noise component is included.

FIG. 4C is based on the contents of FIG.
The output of each of the low-pass filter 105 in the present apparatus 100 and the output of the subtractor 1006 in the conventional apparatus 1000 are shown. Note that the same value is used as the weight coefficient in the low-pass filter 105 and the two-dimensional low-pass filter 1008. As shown in FIG. 4C, the use of the present apparatus 100 can improve the 3.5-point motion component as compared with the absolute value.

As described above, according to the noise reduction apparatus according to the present embodiment, since the image changes abruptly, even if the noise signal is included in the video signal of the previous frame, the motion component can be accurately determined. Detection can be performed well, and suitable noise reduction based on this motion component becomes possible.

(Embodiment 2) In this embodiment,
The configuration of the noise reduction device is the same as that of the first embodiment. Note that an 8-bit luminance signal is used as the input video signal 101, but it is assumed that sampling is performed at an interlace of 30 frames / second and 60 fields / second.

FIG. 5 is a diagram for explaining how to select surrounding pixels for a target pixel in the present embodiment. In this embodiment, as the surrounding pixels, the latest eight pixels in the same field are used. In other words, the pixels 541 to 544, 5
Eight pixels 46 to 549 are selected. In this case, although one pixel is spatially separated vertically, 9 pixels in the same field are separated.
Since pixels are used, they are close in time.

FIG. 6 is a diagram for explaining how to select surrounding pixels for a target pixel in the conventional apparatus 1000. In the conventional device 1000, the eight most recent pixels in the first field and the second field in the same frame are used. That is, for the pixel of interest 545,
Eight pixels 544 and 546 and pixels 534 to 539 are selected. In this case, the nine pixels including the target pixel 545 are spatially closest, but since pixels of two different fields are used, a time difference of 1/60 seconds occurs.

Next, the operation of the present embodiment will be described using specific luminance signal data. FIG.
(a) and (b) show data examples of the luminance signal of each pixel used in the present embodiment and the conventional device 1000. FIG. 7 (a)
6 shows data of a luminance signal related to 15 pixels 514 to 519 and 521 to 529 in a first field and a second field in a first frame. FIG. 7B shows 15 pixels 53 of the first field and the second field in the second frame.
4 shows data of luminance signals of 4 to 539 and 541 to 549. As can be seen from FIG. 7B, in the first field and the second field of the second frame, there is a luminance change of about 20 points.

FIGS. 8A to 8C show luminance signal data corresponding to 9 pixels of 3.times.3 and the results of calculation of the luminance signals, which are applied to the present embodiment and the conventional apparatus 1000, respectively. FIG. 8A shows nine pixels 521 to 529 in the second field of the first frame applied to the present embodiment, and a target pixel 545 in the second field of the second frame.
5 shows luminance signal data for each of the nine pixels 541 to 549 including. In this case, the data of the luminance signal relating to the nine pixels in the second field of the first frame is data output from the frame memory 102b, and the data of the luminance signal relating to the nine pixels in the second field of the second frame is:
This is data output from the frame memory 102a.

FIG. 8B shows nine pixels 514 to 519 and 524 in the first frame applied to the conventional apparatus 1000.
526 and luminance signal data of nine pixels 534 to 539 and 544 to 546 including the target pixel 545 in the second frame. In this case, the data of the luminance signal corresponding to the nine pixels in the first frame is data provided from the frame memory 1004,
The luminance signal data corresponding to the nine pixels of the frame is data input as the input video signal 1001. As described above, three pixels 544 to 546 including the pixel of interest 545,
There is a luminance difference of about 20 points between the front and rear six pixels 534 to 539.

FIG. 8C shows this embodiment and the conventional device 100.
The calculation result of the luminance signal related to the target pixel 545 at 0 is shown. Of these calculation results, the output of the low-pass filter 105 in this embodiment is the
0 is the output of the subtractor 906. As can be seen from these results, in the case of the conventional apparatus 1000, if there is a sudden change in luminance between fields, the luminance signal is smoothed with the luminance signal of the previous field, so that the luminance signal is lower by about 11 points. This corresponds to the aforementioned “afterimage phenomenon”.

