JP2008153726A - Noise removing device, noise removing method, and video signal display device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain an optimum noise removing effect on any video signal by obtaining the amount of noise included in a real video signal, and also to obtain an excellent noise removing effect by reducing harmful effects such as tailing and an afterimage. <P>SOLUTION: A noise removing device has: a motion difference detecting means 50 of outputting a motion difference signal representing motion of a video signal included in a frame difference of the video signal; a frame noise extracting means 39 of extracting only a noise component from the frame difference and the motion difference signal by removing a motion part of the frame difference and outputting the noise component by frames; a noise amount converting means 38 of converting the frame noise extraction result into a noise amount and outputting it; a noise removal processing control means 40 of generating a control signal for controlling the extent of noise removal according to the noise amount; and a noise removing means of performing the noise removal processing by varying a coefficient for determining the extent of the noise removal according to the control signal. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a noise removal device, a noise removal method, and a video signal display device to which the noise removal device or method is applied for removing noise in a video signal, and in particular, noise removal is performed according to the amount of noise in the video signal. The present invention relates to a noise removing device, a noise removing method, and a video signal display device.

  In recent years, when the display screen of a television receiver or the like is enlarged and the image quality of the video signal is increased, it is included in the input video signal (also referred to as the input image signal) when the video signal is processed and displayed with high image quality. Unnecessary components, that is, noise (also referred to as noise) components are becoming conspicuous, and high signal reliability is required in video signal processing. In order to reduce the noise component contained in this video signal and obtain a high-quality video, there is a noise removal (also called noise reduction, abbreviated as NR) device, for example, correlation between frames (frame correlation) on the time axis. Many three-dimensional noise removal apparatuses have been proposed for removing noise components without).

  The three-dimensional noise removal apparatus delays the output signal after noise removal by one frame, and uses a difference between frames of the signal before this one frame and the signal of the current input frame to use a certain coefficient (cyclic coefficient). A cyclic noise removal device (called a frame cyclic noise removal device) that removes the noise component by adding or subtracting the product of the current input frame to the current input frame signal, or a plurality of input frame signals that are delayed by a frame. There is a non-cyclic type noise removing device (referred to as a non-cyclic type noise removing device) that removes noise components by multiplying a frame signal by a coefficient and filtering between signals of a plurality of frames. In such a three-dimensional noise removal apparatus, while a noise removal effect can be obtained, a tailing or an afterimage may occur in a portion with motion, and the motion obtained from the coefficient or filter coefficient based on the difference between frames or the like. Calculation is performed using the detection information, and the coefficient is reduced in a portion where there is a motion to weaken the effect of noise removal, and tailing and afterimages are reduced (see, for example, Patent Document 1).

  On the other hand, the input video signal contains a lot of noise components depending on the reception status, input status, etc., and noise is mixed when the video signal is generated. ) May be a very good video signal. When there are many noise components, noise removal processing can be achieved by performing noise removal processing with a three-dimensional noise removal device, but the effect on the still image portion is small for a screen that does not contain noise components. , Tailing and afterimages in areas with movement will stand out. Therefore, by detecting the amount of noise components included in the video signal and controlling the noise removal effect according to the amount of noise, tailing and afterimage are reduced, and the appropriate noise removal effect is achieved regardless of the image. Is designed to get.

  In the conventional noise elimination device, the noise component is extracted by extracting the noise component contained in the synchronization signal portion, blanking portion, etc. of the video signal, and the cyclic coefficient of the circulating noise is set based on the detection result of the noise amount. Thus, the electric field strength of the input signal is determined from the level of the input video signal within a detection period provided in a device for performing noise removal (see, for example, Patent Document 2) and a vertical synchronization period and a vertical equalization period of the video signal. However, there is a proposal of an apparatus (see, for example, Patent Document 3) that adaptively controls a cyclic coefficient for frame cyclic noise removal and performs noise removal.

  Proposal of a device that performs noise removal processing with improved noise removal response by detecting the amount of noise in the horizontal sync signal section separated from the input video signal and controlling the noise cyclic coefficient based on the detected noise amount (For example, refer to Patent Document 4).

  Furthermore, by measuring the amount of noise based on the difference between images with high correlation in the previous and subsequent frames and controlling the cyclic coefficient according to the measured amount of noise, an apparatus that appropriately adjusts the noise reduction effect There is also a proposal (for example, refer to Patent Document 5).

JP-A-9-81754 (FIG. 1) Japanese Patent Laid-Open No. 3-69276 (FIG. 1) JP 2000-134510 A (FIGS. 1 and 2) JP-A-9-284608 (FIG. 1) Japanese Patent Laying-Open No. 2005-229166 (FIGS. 1 and 4)

  As described above, in the conventional apparatus, the cyclic coefficient for noise is controlled in accordance with the result of motion detection based on the difference between frames, and a noise component included in a signal of a known level in a synchronization signal portion or the like. In addition, the amount of noise is detected from the difference between images having high correlation in the preceding and following frames, and the cyclic coefficient of the noise is controlled by the detected amount of noise to reduce the tailing and afterimage of the moving part. On the other hand, the difference and correlation between frames include a difference due to “motion” of the video signal between the previous and subsequent frames and a “noise” component in the input video signal. Therefore, in a video signal that contains a lot of noise components, if a video signal with a large noise component value is input, for example, when the gain of the input video signal level is increased, the difference due to noise is erroneously detected as motion. In the case of an image that cannot be detected as a noise amount and does not include a noise component, the noise removal process is performed even though there is no noise component with the motion as noise. There is a problem in that tailing and afterimages in parts are conspicuous and appropriate noise removal cannot be performed.

  Furthermore, the noise component included in the signal of a known level in the synchronization signal portion or the like is a noise component outside the effective period, and the noise component limited to the portion without motion is detected, so that the original effective video There is a problem in that the amount of noise included in the signal cannot be detected, and good noise removal that appropriately detects the amount of noise cannot be performed.

  Therefore, the present invention has been made to solve the above-described problem, and its purpose is to separate the motion of the video signal and the noise component included in the difference between frames into a moving real video signal. The amount of noise included can be obtained, and by controlling the noise removal processing according to the amount of noise in the video signal, an optimal noise removal effect can be obtained for any video signal, and tailing and An object of the present invention is to provide a noise removal device, a noise removal method, and a video signal display device that can reduce the adverse effect of afterimages and obtain a good noise removal effect.

The present invention
A three-dimensional noise removing apparatus that removes a noise component having no correlation between frames from a correlation between frames of a video signal,
A frame delay means for delaying an input video signal or a video signal from which noise, which is an output of a noise removal device, has been removed;
A frame difference detection means for inputting a video signal of an input frame and a video signal delayed by the frame delay means, obtaining a signal difference between the frames, and outputting a frame difference;
Motion difference detection means for detecting a motion of a video signal included in the frame difference from the frame difference and outputting a motion difference signal indicating a difference corresponding to the motion;
A frame indicating the degree of the noise component included in the frame by extracting only the noise component in the video signal included in the frame difference by removing the motion difference in the frame difference from the frame difference and the motion difference signal. Frame noise extraction means for outputting the noise extraction result in frame units;
A noise amount converting means for converting the frame noise extraction result into a value indicating a noise amount in a frame of the video signal, detecting the noise amount in the video signal, and outputting;
Noise removal processing control means for generating a control signal for controlling the strength of noise removal according to the amount of noise;
There is provided a noise removal apparatus comprising noise removal means for performing noise removal processing by changing a coefficient for determining noise removal strength in accordance with the control signal.

  According to the present invention, the motion of the video signal and the noise component included in the difference between frames can be separated to appropriately obtain the amount of noise included in the video signal, and the noise removal process is controlled according to the amount of noise. Therefore, it is possible to obtain an optimal noise removal effect for both video signals that contain a lot of noise components and video signals that do not contain noise components. There is an effect that a noise removal effect can be obtained.

  In addition, a video signal display device incorporating the above-described noise removal device has an effect of being able to display a high-quality video based on a video signal from which a good noise removal effect is obtained.

Embodiment 1 FIG.
FIG. 1 is a block diagram showing a configuration of a noise removal apparatus according to the first embodiment of the present invention (that is, an apparatus that can implement the noise removal method according to the first embodiment). The illustrated noise removal apparatus 1 is a three-dimensional noise removal apparatus that removes a noise component from a signal of one frame before and a signal of the current input frame, and delays the output signal after noise removal by one frame to generate the next frame. A cyclic noise removing device (referred to as a frame cyclic noise removing device) used for the above process.

  In the noise removal apparatus 1 of Embodiment 1 shown in FIG. 1, an input video signal (also referred to as an input signal, an input frame signal, or a current frame signal) Di0 is sequentially input, and a frame memory 11 and a timing signal generating means 12, a noise removal unit 20, a noise amount detection unit 30, and a noise removal process control unit 40.

  The frame memory 11 is used as a frame delay means for delaying the video signal by one frame. The timing signal generation unit 12 generates a timing signal from the synchronization signal SY in the video signal. The noise removing unit 20 performs a process of removing a noise component based on a difference between frames. The noise amount detection means 30 detects the amount of the noise component included in the input video signal. The noise removal processing control unit 40 generates a signal for controlling the effect of noise removal according to the noise amount detection result from the noise amount detection unit 30.

  The noise removing means 20 determines whether the difference between the input frame signal and the signal one frame before corresponds to “motion” (due to “motion”) or “noise” (due to “noise”). A motion degree signal indicating the degree of motion is obtained by detection (judgment), converted to a cyclic coefficient for the noise component based on this motion degree signal, and noise removal is performed. Subtracting means 21 and amplitude limiting means 22, motion detection means 23, cyclic coefficient generation means 24, multiplication means 25, and calculation means 26.