As described above, according to the noise reduction device of the present embodiment, even if the image changes rapidly between fields, only the motion component of the image is not affected by the field before the change. Can be detected with high accuracy.

In the present invention, a two-dimensional symmetric Gaussian filter of 3.times.3 pixels is used as the low-pass filter. However, the present invention can be implemented by using another filter of 5.times.5 pixels, for example. A one-dimensional low-pass filter may be used. Also, the present invention is different from the low-pass filter 105 of the noise reduction device 900 in FIG. 9 in place of the low-pass filter 105 of the noise reduction device 100 in FIG.
Even if it is realized by the configuration like 01, 902, the same effect as the result shown in FIG.

Finally, the noise reduction device according to the present invention
It goes without saying that it can be realized as a program executed on a general-purpose computer system. Therefore, it is possible to store and distribute the noise reduction program according to the present invention on a recording medium such as a CD-ROM.

[0039]

As is apparent from the above description, the noise reduction apparatus according to the present invention is an apparatus for reducing noise included in video data of a moving image represented by continuous frames, and is a processing target. A first storage unit for storing video data of a frame; a second storage unit for storing video data of a frame immediately before the frame to be processed; and all the pixels constituting the frame to be processed. This is the absolute value of the difference between the video data of a plurality of pixels surrounding the pixel including the pixel of interest and the corresponding video data of the plurality of pixels in the immediately preceding frame. Difference calculating means for calculating difference data, smoothing means for smoothing the difference data, and video data relating to the pixel of interest in the frame to be processed. The video data relating to the pixel corresponding to the pixel of interest in the immediately preceding frame is synthesized at a ratio determined by the size of the data obtained by the smoothing means, and the synthesized video data is combined with the pixel of interest in the frame to be processed. And output means for outputting the image data as noise-reduced video data.

Thus, the difference between the video data of the frame to be processed and the video data of the immediately preceding frame, that is, the motion component is subjected to the smoothing process. It is possible to reduce the noise component based on the motion component with higher accuracy than in the case of performing the processing.

Further, the difference calculation means is configured to sequentially select all pixels constituting the processing target frame stored in the first storage means as the target pixel and calculate difference data. Can also. This allows
Since the noise reduction processing for one target pixel is applied to all pixels, the noise reduction processing can be performed on the video signal of the entire frame.

The first and second storage means store two fields for interlacing as one frame, and the difference calculation means includes:
From one field including the pixel of interest, video data relating to a plurality of pixels around the pixel including the pixel of interest is read, and the second storage unit reads a plurality of pixels corresponding to the field from the field corresponding to the one field in the immediately preceding frame. Such video data may be read, and the difference data may be calculated from the read video data. As a result, since the smoothing is performed by the difference data based on the pixels of the corresponding same fields, even if there is a sudden change between the fields, it is not affected by the pixels of the field before the change. A motion component can be detected with high accuracy.

Further, the output means determines whether or not the video data relating to the pixel of interest has changed with respect to the video data relating to the corresponding pixel in the immediately preceding frame. A determining unit that determines by comparing the smoothed data with a predetermined threshold, and sets the ratio to 1 when it is determined that there is a change.
It may be configured to include a synthesizing unit that sets the ratio to (0−K) and sets the ratio to (1−K) to K (0 <K <1) when it is determined that there is no change. Accordingly, the motion component of the image is determined based on the smoothed difference data, and the reduction of the noise component is controlled based on the motion component. Therefore, the noise component is reduced while minimizing the afterimage phenomenon. be able to.

Further, the noise reduction device is a device for reducing noise contained in video data of a moving image represented by a continuous frame, wherein the first storage means stores video data of a frame to be processed. A second storage unit for storing video data of a frame immediately before the frame to be processed; and a pixel to be focused out of all the pixels constituting the frame to be processed, A first smoothing means for smoothing video data relating to a plurality of pixels surrounding the pixel, and a plurality of pixels corresponding to a plurality of pixels including a pixel of interest in the frame to be processed. A second smoothing unit for smoothing the video data, and a combination of the data obtained by the first smoothing unit and the data obtained by the second smoothing unit. A difference calculating unit that calculates difference data that is an absolute value of the minute, video data related to the pixel of interest in the frame to be processed, and video data related to a pixel corresponding to the pixel of interest in the immediately preceding frame. Output means for synthesizing at a ratio determined by the size of the data obtained by the calculation means, and outputting the synthesized video data as noise-reduced video data relating to the pixel of interest in the frame to be processed. You can also.