The subtracting means 21 subtracts the current frame signal Di0 from the previous frame signal Im1 to obtain a frame difference Diff between the frames of the current frame signal Di0 and the previous frame signal Im1.
The amplitude limiting unit 22 limits the amplitude within a predetermined value (for example, within ± dTh) with respect to the frame difference Diff from the subtracting unit 21, and the amplitude limited difference value Dfn in the pixel of the input signal Di0. Output as a noise component.
The motion detector 23 outputs a motion level signal mds indicating the level of motion from the difference between the input frame signal Di0 and the signal Im1 one frame before.

The cyclic coefficient generation unit 24 is further controlled by the control signal efcnt from the noise removal processing control unit 40 according to the change in the signal mds from the motion detection unit 23, and generates a cyclic coefficient Km for noise.
The multiplying unit 25 multiplies the difference Dfn from the amplitude limiting unit 22 and the cyclic coefficient Km and outputs the result as a noise cyclic amount Nd.
The arithmetic means 26 adds the noise circulation amount Nd to the input signal Di0 from the multiplication means 25 (when the sign of the noise circulation amount Nd is positive, the absolute value of the noise circulation amount Nd is added to the input signal Di0, and the noise circulation When the sign of the quantity Nd is negative, a process of subtracting the absolute value of the noise circulation quantity Nd from the input signal Di0 is performed).

The noise amount detection means 30 obtains the amount of noise in the video signal by subtracting the difference corresponding to “motion” from the difference between the input frame signal and the signal one frame before. Detection means 33, motion difference detection means 50, frame noise extraction means 39, and noise amount conversion means 38 are provided.
The frame difference detection means 33 detects the difference between one frame of the video signal.
The motion difference detection means 50 detects a motion from the inter-frame difference of the video signal and outputs a motion difference signal indicating the motion.

The frame noise extraction unit 39 includes a motion subtraction unit 35, a noise difference accumulation calculation unit 36, and a frame average accumulation value calculation unit 37 that calculates an average of noise difference accumulation values for a plurality of frames. Only the difference having the noise component is extracted, and the accumulated value of the difference in units of one frame (that is, the accumulated noise difference value) is output as the extraction result of the noise component in the frame (referred to as the frame noise extraction result).
The noise amount conversion means 38 converts the averaged noise difference cumulative value frnm as a value flnd indicating the amount of noise in the video signal, and outputs it as a noise amount.

  The frame memory 11 is also used as a frame delay means for delaying and outputting the video signal by one frame. The frame memory 11 delays the output signal Do0 (the video signal after noise removal) of the noise removing device 1 by one frame, For the input video signal Di0, the video signal Im1 of the previous frame (also referred to as the previous frame signal) Im1 is output. The frame memory 11 is called a field memory when the input video signal is a field unit signal such as an interlaced scanning (interlace) signal, and is called a field memory. The delay for one frame is a delay for two fields. The frame memory 11 outputs the pixel of the signal Im1 one frame before at the same position as the pixel of the input signal Di0 of the current frame.

  The timing signal generation unit 12 receives the vertical and horizontal synchronization signals SY in the input video signal, and generates a timing signal based on the synchronization signal SY. Here, for example, a vertical synchronization pulse signal frp indicating a break of one frame period and a frame effective area signal erct for indicating a predetermined area within the effective period in the video signal of one frame are generated. The frame effective area signal erct may be a signal indicating the entire video signal effective period, not a predetermined area in the effective period of the video signal. In addition, the period delimited by one vertical synchronization signal in the interlace signal is one field period, that is, a field unit. Here, the period delimited by one vertical synchronization signal is referred to as one frame.

  The noise removing means 20 receives the current frame signal Di0, the previous frame signal Im1, and a signal efcnt for controlling the effect of noise removal. The noise removing means 20 obtains a difference value (frame difference) Diff between the current frame signal Di0 and the previous frame signal Im1, and the difference corresponds to “motion” or “noise” from the frame difference Diff. A motion degree signal mds indicating the degree of motion is obtained and a cyclic coefficient Km for the noise component is obtained based on the motion degree signal mds, and noise removal processing is performed based on the difference Diff and the cyclic coefficient Km. The cyclic coefficient Km that determines the effect of noise removal is controlled not only by the motion level signal mds but also by a control signal efcnt from the noise amount detection result by the noise removal processing control means 40. The detailed configuration of the noise removing unit 20 will be described later.

The noise amount detection means 30 receives the current frame signal Di0, the previous frame signal Im1, and the vertical synchronization pulse signal frp and the frame effective area signal erct from the timing signal generation means 12.
The noise amount detection means 30 detects the motion in the video signal from the frame difference Dn, and removes the difference corresponding to the “motion” mixed with the noise component from the difference Dn, and calculates the noise amount in the video signal. A signal (noise amount) flnd indicating the amount of noise of the video signal is output in units of frames. The noise amount flnd is obtained by averaging a signal nmct (described later) representing the degree of the noise component of each frame over a plurality of frames, and the larger the value, the more noise components are included in the video signal, and the value is zero. The closer it is, the less the noise component.
The noise amount flnd from the noise amount detection means 30 is sent to the noise removal processing control means 40. Further, in order to use the noise amount in the external processing, the noise amount flnd can be sent to a block outside the noise removing apparatus 1 as shown in the figure.

  The configuration of the noise amount detection means 30 will be described in detail below with reference to FIG. 1 and FIGS. 2, 3, and 4.

The frame difference detection means 33 in the noise amount detection means 30 is supplied with the current frame signal Di0 and the previous frame signal Im1, and the frame difference detection means 33 is connected between one frame between the previous frame signal Im1 and the current frame signal Di0. The difference value Dn is obtained.
The difference Dn between the frames includes a “motion” component and a “noise” component in the video signal. When the frame difference Dn is 0, the frame difference Dn is completely stationary or includes a noise component. If there is a motion or noise component, the absolute value of the value increases.

As described above, the frame difference detection unit 33 includes the difference calculation unit 31, the absolute value calculation unit 32 that calculates the absolute value of the interframe difference, and the difference sensitivity conversion unit 34 that nonlinearly converts the difference absolute value. The absolute value (difference absolute value) Dabs of one frame difference and the difference signal ndf obtained as a result of nonlinear transformation of the difference absolute value Dabs are used as one frame difference detection result, and the difference absolute value Dabs is sent to the motion difference detection means 50 as a difference signal. The ndf is output to the frame noise extraction means 39.
Instead of the difference signal ndf, the difference absolute value Dabs before the nonlinear conversion by the difference sensitivity conversion means 34 may be supplied to the frame noise extraction means 39 as it is.

  The difference calculation means 31 receives the current frame signal Di0 and the previous frame signal Im1, performs a subtraction process between the previous frame signal Im1 and the current frame signal Di0, and calculates a difference value between the current frame signal Di0 and the previous frame signal Im1. (Frame difference) Dn is obtained. The frame difference Dn includes a “motion” component and a “noise” component in the video signal. When the frame difference Dn is 0, the frame difference Dn is completely stationary or includes a noise component. If there is a motion or noise component, the absolute value of the frame difference Dn increases.

  The absolute value calculation means 32 receives the frame difference Dn from the difference calculation means 31. The absolute value calculation means 32 calculates the absolute value of the frame difference Dn and outputs the absolute value (difference absolute value) Dabs of the frame difference.

  The difference absolute value Dabs from the absolute value calculation means 32 is input to the difference sensitivity conversion means 34. In the differential sensitivity conversion means 34, for example, the offset value is subtracted from the absolute difference value Dabs, multiplied by a predetermined sensitivity magnification, and the calculation result is limited within a predetermined range (for example, a range from 0 to the upper limit value dN). As a result, nonlinear conversion is performed, and the difference signal ndf is output to the motion subtraction means 35 in the frame noise extraction means 39. This difference signal ndf includes a motion component and a noise component. (Note that after the absolute difference Dabs is multiplied by a predetermined sensitivity magnification, the offset value is subtracted from the multiplication result to limit the calculation result within a predetermined range (for example, a range from 0 to the upper limit value dN). (The same effect can be obtained by performing non-linear transformation with.)

  FIG. 2 shows an example of the relationship (input / output characteristics) between the differential absolute value Dabs (shown on the horizontal axis) as an input of the differential sensitivity conversion means 34 and the differential signal ndf (shown on the vertical axis) as an output. . In the illustrated example, when the difference absolute value Dabs is equal to or greater than the predetermined value Nm, the difference signal ndf is maintained at the upper limit value dN, and there is no noise component or a minute noise component in the range up to the offset value Nof. The difference signal ndf = 0 is output. Further, in the range where the difference absolute value Dabs is from the offset value Nof to the predetermined value Nm, the difference signal ndf increases linearly from 0 to the value dN. The difference signal ndf indicates the value of motion or noise.

  Increasing the sensitivity magnification of the difference sensitivity conversion means 34 increases the slope in the linearly increasing range (from Nof to Nm), and the frame difference absolute value Dabs is detected as a difference containing a large amount of noise components. On the other hand, by increasing the offset value Nof, the range detected as a minute noise component (or a portion without the noise component) is expanded (the upper limit of the range is increased). Therefore, the sensitivity for detecting a difference as noise is adjusted by the sensitivity magnification and the offset value.

By providing the difference sensitivity conversion means 34, the sensitivity of the frame difference when obtaining the amount of noise can be adjusted independently of the motion.
Instead of the non-linearly converted difference signal ndf, the absolute difference value Dabs may be supplied to the frame noise extraction means 39 as it is.

  The frame difference absolute value Dabs from the absolute value calculation means 32 is input to the motion difference detection means 50 in the noise amount detection means 30. The motion difference detecting means 50 extracts a difference corresponding to the motion from the difference absolute value Dabs, and outputs a motion difference signal mv indicating the motion difference value.

The motion difference detection unit 50 includes a motion sensitivity conversion unit 502 and a motion difference value generation unit 503, and the difference absolute value Dabs from the absolute value calculation unit 32 is input to the motion sensitivity conversion unit 502.
The motion sensitivity conversion means 502 performs non-linear conversion for detecting motion from the difference absolute value Dabs, for example, subtracts an offset value from the difference absolute value Dabs, multiplies by a predetermined sensitivity magnification, and calculates the calculation result to a predetermined value. Non-linear transformation is performed by limiting within a range (for example, a range from 0 to the upper limit value DMUL), and the difference signal Dm that is a difference indicating motion is sent to the motion difference value generation unit 503. (Note that after the difference absolute value Dabs is multiplied by a predetermined sensitivity magnification, the offset value is subtracted from the multiplication result, and the calculation result is limited to a predetermined range (for example, a range from 0 to the upper limit value DMUL). (The same effect can be obtained by performing non-linear transformation with.)

  FIG. 3 shows an example of the relationship (input / output characteristics) between the differential absolute value Dabs (shown on the horizontal axis) as an input of the motion sensitivity conversion means 502 and the differential signal Dm (shown on the vertical axis) as an output. ing. In the illustrated example, when the absolute difference value Dabs is equal to or greater than the predetermined value Tm, the differential signal Dm is maintained at the upper limit value DMUL, and becomes a minute noise component in the range up to the offset value Tof, and is output as the differential signal Dm = 0. Is done. Further, in the range where the difference absolute value Dabs is from the offset value Tof to the predetermined value Tm, the difference signal Dm increases linearly from 0 to the upper limit value DMUL. The difference signal Dm indicates the value of motion or noise.

  Increasing the sensitivity magnification of the motion sensitivity conversion means 502 increases the slope in the linearly increasing range (from Tof to Tm), making it easier to detect the frame difference absolute value Dabs as motion (the motion is reduced). On the other hand, by increasing the offset value Tof, the range detected as a minute noise component (or a stationary portion) is expanded (the upper limit of the range is increased). Therefore, the sensitivity of motion detection is adjusted by the sensitivity magnification and the offset value.

  By providing the motion sensitivity conversion means 502, the sensitivity of the frame difference for detecting motion can be adjusted independently of the difference sensitivity conversion means 34.

  The difference signal Dm from the motion sensitivity conversion unit 502 is input to the motion difference value generation unit 503. The motion difference value generation unit 503 determines whether the motion of the difference signal, such as isolated point removal, is noise, thereby converting the nonlinearly converted difference signal into a motion difference value (hereinafter referred to as “motion amount”, “motion difference”). It is converted into a motion difference indicating that it is output. More specifically, the motion difference value generation 503 is performed so that there is no isolated value by determining the majority of the difference signal and the value within a predetermined peripheral pixel range with respect to the difference signal Dm at the target pixel. By performing processing such as isolated point removal processing for removing points, noise is determined from the value of the high frequency component of the difference signal Dm, and the difference signal Dm is adjusted, the difference signal at the pixel is corrected, and the difference signal Dm A difference signal (motion difference signal mv) representing a difference value indicating motion is generated and output by unifying the variation with the values in the peripheral pixels, correcting erroneous detection of motion.

Thereby, the motion difference detection means 50 detects the motion in the video signal from the difference signal Dm, and outputs the motion difference signal mv indicating the motion in the frame difference, and detects that the difference signal Dm is due to the motion. If it is detected that the value of the motion difference signal mv is large and is a still image portion, or if the difference is not caused by motion and is detected as a noise component, zero (mv = 0, complete still pixel) ) And converted into a motion difference signal mv indicating the amount of motion.
The motion difference signal mv from the motion difference value generation unit 503 is output to the motion subtraction unit 35 in the frame noise extraction unit 39.

  To the frame noise extraction means 39, the difference signal ndf from the difference sensitivity conversion means 34, the motion difference signal mv from the motion difference value generation means 503, the vertical synchronization pulse signal frp from the timing signal generation means 12, and the frame effective area The signal erct is input.

  As described above, the frame noise extraction unit 39 includes the motion subtraction unit 35, the noise difference accumulation calculation unit 36, and the frame average accumulation value calculation unit 37. First, the motion difference signal is obtained from the difference signal ndf that is a frame difference. The difference signal nmd of only the noise component is obtained except for mv, and the value of the difference signal nmd of only the noise component is accumulated in the area of one frame to extract the extent that the noise component exists in one frame. A value (noise difference accumulated value nmct) corresponding to the degree of the noise component in one frame is obtained. Then, the noise difference accumulated value nmct is obtained for a plurality of frames, an average value (average noise difference accumulated value frnm) in the noise difference accumulated values in the plurality of frames is calculated, and output as a frame noise extraction result.

  The motion subtraction unit 35 receives the difference signal ndf from the difference sensitivity conversion unit 34 and the motion difference signal mv from the motion difference value generation unit 503. The motion subtracting means 35 subtracts the value of the motion difference signal mv from the difference signal ndf to obtain a difference signal nmd having only a noise component. If the result of subtracting the value of the motion difference signal mv, which is a difference value indicating motion, from the difference signal ndf is a negative (−) value, and if the amount of motion is greater than the difference signal ndf, is there a noise component? Alternatively, the differential signal nmd = 0 is output as a part where noise detection is invalid.

  The difference signal ndf includes a difference due to noise and a difference due to motion in the video signal. Therefore, by subtracting the motion difference signal mv indicating the degree of motion from the difference signal ndf detected from the frame difference, when the motion is detected as noise, the difference value is reduced and the motion part is represented as a noise component. The difference signal nmd as a noise component can be obtained by removing the motion included in the difference signal ndf.

  The difference signal ndf from the motion subtracting means 35, the vertical synchronizing pulse signal frp from the timing signal generating means 12, and the frame effective area signal erct are input to the noise difference accumulation calculating means 36. The noise difference accumulation calculating means 36 accumulates the value of the difference signal nmd from the motion subtracting means 35 in a predetermined area in one frame indicated by the frame effective area signal erct from the timing signal generating means 12, and the area of one frame The difference accumulated value in the frame is output as the noise difference accumulated value nmct in one frame for each frame unit indicated by the vertical synchronization pulse signal frp. The accumulated noise difference value nmct, which is the accumulation result, is based on the difference signal nmd obtained as a noise component from the frame difference, and the degree to which the noise component exists in one frame is extracted, and the video signal in one frame is extracted. This is a value indicating the degree of the noise component included, that is, the amount of noise. When there is no noise included in one frame in the noise difference cumulative value nmct (that is, when there is no noise component or there is a large movement and noise detection is invalid), the noise difference cumulative value nmct in one frame is When the value decreases and there are many noise components (that is, when the noise is large), the noise difference cumulative value nmct in one frame increases.

  Here, the video signal has a synchronizing signal period and a blanking period, and may be sent as a movie material in which a black signal portion (no-image portion) exists in the upper and lower band portions called letterboxes. In the frame effective area signal erct, for example, only the pixels in the video signal effective period excluding the image synchronization signal period and the blanking period, and the upper and lower band portions in one frame and the pixel period excluding the image blanking period are shown. The signal is generated in the timing signal generation means 12 as a signal. Therefore, the noise difference accumulation calculation means 36 outputs the difference accumulation value as the noise difference accumulation value nmct in the effective area of the video signal excluding the upper and lower band portions and the image blanking period, thereby preventing erroneous determination. it can.

  The frame average cumulative value calculation unit 37 receives the noise difference cumulative value nmct output from the noise difference cumulative calculation unit 36 and the vertical synchronization pulse signal frp from the timing signal generation unit 12. The frame average accumulated value calculating means 37 delays the noise difference accumulated value nmct in accordance with the vertical synchronization pulse signal frp, obtains a noise difference accumulated value for a plurality of frames, and is an average that is an average value of the noise difference accumulated values in a plurality of frames. The noise difference accumulated value frnm is output as a frame noise extraction result. This prevents the effects of instantaneous noise changes and video movement changes (such as movements due to scene changes and sudden changes), and prevents the average noise contained in the input video signal. Cumulative values can be obtained. The number of frames on which the average value is calculated is, for example, in the range of 32 frames, and the average value in the 32 frame period is output as the average noise difference accumulated value frnm. Note that the range of frames to be averaged is not limited to 32 frames, but may be set to 16 frames, 8 frames, or more than 32 frames.

The average noise difference cumulative value frnm, which is the frame noise extraction result from the frame average cumulative value calculation unit 37, is input to the noise amount conversion unit 38.
The noise amount conversion means 38 converts the average noise difference cumulative value frnm, which is the averaged noise difference cumulative value from the frame average cumulative value calculation means 37, as a value indicating the amount of noise in the video signal, and A signal indicating the amount of noise (noise amount flnd) is output.

The noise amount conversion means 38 performs non-linear conversion by subtracting the offset value from the average noise difference accumulated value frnm, multiplying by a predetermined sensitivity magnification, and limiting the calculation result within a predetermined range to obtain the noise amount flnd. Output. (Note that the same effect can be obtained by performing nonlinear conversion by multiplying the average noise difference accumulated value frnm by a predetermined sensitivity magnification, then subtracting the offset value from the multiplication result, and limiting the calculation result within a predetermined range. Is obtained.)
The average noise difference accumulated value frnm is a cumulative value of noise components extracted from the difference signal Dn in units of frames. As the average noise difference accumulated value frnm is larger, the amount of noise in the video signal is increased, and the sensitivity magnification is increased. Increasing the value makes it easier to obtain the average noise difference accumulated value frnm as the amount of noise.