As a result, since the smoothing process is performed on the video data of the frame to be processed and the immediately preceding frame, it is possible to reduce the noise component based on the motion component with higher accuracy.

The recording medium on which the noise reduction program according to the present invention is recorded is a computer-readable recording medium on which a program for reducing noise in an image represented by continuous frames is recorded, and is sequentially stored in the memory. For the video data of the frame to be processed and the video data of the immediately preceding frame, one pixel of interest is specified from all the pixels constituting the frame to be processed, and the pixel of interest is A difference calculation step of calculating difference data that is an absolute value of a difference between video data related to a plurality of pixels surrounding the video data and video data related to a plurality of pixels in the immediately preceding frame, including the difference; A smoothing step of smoothing data; and a video data relating to the pixel of interest in the frame to be processed. And the video data relating to the pixel corresponding to the pixel of interest in the immediately preceding frame are synthesized at a ratio determined by the size of the data obtained in the smoothing step, and the synthesized video data is processed in the frame to be processed. And an output step of outputting as noise-reduced video data relating to the pixel of interest.

Thus, the computer that executes the noise reduction program performs the smoothing process on the difference between the video data of the frame to be processed and the immediately preceding frame, that is, the motion component. It is possible to reduce a noise component based on a motion component with higher accuracy as compared with the case where the smoothing process is performed on the video data itself.

[Brief description of the drawings]

FIG. 1 is a schematic block diagram of a noise reduction device 100 according to the present invention.

FIG. 2 is a diagram illustrating an example of peripheral pixels and a value of a weight coefficient of each pixel when a low-pass filter is applied;

FIG. 3 is a diagram illustrating a third embodiment of the noise reduction device 100 according to the present invention.
It is an example of the characteristic curve of the kind coefficient K.

FIG. 4A is a data example of a luminance signal applied to the first embodiment; (b) is the absolute value of the difference between the luminance signals of the first and second frames in FIG. (c) is the figure
The luminance signal of FIG.
7 shows a calculation result of a luminance signal relating to a pixel of interest when applied to the conventional device 1000 and when applied to the conventional device 1000.

FIG. 5 is an explanatory diagram regarding a method of selecting a target pixel to be calculated and peripheral pixels in a second embodiment.

FIG. 6 is a diagram illustrating a method of selecting a pixel of interest to be calculated and surrounding pixels in the conventional apparatus 1000.

FIG. 7A is a data example of a luminance signal of a first frame applied to the second embodiment and the conventional apparatus 1000;
(b) is a data example of the luminance signal of the second frame applied to the second embodiment and the conventional apparatus 1000.

FIG. 8A shows a second embodiment based on FIGS. 7A and 7B.
9 is an example of data of a luminance signal of 9 pixels applied to. (b)
7A is a data example of a luminance signal of 9 pixels applied to the conventional device 1000 based on FIGS. 7A and 7B. (c) is the figure
2 and the conventional device 1 based on FIG.
9 shows a calculation result of the luminance value of the pixel of interest at 000.

FIG. 9 is a schematic block diagram of a noise reduction device 900 according to another embodiment of the present invention.

FIG. 10 is a schematic block diagram of a conventional noise reduction device 1000.

11 is an example of a characteristic curve of a coefficient K in the conventional device 1000. FIG.