  FIG. 4 shows an example of the relationship (input / output characteristics) between the average noise difference cumulative value frnm (shown on the horizontal axis) as an input of the noise amount conversion means 38 and the noise amount flnd (shown on the vertical axis) as an output. Is shown. In the illustrated example, when the average noise difference accumulated value frnm is equal to or larger than the predetermined value frn1, the noise amount flnd is maintained at the upper limit value flmax, and in the range up to the offset value frn0, a minute noise component is obtained, and the noise amount flnd = 0. Is output as Further, in the range where the average noise difference cumulative value frnm is between the offset value frno and the predetermined value fln1, the noise amount flnd increases linearly from 0 to the upper limit value flmax. The noise amount flnd indicates the amount of noise excluding movement in the video signal.

  By increasing the sensitivity magnification of the noise amount conversion means 38, the average noise difference cumulative value frnm can be easily detected as the noise amount. On the other hand, by increasing the offset value, a video signal without noise components (or a large movement). , Noise detection is disabled), and the range to be detected is expanded (the upper limit of the range is increased). Therefore, the sensitivity of detecting the noise amount is adjusted by the sensitivity magnification and the offset value.

  The noise amount flnd obtained by the noise amount conversion means 38 indicates that the larger the value, the more noise components are included in the video signal, and the closer the value is to 0, the smaller the noise components. The noise amount from the noise amount conversion means 38 The flnd is sent to the noise removal processing control means 40. Furthermore, since the noise amount is used in external processing, the noise amount flnd can be sent to a block outside the noise removing apparatus 1.

  A noise amount flnd as a noise amount detection result from the noise amount detection unit 30 is input to the noise removal processing control unit 40. The noise removal processing control means 40 indicates the level of the noise removal effect according to the noise amount flnd, and generates and outputs a control signal efcnt for controlling the noise removal effect.

  The control signal efcnt in the noise removal processing control means 40 is generated as a control signal for controlling a noise cyclic coefficient that determines the effect of noise removal, for example, depending on the noise amount flnd in the video signal, for example, from 0 to 8 It is assumed that the control signal efcnt is a 9-step value. That is, if the noise amount flnd is zero (flnd = 0), the control signal efcnt = 0, and if the noise amount flnd is greater than or equal to a predetermined value (first predetermined value), the control signal efcnt = 8 is converted. A control signal efcnt corresponding to the amount of noise is generated with a value between 0 and 8. When this control signal efcnt has a large noise amount flnd and is detected that there are many noise components in the video signal, the control signal efcnt = 8 (noise component “large” image), and the value of the noise amount flnd is zero. Yes (flnd = 0), when there is no noise component in the video signal, or the motion is large and noise detection is detected to be invalid, the control signal efcnt = 0 (no-noise image), between 0 and 8 This value indicates the level of the noise removal effect.

  In the above example, the setting of the control signal efcnt has been described as nine stages from 0 to 8 indicating the level of the noise removal effect, but is not limited to this, and may be a value indicating the level of the noise removal effect. For example, it may be 9 levels or less.

  Then, the control signal efcnt from the noise removal processing control means 40 is sent to the cyclic coefficient generation means 24 in the noise removal means 20.

  Next, the configuration of the above-described noise removing unit 20 will be described in detail with reference to FIGS.

  The noise removal means 20 is supplied with the current frame signal Di0, the previous frame signal Im1, and the control signal efcnt indicating the level of the noise removal effect from the noise removal processing control means 40.

  The subtracting means 21 receives the current frame signal Di0 and the previous frame signal Im1, and subtracts the current frame signal Di0 from the previous frame signal Im1, thereby obtaining a frame difference Diff between the frames of the current frame signal Di0 and the previous frame signal Im1. It is done. This frame difference Diff includes a “motion” component and a “noise” component in the video signal, similarly to the frame difference Dn by the difference calculation means 31 described above. When the frame difference Diff is 0, the frame difference Diff is completely It is a portion that is stationary or has no noise component, and if there is a motion component or noise component, the value increases.

  The amplitude limiting unit 22 limits the amplitude within a predetermined value (for example, within ± dTh) with respect to the frame difference Diff from the subtracting unit 21, and the amplitude limited difference value Dfn in the pixel of the input signal Di0. Output as a noise component.

The motion detection unit 23 receives the frame difference Diff from the subtraction unit 21. The motion detector 23 detects a motion in the video signal from the frame difference Diff between the input frame signal Di0 and the signal Im1 one frame before, and outputs a motion degree signal mds indicating the degree of motion in the frame difference. To do.
The movement degree signal mds represents the degree of movement in (MD1 + 1) stages. For example, MD1 is set to 8.

  When it is detected that the frame difference Diff is due to motion, that is, when it is determined that the pixel is a complete motion pixel, the value of mds is set to the maximum value MD1 (for example, = 8), which is a still image portion, or When it is detected that the difference is not a motion but a noise component, the value of mds is set to zero. The motion degree signal mds from the motion detector 23 is sent to the cyclic coefficient generator 24.

  The setting of the motion level signal mds from the motion detection means 23 has been described as 9 stages from 0 to 8 indicating the degree of motion, but is not limited to this, and any value indicating the degree of motion can be used. It may be 9 levels or less.

The cyclic coefficient generation unit 24 receives the motion degree signal mds from the motion detection unit 23 and the control signal efcnt indicating the level of the noise removal effect from the noise removal processing control unit 40.
The cyclic coefficient generation means 24 generates a cyclic coefficient Km for noise in accordance with the motion degree signal mds and is controlled by the control signal efcnt from the noise removal processing control means 40.

  The cyclic coefficient Km generated by the cyclic coefficient generating means 24 determines the cyclic value (degree of cyclic) of noise when performing noise removal, and is set in the range of 0 ≦ Km ≦ 1. The closer the value of the cyclic coefficient Km is to 1, the greater the cyclic value of noise for noise removal, and the higher the noise removal effect. On the other hand, when the value of the cyclic coefficient Km is zero (Km = 0), the cyclic value of noise is zero and noise removal is not performed.

  FIG. 5 is a block diagram showing an example of the configuration of the cyclic coefficient generator 24. The illustrated cyclic coefficient generation means 24 includes a coefficient calculation means 241, a coefficient restriction means 242, a slope setting means 243, a coefficient maximum value setting means 244, and a noise amount coefficient control means 245.

  First, the cyclic coefficient generation means 24 further obtains Km0 indicating the level of the noise removal effect according to the motion degree signal mds from the motion detection means 23 in order to reduce tailing and afterimage due to motion in the video signal. Further, the cyclic coefficient Km is obtained by multiplying the cyclic coefficient Km0 by the control signal efcnt.

The coefficient calculation means 241 receives the motion degree signal mds from the motion detection means 23 and the inclination setting value Ef from the inclination setting means 243. The coefficient calculating unit 241 calculates a coefficient value pkm corresponding to the cyclic coefficient based on the value of the motion degree signal mds and the inclination setting value Ef by the inclination setting unit 243. As described above, for example, when the motion degree signal mds is set in the range of 0 to 8, the coefficient value pkm is calculated by setting the value and the slope obtained by subtracting the motion degree signal mds from the maximum value 8. By multiplying the value Ef, ie
pkm = (8−mds) × Ef
Is required. Therefore, the coefficient value pkm is
0 ≦ pkm ≦ 8 × Ef
Meet. Instead of obtaining the coefficient value by multiplication, the coefficient calculating means 241 is composed of, for example, a ROM (Read Only Memory) or the like, and the coefficient value pkm is generated using the values of the inclination setting value Ef and the motion degree signal mds as addresses. Also good.

  The inclination setting unit 243 sets an inclination (multiplication value) Ef used in the multiplication process in which the coefficient calculation unit 241 calculates the coefficient value pkm from the motion degree signal mds. The change amount of the cyclic coefficient Km0 with respect to the change of the value of the motion degree signal mds is set by the inclination setting value Ef set by the inclination setting means 243.

  The coefficient value pkm calculated by the coefficient calculating unit 241 is output to the coefficient limiting unit 242. The coefficient limiting unit 242 receives the coefficient value pkm from the coefficient calculating unit 241 and the coefficient maximum value Kmax (0 <Kmax ≦ 1) from the coefficient maximum value setting unit 244. The coefficient limiting unit 242 limits the coefficient value pkm from the coefficient calculating unit 241 to the coefficient maximum value Kmax or less and outputs it as a cyclic coefficient Km0. Therefore, the cyclic coefficient Km0 is 0 ≦ Km0 ≦ Kmax, and is a cyclic coefficient generated so as to change according to the motion degree signal mds.

  The coefficient maximum value setting means 244 sets the maximum value Kmax of the cyclic coefficient Km0 for noise. The cyclic coefficient Km0 output from the coefficient limiting unit 242 is limited to the maximum value Kmax. The cyclic coefficient Km0 determines a cyclic value for noise removal, and the strength of the noise removal effect can be adjusted by the maximum value Kmax.

  Note that the cyclic coefficient Km0 corresponding to the degree of motion is obtained by the arithmetic processing by the coefficient calculation means 241 and the coefficient restriction means 242 and the restriction of the maximum value. For example, a ROM having the motion degree signal mds as an address is used. And the cyclic coefficient Km0 may be obtained. Furthermore, not only the motion level signal mds but also the degree of change of the cyclic coefficient with respect to the motion level signal mds (for example, the slope setting value Ef) and the maximum value of the cyclic coefficient are supplied as ROM addresses, and the corresponding cyclic coefficients Even if Km0 is output, the same effect as described above can be obtained.

FIG. 6 shows the relationship between the motion degree signal mds (shown on the horizontal axis) as an input to the combination of the coefficient calculating means 241 and the coefficient limiting means 242 and the cyclic coefficient Km0 (shown on the vertical axis) as an output (input). Two examples of output characteristics are shown. The case of the inclination setting Ef = Ef1 by the inclination setting means 243 is indicated by a solid line, and the case of the inclination setting Ef = Ef2 (Ef2 <Ef1) is indicated by a broken line.
When the motion level signal mds is the maximum value MD1 (for example, 8) and indicates motion (when it is determined to be a complete motion pixel), the cyclic coefficient Km = 0, the motion level is low, and a still image or noise component is used. When mds = 0 is detected as being present, Km0 = Kmax (Kmax ≦ 1) is set, and the degree of movement (value of the movement degree signal mds) increases in the range where the movement degree signal mds is larger than 0 and smaller than MD1. Accordingly, by changing the value of the cyclic coefficient Km0 to be small, a cyclic coefficient for noise considering movement can be obtained.

  Then, the cyclic coefficient Km0 limited to the maximum value Kmax or less by the coefficient limiting unit 242 is output to the noise amount coefficient control unit 245.

  The cyclic coefficient Km0 generated so as to change according to the motion degree signal mds from the coefficient limiting unit 242 and the control signal efcnt from the noise removal processing control unit 40 are input to the noise amount coefficient control unit 245. The noise amount coefficient control unit 245 multiplies the cyclic coefficient Km0 from the coefficient limiting unit 242 by, for example, a predetermined magnification that changes corresponding to the control signal efcnt, and then limits the cyclic coefficient Km0 within a predetermined range. The cyclic coefficient Km, which is the result controlled by the control signal efcnt, is output.

  Then, the cyclic coefficient Km generated by the cyclic coefficient generation means 24 so as to change according to the amount of noise in the video signal in accordance with the motion level signal mds, 0 ≦ Km ≦ Kmax, is output to the multiplication means 25. Is done.

  As described above, for example, the control signal efcnt from the noise removal processing control means 40 is a value between 0 and 8 and indicates the level of the noise removal effect in the frame, that is, the noise amount flnd is zero (flnd = 0), the control signal efcnt = 0, and in the case of a value equal to or greater than a predetermined value, the control signal efcnt = 8, which is a value between 0 and 8 depending on the amount of noise.

In such a case, the cyclic coefficient Km0 is multiplied by a magnification efcnt / 8 that changes according to the control signal efcnt, that is,
(Efcnt / 8) × Km0
After the above calculation is performed, the cyclic coefficient Km is obtained by limiting the value within a predetermined range. Since the control signal efcnt is generated in the noise removal processing control unit 40 according to the noise amount flnd detected by the noise amount detection unit 30, the obtained cyclic coefficient Km has a large noise amount flnd (first If the video signal has a large amount of noise component (detected (determined) as the maximum level), the control signal efcnt = 8 (noise component “large” image), and noise amount coefficient control The cyclic coefficient Km from the means 245 is the cyclic coefficient Km = Km0, and the noise removal effect is determined by Km0.
On the other hand, when the value of the noise amount flnd is small and the noise component in the video signal is small, the control signal efcnt is a small value, the Km is also a small value, and the degree of noise removal is reduced. Further, when the amount of noise flnd is zero (flnd = 0) and there is no noise component in the video signal, the control signal efcnt = 0 (image without noise) and Km = 0 and noise removal is not performed. Therefore, when the value of the cyclic coefficient Km changes corresponding to the control signal efcnt and the noise amount flnd is large, the cyclic coefficient Km having a relatively large value is output. The smaller the value of the noise amount flnd, the cyclic The value of the coefficient Km is made smaller.

  Note that although it has been described that the coefficient value is controlled by multiplying the magnification efcnt / 8 that changes by the control signal efcnt and limiting the value within a predetermined range, the setting of the magnification is not limited to this, and for example, the value of the control signal efcnt Multiplying by a predetermined magnification according to the control signal or limiting to within a predetermined value corresponding to the value of the control signal efcnt. Furthermore, the noise amount coefficient control means 245 may be, for example, a ROM (Read Only Memory). The cyclic coefficient Km may be generated using the control signal efcnt and the value of the cyclic coefficient Km0 as addresses.

  FIG. 7 is a block diagram showing a cyclic coefficient generation means 24 b that can be used in place of the cyclic coefficient generation means 24. Similar to the cyclic coefficient generator 24, the cyclic coefficient generator 24b shown in the figure receives the motion degree signal mds from the motion detector 23 and the control signal efcnt from the noise removal processing controller 40, and receives the motion degree signal mds and noise. A cyclic coefficient Km for noise corresponding to the control signal efcnt from the removal processing control means 40 is generated. In FIG. 7, the same or corresponding components as those in FIG. 5 are denoted by the same reference numerals.

In FIG. 7, the cyclic coefficient generation means 24b is replaced with the inclination setting means 243b and the coefficient maximum value setting means in place of the inclination setting means 243 and the coefficient maximum value setting means 244 in the configuration of the cyclic coefficient generation means 24 (see FIG. 5). 244b is provided.
The set values in the slope setting means 243b and the coefficient maximum value setting means 244b are controlled by the control signal efcnt, whereby the cyclic coefficient Km is changed according to the noise amount flnd. The configuration and operation other than those relating to the inclination setting means 243b and the coefficient maximum value setting means 244b are the same as those shown in the cyclic coefficient generation means 24, and detailed description thereof will be omitted.

  In the coefficient calculation means 241 in the cyclic coefficient generation means 24b, the coefficient value pkm corresponding to the cyclic coefficient is calculated from the value of the motion degree signal mds from the motion detection means 23 and the slope setting value Ef from the slope setting means 243b. The limiting unit 242 limits the value of the coefficient value pkm from the coefficient calculating unit 241 to the coefficient maximum value Kmax or less by the coefficient maximum value setting unit 244b and outputs it as a cyclic coefficient Km.

  The inclination setting unit 243b sets an inclination (multiplication value) Ef used in the multiplication process in which the coefficient calculation unit 241 calculates the coefficient value pkm from the motion degree signal mds. The inclination setting value Ef is a noise removal process control. It changes according to the value of the control signal efcnt from the means 40. Here, the control signal efcnt is a value between 0 and 8 and indicates the level of the noise removal effect in the frame. If the noise amount flnd is zero (flnd = 0), the control signal efcnt = 0 is greater than or equal to a predetermined value. In the case of a value, the control signal efcnt = 8 and a value between 0 and 8 is changed according to the amount of noise. In such a case, the inclination setting value Ef is set by multiplying the maximum value Efmax of the inclination setting value by, for example, a magnification efcnt / 8 that changes according to the control signal efcnt. Since the control signal efcnt is generated according to the noise amount flnd, when the noise amount flnd is large and it is detected that there are many noise components in the video signal, the slope set value Ef is increased, while the noise amount If the value of flnd is small, the slope setting value Ef is also reduced accordingly, and when the amount of noise flnd is zero, the slope setting value Ef is also zero, thereby preventing noise removal. As described above, in response to the control signal efcnt, when the value of the inclination setting value Ef changes and the noise amount flnd is large, the inclination setting value Ef is output as a large inclination, and when the value of the noise amount flnd is small, The inclination setting value Ef becomes smaller.

  The inclination setting value Ef corresponding to the amount of noise is set by the inclination setting unit 243b, and the coefficient calculation unit 241 outputs the change in the value of the motion degree signal mds and the coefficient value pkm corresponding to the amount of noise to the coefficient limiting unit 242. Is done.

  The coefficient maximum value setting means 244b sets the maximum value Kmax of the cyclic coefficient Km with respect to noise, and this maximum value Kmax changes according to the value of the control signal efcnt from the noise removal processing control means 40. That is, the control signal efcnt is a value between 0 and 8 and indicates the level of the noise removal effect in the frame. For example, a value is set so that the maximum value Kmax is changed corresponding to the control signal efcnt. Since the control signal efcnt is generated according to the noise amount flnd, the maximum value Kmax is increased when the noise amount flnd is large and it is detected that there are many noise components in the video signal. On the other hand, when the value of the noise amount flnd is small and the noise component in the video signal is small, the maximum value Kmax is set to a relatively small value so as to reduce the degree of noise removal, and the value of the noise amount flnd Is zero and there is no noise component in the video signal, the maximum value Kmax = 0 is set so that noise removal is not performed. Therefore, the value of the maximum value Kmax changes corresponding to the control signal efcnt. When the noise amount flnd is large, the maximum value Kmax is large. When the noise amount flnd is small, the maximum value Kmax is small. When the value of the noise amount flnd is zero, the maximum value Kmax is also zero. Since the cyclic coefficient Km is limited to the maximum value Kmax, the strength of the noise removal effect can be adjusted by the maximum value Kmax.

  From the above, the inclination setting value Ef and the maximum value Kmax corresponding to the amount of noise are set by the inclination setting unit 243b. Therefore, the cyclic coefficient Km is 0 ≦ Km ≦ Kmax depending on the motion level signal mds and the amount of noise in the video signal. Generated to change.

  The setting by the slope setting means 243b and the coefficient maximum value setting means 244b is constituted by, for example, a ROM using the control signal efcnt from the noise removal processing control means 40 as an address, and obtains the slope setting value Ef and the maximum value Kmax. May also be generated by arithmetic processing such as multiplication. In FIG. 7, the configuration in which the set value is controlled by the control signal efcnt has been described as being performed by both the slope setting unit 243b and the coefficient maximum value setting unit 244b. However, it is also possible to perform the configuration using only one of them. is there.