[Explanation of symbols]

100, 900, 1000 Noise reduction device 101, 1001 Input video signal 102a, 102b, 1004 Frame memory 103, 1006 Subtractor 104 Absolute value calculator 105, 901, 902 Low-pass filter 106 Motion determiner 107, 108 Multiplier 109, 1003 Adder 110 Selector 111, 1009 Output video signal 1002, 1005 Variable attenuator 1007 Motion discrimination circuit 1008 Two-dimensional low-pass filter

 ────────────────────────────────────────────────── ─── Continued on the front page F term (reference) 5C021 PA17 PA34 PA53 PA57 PA62 PA66 PA67 PA76 PA79 RB06 RB08 SA23 SA24 YA01 YC13

Claims (6)

[Claims] 1. An apparatus for reducing noise included in video data of a moving image represented by continuous frames, comprising: a first storage unit that stores video data of a frame to be processed; A second storage unit for storing video data of a frame immediately before the frame; a pixel of interest among all pixels constituting the frame to be processed, and a plurality of pixels surrounding the pixel of interest including the pixel of interest Difference calculation means for calculating difference data that is the absolute value of the difference between the video data related to the pixel and the video data related to a plurality of pixels in the immediately preceding frame, and a smoothing unit that smoothes the difference data Image data relating to the pixel of interest in the frame to be processed, and image data relating to a pixel corresponding to the pixel of interest in the immediately preceding frame. Data were synthesized at a ratio determined by the size of the data obtained by the smoothing means,
An output unit that outputs the synthesized video data as noise-reduced video data related to a pixel of interest in the frame to be processed. 2. The method according to claim 1, wherein the difference calculating unit sequentially selects all pixels constituting the processing target frame stored in the first storage unit as the target pixel and calculates difference data. The noise reduction device according to claim 1. 3. The first and second storage means store two fields for interlacing as one frame, and the difference calculation means stores the first storage means from one field including a pixel of interest in the first storage means. Reading video data relating to a plurality of pixels surrounding the pixel including the pixel of interest; reading, in the second storage means, video data relating to a plurality of pixels corresponding to the field corresponding to the one field in the immediately preceding frame; The noise reduction apparatus according to claim 1, wherein the difference data is calculated from the obtained video data. 4. The output means determines whether or not the video data of the pixel of interest has changed with respect to the video data of the corresponding pixel of the immediately preceding frame, based on a plurality of surrounding pixels including the pixel of interest. A determining unit that determines by comparing the smoothed data with a predetermined threshold value; and when it is determined that there is a change, the ratio is set to 1 to 0, and when it is determined that there is no change, Represents the ratio (1-
The noise reduction device according to any one of claims 1 to 3, further comprising: K) a combination unit that performs the combination as K (0 <K <1). 5. An apparatus for reducing noise included in video data of a moving image represented by continuous frames, comprising: first storage means for storing video data of a frame to be processed; A second storage unit for storing video data of a frame immediately before the frame; a pixel of interest among all pixels constituting the frame to be processed, and a plurality of pixels surrounding the pixel of interest including the pixel of interest A first smoothing unit for smoothing the video data related to the pixel, and smoothing the video data related to the plurality of pixels in the immediately preceding frame corresponding to the plurality of pixels including the pixel of interest in the frame to be processed. A second smoothing means; and a difference data which is an absolute value of a difference between the data obtained by the first smoothing means and the data obtained by the second smoothing means. A difference calculation unit that calculates the video data of the pixel of interest in the frame to be processed, and video data of a pixel corresponding to the pixel of interest in the immediately preceding frame. A noise reduction device that combines the video data at a ratio determined by the size and outputs the synthesized video data as noise-reduced video data related to a pixel of interest in the frame to be processed. 6. A computer-readable recording medium on which a program for reducing noise included in video data in a moving image represented by continuous frames is recorded, the computer-readable recording medium comprising: For the video data and the video data for the immediately preceding frame, one pixel of interest is specified from among all the pixels constituting the frame to be processed, and a plurality of pixels around the pixel including this pixel of interest are specified. Calculating the difference data which is the absolute value of the difference between the video data according to the above and the video data corresponding to the plurality of pixels in the immediately preceding frame, and the smoothing step for smoothing the difference data Video data relating to the pixel of interest in the frame to be processed, and a note in the immediately preceding frame. The video data related to the pixel corresponding to the pixel is synthesized at a ratio determined by the size of the data obtained in the smoothing step, and the synthesized video data is obtained after the noise reduction related to the pixel of interest in the frame to be processed. A recording medium for recording a program for causing a computer to execute an output step of outputting as video data.