  As described above, the cyclic coefficient Km generated by the cyclic coefficient generation means 24b so as to change with 0 ≦ Km ≦ Kmax corresponding to the amount of noise in the video signal in accordance with the motion degree signal mds is the cyclic coefficient shown in FIG. Instead of the cyclic coefficient Km from the coefficient generation means 24, it is output to the multiplication means 25.

In addition to the cyclic coefficient Km from the cyclic coefficient generation means 24, the multiplication means 25 is supplied with the difference value Dfn from the amplitude limiting means 22. In the multiplying means 25, the difference value Dfn from the amplitude limiting means 22 is multiplied by the cyclic coefficient Km, and the noise cyclic amount Nd is calculated as Nd = Km × Dfn.
Ask for. The obtained noise circulation amount Nd is output to the computing means 26.

  Here, the cyclic coefficient Km is calculated in accordance with the value of the motion degree signal mds, and is further calculated by being controlled by the control signal efcnt indicating the level of the noise removal effect. Therefore, when the motion level signal mds indicates motion, and when there is no noise component in the video signal, or when motion is large and noise detection is detected as invalid, the cyclic coefficient Km = 0. In this case, the cyclic amount Nd of noise is Nd = 0, and noise removal is not performed. On the other hand, when it is detected that the degree of motion is low and it is detected that there are many noise components in the video signal, since Km = Kmax, the noise circulation amount Nd is set to the maximum value, and the noise removal effect The noise removal effect is changed step by step according to the degree of movement.

  The calculation unit 26 receives the noise circulation amount Nd and the input signal Di0 from the multiplication unit 25. The calculating means 26 adds the noise circulation amount Nd to the input signal Di0, thereby removing noise in the video signal and obtaining the output signal Do0 after noise removal. In this way, the noise cyclic amount Nd is added (the absolute value is added when the sign of the noise cyclic amount Nd is positive, and the absolute value is subtracted when the sign of the noise cyclic amount Nd is negative). Therefore, noise is removed.

  Next, in the noise removal apparatus 1 according to the first embodiment, the difference value for the motion of the video signal is detected from the difference between the frames, and the detection result of the motion difference and the frame difference are used for each frame included in the video signal. Specifically, the operation of the noise amount detection means 30 for detecting the amount of noise and the operation in the noise removal means 20 for performing noise removal by setting a coefficient for noise in accordance with a signal indicating the noise amount will be specifically described. To do.

  FIG. 8 is a flowchart for explaining noise amount detection and noise removal operations in the noise removal apparatus 1. Hereinafter, a description will be given with reference to FIG.

  The video signal Di0 input to the noise removal apparatus 1 and the previous frame signal Im1 after noise removal delayed by one frame by the frame memory 11 are input to the difference calculation means 31, and in the difference calculation means 31, the current frame signal Di0 and A frame difference Dn between the previous frame signals Im1 is obtained. Then, the absolute value calculation means 32 calculates the absolute value of the frame difference Dn to obtain the difference absolute value Dabs (step S1).

  The video signal Di0 input to the noise removal apparatus 1 and the previous frame signal Im1 after noise removal delayed by one frame by the frame memory 11 are also supplied to the subtraction means 21, and the subtraction means 21 also changes the current frame signal Di0 and the previous frame signal Di0. The frame difference Diff between the frame signals Im1 is obtained, and the amplitude limiting unit 22 limits the amplitude within a predetermined value with respect to the frame difference Diff, and outputs the amplitude-limited difference value Dfn as a noise component. To do. Then, the frame difference Diff is sent to the motion detection means 23, and the motion detection means 23 detects the motion in the video signal, and obtains a motion degree signal mds indicating the degree of motion in the frame difference (step S14).

  The difference absolute value Dabs is nonlinearly converted in the difference sensitivity conversion means 34, and the difference signal ndf after the nonlinear conversion is sent to the motion subtraction means 35 (step S2). This difference signal ndf includes a motion component and a noise component.

Also, the absolute difference value Dabs is input to the motion difference detection means 50, performs nonlinear conversion for detecting motion from the absolute difference value Dabs, corrects erroneous motion detection by isolated point removal processing, etc. A corresponding difference is extracted, and a motion difference signal mv indicating a difference in motion is output (step S3).
As a result, the motion difference detection means 50 obtains a motion difference signal mv indicating the difference value of the motion in the frame difference excluding the noise component from the difference absolute value Dabs.

Next, the motion subtraction means 35 receives the difference signal ndf from the difference sensitivity conversion means 34 and the motion difference signal mv from the motion difference detection means 50, and subtracts the value of the motion difference signal mv from the difference signal ndf. That is, the motion difference signal mv that is the difference corresponding to the “motion” mixed with the noise component is excluded from the difference signal ndf that is the frame difference, and the difference signal nmd including only the noise component is obtained (step S4).
If the result obtained by subtracting the value of the motion difference signal mv, which is a difference corresponding to motion, from the difference signal ndf is a negative (−) value, and the amount of motion is greater than the difference signal ndf, there is no noise component, or The difference signal nmd = 0 is output as a part where noise detection is invalid.

  In the difference signal ndf, there are a difference due to noise and a difference due to motion in the video signal, but when motion is detected as noise by subtracting the motion difference signal mv indicating the degree of motion, By removing the motion included in the difference signal ndf, the difference signal nmd as a noise component can be obtained.

In the noise difference accumulation calculation means 36, the value of the difference signal nmd from the motion subtraction means 35 is accumulated in a predetermined area in one frame indicated by the frame effective area signal erct, and the difference accumulation value in the area of one frame is set to 1. The noise difference cumulative value nmct in the frame is output for each frame unit (step S5).
The noise difference accumulated value nmct is based on the difference signal nmd obtained as a noise component from the frame difference, becomes a value indicating the amount of noise included in the video signal in one frame, and extracts the noise component in the frame. It becomes.

The frame average cumulative value calculation means 37 obtains the noise difference cumulative value for a plurality of frames from the noise difference cumulative value nmct for one frame period, and calculates the average noise difference cumulative value frnm in the noise difference cumulative value for the plurality of frames. (Step S6).
This prevents the effects of instantaneous noise changes and video movement changes (such as movements due to scene changes and sudden changes), and prevents the average noise contained in the input video signal. Cumulative values can be obtained.

The average noise difference accumulated value frnm from the frame average accumulated value calculating means 37 is nonlinearly converted as a value indicating the noise amount in the video signal by the noise amount converting means 38 to obtain the noise amount flnd in the frame of the video signal ( Step S7).
The average noise difference cumulative value frnm is a cumulative value of noise components in units of frames extracted from the difference signal Dn. The larger the average noise difference cumulative value frnm, the greater the amount of noise in the video signal. The means 38 adjusts the sensitivity of detection of the noise amount by the sensitivity magnification and the offset value in the conversion.

  Next, the noise amount flnd is sent to the noise removal processing control means 40. The noise removal processing control means 40 generates and outputs a control signal efcnt indicating the level of the noise removal effect and controlling the noise removal effect from the noise amount flnd (step S8).

  The control signal efcnt is generated, for example, as a control signal for controlling a noise cyclic coefficient that determines the effect of noise removal. If the noise amount flnd is zero (flnd = 0), the control signal efcnt = 0 is set to a predetermined value. In the case of the above values, the control signal efcnt is generated with a value between 0 and 8 by converting the control signal efcnt = 8. When this control signal efcnt has a large amount of noise flnd and is detected that there are many noise components in the video signal, the control signal efcnt = 8 (noise component “large” image) and the value of the amount of noise flnd is small. When the noise component in the video signal is small, the control signal efcnt is also a small value, the value of the noise amount flnd is zero, and there is no noise component in the video signal, or the movement is large and noise detection is detected to be invalid. In this case, the control signal efcnt = 0 (image without noise), and a value between 0 and 8 indicates the level of the noise removal effect.

A control signal efcnt for controlling the effect of this noise removal is sent to the cyclic coefficient generating means 24. The cyclic coefficient generation means 24 generates a cyclic coefficient Km for noise corresponding to the motion degree signal mds from the motion detection means 23, and this cyclic coefficient Km is also controlled by a control signal efcnt indicating the level of the noise removal effect. Is done.
In the cyclic coefficient generation means 24, first, a coefficient Km0 that changes in accordance with the degree of movement is calculated in accordance with the degree of movement indicated by the value of the movement degree signal mds (step S9).
Next, the cyclic coefficient Km, which is the result of controlling the coefficient Km0 with the control signal efcnt, is output by, for example, multiplying by a predetermined magnification that changes corresponding to the control signal efcnt (step S10).

  For example, the coefficient Km0 that changes in accordance with the value of the motion level signal mds is, as shown in FIG. 6, when the motion level signal mds indicates motion from the slope setting value Ef and the maximum value Kmax (Kmax ≦ 1). When the cyclic coefficient Km0 = 0, the degree of motion is low (below a predetermined value), and mds = 0 detected as a noise component, Km0 = Kmax and the value of the motion degree signal mds increases. By changing the value of the cyclic coefficient Km0 so as to be small, the cyclic coefficient Km0 for noise considering movement is obtained. The cyclic coefficient Km0 is controlled by the control signal efcnt by, for example, multiplying by a magnification efcnt / 8 that changes by the control signal efcnt, or by limiting the value by a predetermined value corresponding to the value of the control signal efcnt. The obtained cyclic coefficient Km is obtained.

  Since the control signal efcnt is generated according to the noise amount flnd detected by the noise amount detection means 30, the obtained cyclic coefficient Km is detected when the noise amount flnd is large and there are many noise components in the video signal. If it is, the cyclic coefficient Km = Km0, and the noise removal effect is determined by Km0. On the other hand, when the value of the noise amount flnd is zero (flnd = 0) and there is no noise component in the video signal, noise removal is not performed with Km = 0. Accordingly, when the value of the cyclic coefficient Km changes corresponding to the control signal efcnt and the noise amount flnd is large, the cyclic coefficient Km having a large value is output, and when the value of the noise amount flnd is small, the cyclic coefficient Km The value of becomes smaller.

The cyclic coefficient Km obtained according to the motion level signal mds indicating the level of motion and the control signal efcnt indicating the level of the noise removal effect in this way is input to the multiplying unit 25. In the multiplying means 25, the difference value Dfn from the amplitude limiting means 22 is multiplied by the cyclic coefficient Km, and the noise cyclic amount Nd is calculated as Nd = Km × Dfn.
And the obtained noise circulation amount Nd is sent to the calculating means 26 (step S11).

  In the calculating means 26, the noise circulating amount Nd from the multiplying means 25 is added to the input signal Di0 (step S12), thereby removing the noise in the video signal and obtaining the output signal Do0 after noise removal. (Step S13).

  Here, the cyclic coefficient Km is calculated according to the value of the motion degree signal mds, and is also controlled and calculated by the control signal efcnt indicating the level of the noise removal effect. Therefore, when the motion level signal mds indicates movement and when there is no noise component in the video signal, the cyclic coefficient Km = 0 is sent, so the noise cyclic amount Nd in this case is Nd = 0, Noise removal is not performed. Since the amount of noise for each frame included in the video signal is obtained, and the cyclic coefficient Km for noise is obtained according to the signal indicating the amount of noise, the noise cyclic amount for video signals with less noise components The value of Nd becomes small, and the adverse effects of tailing and afterimage can be reduced to make it inconspicuous.

  On the other hand, when it is detected that the degree of motion is low (for example, below a predetermined value) and there are many noise components in the video signal, since Km = Kmax, noise other than the motion portion is detected. The traveling amount Nd can be set to the maximum value, and the noise removal effect can be increased.

  As described above, according to the noise removing apparatus 1 of the first embodiment, in the noise amount detection unit 30, the motion difference detection unit 50 changes the motion in the frame difference with respect to the frame difference Dn obtained by the frame difference detection unit 33. A corresponding difference is extracted, a motion difference signal mv corresponding to the motion is obtained, and the frame noise extraction means 39 subtracts the value of the motion difference signal mv corresponding to the motion from the frame difference to thereby obtain a noise component. Only the difference signal nmd is obtained, and the average noise difference cumulative value frnm in units of frames is output as the frame noise extraction result from the difference signal nmd of only the noise component. Then, a noise amount flnd in the frame of the video signal is obtained from the average noise difference accumulated value frnm, a control signal efcnt for controlling the effect of noise removal is generated, and the cyclic coefficient Km is controlled based on the control signal efcnt. Thus, the noise circulation amount of noise removal is obtained.

  Therefore, the motion component and the noise component of the video signal included in the difference between frames are separated to obtain the amount of noise contained in the frame of the video signal, and the cyclic coefficient corresponding to the amount of noise in the video signal is obtained. Can do. Therefore, the cyclic coefficient can be increased for video signals that contain a lot of noise components, and the noise cyclic amount can be increased for video signals and motion parts that have less noise components. It is possible to reduce tailing and afterimages, and to eliminate the tailing and afterimages by reducing the noise circulation amount to 0 for video signals and moving parts that do not contain noise components. Accordingly, an optimum noise removal effect can be obtained for a video signal having any amount of noise, and a good noise removal effect can be obtained by reducing adverse effects such as tailing and afterimage.

  In the above embodiment, if the noise amount flnd is zero (flnd = 0), the control signal efcnt is set to zero. However, the noise amount flnd is not limited to zero, but a predetermined value (the first predetermined value). (Second predetermined value different from), the control signal efcnt may be set to zero, and the cyclic coefficient Km may be set to zero, assuming that the noise component in the video signal is small (at the minimum level).

  Although the frame memory 11 has been described as delaying the video signal by one frame, the input video signal may be an interlaced signal or a progressive scanning signal. If the signal is an interlaced signal, a delay of two fields is performed, and if the signal is a progressive signal, the same configuration can be achieved by delaying by one frame. Further, the frame delay in the frame memory 11 is not limited to the delay for one frame, and the video signal may be configured to be delayed by two frames. In this case, the calculation is performed between frames that are separated in time, A noise component having no frame correlation in the time axis is obtained, and the noise removal effect can be further increased.

  Moreover, although the case where the characteristic of the cyclic coefficient Km0 obtained according to the motion degree signal mds in the cyclic coefficient generation means 24 is as shown in FIG. 6 has been described, the present invention is not limited to this and can be obtained. In accordance with the motion level signal mds, for example, when the predetermined motion level signal mds is smaller than the predetermined value MD2, the degree of motion from the predetermined value MD2 to the maximum value of the motion level signal indicating the motion, with Km0 being a predetermined fixed value. It may be configured such that Km0 is gradually decreased as the value increases, and conversion is performed so that Km0 = 0 at the maximum of the motion degree signal.

  In the configuration in which the cyclic coefficient Km is controlled by the control signal efcnt indicating the level of the noise removal effect, the cyclic coefficient Km is set to zero (Km = 0) when the control signal efcnt is equal to or smaller than a predetermined value and the noise amount is zero or small. ) And noise removal may not be performed. That is, the same effect can be obtained if the cyclic coefficient Km changes according to the control signal efcnt obtained from the noise amount flnd and the cyclic coefficient for noise considering the noise amount of the video signal can be obtained.

  Furthermore, although each component of the noise removal apparatus 1 is described as being configured by hardware, each component may be configured to be realized by software processing in program control.

Embodiment 2. FIG.
In the noise removal apparatus 1 according to the first embodiment, the noise removal unit 20 is configured as a frame cyclic type noise removal in which the output signal Do0 after the noise removal is delayed by one frame, and the frame memory 11 performs an output after the noise removal. The frame difference Diff is obtained by delaying the signal Do0 by one frame. However, as shown in FIG. 9, the acyclic noise for obtaining the frame difference by delaying the input signal Di0 of the current frame by one frame by the frame memory 11 as it is. A removal device (referred to as a non-cyclic noise removal device) can also be configured.

  FIG. 9 is a block diagram illustrating a configuration of the non-cyclic type noise removal apparatus according to the second embodiment (that is, an apparatus that can implement the noise removal method according to the second embodiment). 9, the same or corresponding components as those in FIG. 1 are denoted by the same reference numerals.

  9, the noise removal apparatus 2 according to the second embodiment uses a noise removal means 60 instead of the noise removal means 20 in the noise removal apparatus 1 (see FIG. 1) according to the first embodiment, and the frame memory 11 receives an input signal. Di0 is delayed by one frame, and acyclic noise removal processing is performed by the input signal Di0 and the one-frame delayed signal Im1. Configurations and operations other than those relating to the noise removing means 60 are the same as those shown in the first embodiment, and a detailed description thereof will be omitted.

  Similarly to the noise removing unit 20, the noise removing unit 60 in the second embodiment obtains a movement degree signal mds indicating the degree of movement from the difference between the input frame signal and the signal Im1 one frame before, and this degree of movement. Noise removal is performed by obtaining a filter coefficient for noise removal based on the signal mds and a signal efcnt for controlling the effect of noise removal generated according to the noise amount detection result flnd. Coefficient detectors 61 and 62, a motion detector 63 for outputting a motion degree signal mds indicating the degree of motion from the difference between the input frame signal Di0 and the signal Im1 one frame before, For controlling the signal mds and the effect of noise removal generated according to the noise amount detection result in the noise removal processing control means 40. A filter coefficient generation means 64 for generating a noise removal filter coefficient kn, (1-kn) in accordance with the signal efcnt, and an addition means 65 for adding the signals obtained by coefficient multiplication in the coefficient multiplication means 61 and 62 Prepare.

  The frame memory 11 delays the input video signal Di0 by one frame and outputs the video signal Im1 one frame before the input video signal Di0. The current frame signal Di0 and the previous frame signal Im1 from the frame memory 11 are sent to the difference calculation means 31 in the noise amount detection means 30 and the noise removal means 60.

  The timing signal generation unit 12 generates timing signals erct and frp based on the synchronization signal SY, and the noise amount detection unit 30 calculates the frame difference Dn between the current frame signal Di0 and the previous frame signal Im1, ”Is detected, and the difference mv corresponding to“ motion ”is excluded from the signal ndf obtained by performing sensitivity conversion on the absolute value of the difference Dn, and the amount of noise in the video signal is obtained in units of frames. Thus, the configuration for obtaining the noise amount flnd of the video signal and the noise removal processing control means 40 indicate the level of the noise removal effect according to the noise amount flnd, and generate the control signal efcnt for controlling the noise removal effect The configuration is the same as that shown in the first embodiment, and a detailed description thereof is omitted.

  The noise removal means 60 receives the current frame signal Di0, the previous frame signal Im1, and the control signal efcnt from the noise removal processing control means 40. The noise removing unit 60 detects whether the difference corresponds to “movement” or “noise” from the frame difference between the current frame signal Di0 and the previous frame signal Im1, and generates a movement degree signal mds indicating the degree of movement. In addition, the filter coefficient kn, (1-kn) for noise removal is obtained based on the motion degree signal mds, and the noise removal filter processing is performed. The filter coefficient kn, (1-kn) is the degree of movement. In addition to the signal mds, control is also performed by a control signal efcnt from the noise removal processing control means 40.

  The input frame signal Di0 and the previous frame signal Im1 are input to the motion detection unit 63 in the noise removal unit 60. The motion detector 63 detects a motion in the video signal from the frame difference between the input frame signal Di0 and the signal Im1 one frame before, and outputs a motion level signal mds indicating the degree of motion in the frame difference. .

  The motion degree signal mds generated by the motion detection means 63 represents the degree of motion in (MD1 + 1) stages, similar to the motion degree signal mds generated by the motion detection means 23 in the first embodiment. MD1 is set to 8. When it is detected that the frame difference Diff is due to motion, that is, when it is determined that the pixel is a complete motion pixel, the value of mds is MD1 (eg, = 8), and the still image portion or the difference is due to motion. If the noise component is detected instead, the value of mds is set to zero.

  The motion degree signal mds from the motion detector 63 is sent to the filter coefficient generator 64.

  Note that the setting of the motion level signal mds from the motion detection unit 63 has been described as 9 stages from 0 to 8 indicating the degree of motion, but is not limited to this, and any value that indicates the degree of motion can be used. It may be 9 levels or less.

The filter coefficient generation means 64 receives the motion degree signal mds from the motion detection means 63 and the control signal efcnt indicating the level of the noise removal effect from the noise removal processing control means 40.
The filter coefficient generation means 64 generates filter coefficients kn and (1-kn) of the noise removal filter according to the motion degree signal mds and controlled by the control signal efcnt from the noise removal processing control means 40.

  The filter coefficients kn and (1-kn) of the noise removal filter generated and output by the filter coefficient generation means 64 are for setting the coefficient of the filter for noise removal, and are set in the range of 0 ≦ kn ≦ 1. . The coefficient kn is a coefficient for multiplying the previous frame signal Im1, and the coefficient (1-kn) is a coefficient for multiplying the input current frame signal Di0, the value of this coefficient kn is close to 1. As the noise is removed, the noise removal amount increases, and the noise removal effect increases. On the other hand, if the value of the coefficient kn is zero (kn = 0), the previous frame signal Im1 is not multiplied, the noise removal amount becomes zero, and the input signal is output as it is, and noise removal is performed. Is not done.

The filter coefficient generation means 64 obtains filter coefficients kn and (1-kn) according to the motion degree signal mds from the motion detection means 63 in order to reduce tailing and afterimage due to motion in the video signal. For example,
When the movement degree signal mds is the maximum value MD1 (for example, 8) and indicates movement, the filter coefficient kn = 0 is set.
When mds = 0, which is detected as a still image or a noise component with a low degree of motion, kn = kmax (kmax ≦ 1)
In a range where the value of the motion degree signal mds is larger than 0 and smaller than MD1, the noise is reduced in consideration of the motion by changing the coefficient kn to be smaller as the value of the motion degree signal mds becomes larger. A coefficient indicating the quantity is obtained.

  Further, the control that is detected by the control signal efcnt from the noise removal processing control means 40 when the noise amount flnd is large (greater than or equal to the first predetermined value) and the noise component in the video signal is large (at the maximum level). In the case of the signal efcnt = 8 (noise component “large” image), the control is performed so that the value of the coefficient kn becomes large, while the value of the noise amount flnd is small, the noise component in the video signal is small, and the control signal efcnt Is a small value, the coefficient kn is decreased, and the filter coefficient is controlled so that the degree of noise removal is reduced. Furthermore, when the value of the noise amount flnd is zero (flnd = 0) and the control signal efcnt = 0 (image without noise) having no noise component in the video signal, noise removal is not performed with the coefficient kn = 0. The filter coefficient is controlled as follows. Therefore, when the value of the filter coefficient kn changes in response to the control signal efcnt and the noise amount flnd is large, a large value of the coefficient kn is output. The smaller the value of the noise amount flnd, the smaller the value of the coefficient kn becomes. Made smaller.

  As described in the first embodiment, instead of setting the control signal efcnt to zero only when the noise amount flnd is zero (flnd = 0), the noise amount flnd is set to a predetermined value (the above first value). (The second predetermined value different from the predetermined value) or less, the control signal efcnt may be set to zero and the coefficient kn may be set to zero because the noise component in the video signal is small (at the minimum level). .

The filter coefficient generating means 64 is configured as shown in FIG. 10, for example.
The filter coefficient generating means 64 shown in FIG. 10 includes a coefficient calculating means 641, a coefficient limiting means 642, a slope setting means 643, a coefficient maximum value setting means 644, a noise amount coefficient control means 645, and a subtraction means 646. Is provided.

  The coefficient calculation means 641, coefficient restriction means 642, slope setting means 643, coefficient maximum value setting means 644, and noise amount coefficient control means 645 are the coefficient calculation means 241, coefficient restriction means 242, slope setting means 243 in FIG. It operates in the same manner as the coefficient maximum value setting means 244 and the noise amount coefficient control means 245. However, the signal generated by the noise amount coefficient control means 645 is the coefficient kn, and the signals output from the means 641, 642, 643, and 644 are also different correspondingly. By means 641 to 644, a coefficient kn0 is generated according to the motion degree signal mds (similar to the coefficient Km0 in the first embodiment), and the coefficient kn is generated by being limited by the control signal efcnt. The subtracting unit 646 subtracts the output kn of the noise amount coefficient control unit 645 from the signal representing “1” supplied separately, and outputs a signal representing the coefficient (1−kn).

Instead of the filter coefficient generating means 64, a filter coefficient generating means 64b shown in FIG. 11 can be used.
The filter coefficient generation means 64b shown in FIG. 11 includes a coefficient calculation means 641, a coefficient restriction means 642, a slope setting means 643b, a coefficient maximum value setting means 644b, and a subtraction means 646.

  The coefficient calculation means 641, coefficient restriction means 642, slope setting means 643b, and coefficient maximum value setting means 644b are respectively the coefficient calculation means 241, coefficient restriction means 242, slope setting means 243b, and coefficient maximum value setting means 244b of FIG. Works as well. However, the signal generated by the coefficient limiting means 642 is the coefficient kn, and the signals output from the means 641, 643b, and 644b are also different correspondingly. The means 641, 643b, 644b generates a coefficient kn (similar to the coefficient Km in the first embodiment) according to the motion degree signal mds. In this case, after the coefficient kn is obtained from the control signal efcnt from the noise removal processing control means 40, the value of the coefficient may be changed according to the motion degree signal mds, and the filter may be filtered according to the motion degree signal mds. After calculating the coefficient kn, the coefficient kn may be limited by the control signal efcnt or may be processed. The subtracting unit 646 subtracts the output kn of the noise amount coefficient control unit 645 from the signal representing “1” supplied separately, and outputs a signal representing the coefficient (1−kn).

  In the filter coefficient generation means 64, the filter coefficients kn and (1-kn) generated based on the motion degree signal mds and the noise amount flnd in the video signal are sent to the first and second coefficient multiplication means 61 and 62. Is output.

  To the first coefficient multiplication means 61, the input frame signal Di0 and the filter coefficient (1-kn) from the coefficient conversion means 64 are sent. The coefficient multiplier 61 multiplies the input frame signal Di0 by the filter coefficient (1-kn) and outputs the multiplication result to the adder 65. Further, the previous frame signal Im 1 and the filter coefficient kn from the coefficient conversion means 64 are sent to the second coefficient multiplication means 62. The coefficient multiplying unit 62 multiplies the previous frame signal Im1 by the filter coefficient kn and outputs the multiplication result to the adding unit 65.

  A signal obtained by the filter coefficient multiplication in the first and second coefficient multiplication means 61 and 62 is input to the adding means 65. The adding means 65 obtains the noise removal filter output Nd by adding the signals (the product of the input signal Di0 and the previous frame signal Im1 and the filter coefficient) from the first and second coefficient multiplying means 61 and 62, This is output as an output signal Do0 from which noise in the video signal has been removed.

  Here, the filter coefficients kn and (1−kn) are calculated according to the value of the motion degree signal mds, and are further calculated by being controlled by the control signal efcnt indicating the level of the noise removal effect. Therefore, when the motion level signal mds indicates motion, and when there is no noise component in the video signal, or when motion is large and noise detection is detected to be invalid, the coefficient kn = 0. The signal obtained by the filter coefficient multiplication in the coefficient multiplication means 62 of 2 becomes zero, and noise removal is not performed with the input signal Di0. On the other hand, if it is detected that the degree of motion is low and if there is a large amount of noise component in the video signal, the coefficient kn = 1, so that the signal of the previous frame that is added by filtering as the noise removal amount This ratio is large and is set to the maximum value, so that the noise removal effect can be increased, and the noise removal effect is changed step by step according to the degree of movement. In this way, the noise removal effect can be changed stepwise according to the degree of movement.

  The configuration and operation other than those relating to the noise removing means 60 are the same as those shown in the first embodiment, and detailed description thereof is omitted. However, in the case of the acyclic type, two frames, that is, an input frame and its previous frame The calculation is performed using the frame correlation between the two.

  In the case of a frame cyclic type such as the noise removal apparatus 1 in FIG. 1, since the video signal after noise removal is delayed by a frame, the calculation is performed not only between two frames but also over many frames. The processing of the frame cyclic noise removal device 1 and the non-cyclic noise removal device 2 is the same, but the noise removal effect is large in the case of the cyclic noise removal device 1. On the other hand, in the case of a video signal with a lot of movement, the acyclic noise removal device 11 can reduce tailing and afterimage.

  According to the noise removal apparatus 2 of the second embodiment, the noise amount detection unit 30 obtains the difference signal nmd of only the noise component except for the motion difference signal mv that is a difference corresponding to “motion”, and only the noise component is obtained. An average noise difference cumulative value frnm for each frame is calculated from the difference signal nmd, a noise amount flnd in the frame of the video signal is obtained from the average noise difference cumulative value frnm, and the noise removal effect is controlled from the noise amount flnd. Control signal efcnt is generated, and filter coefficients kn and (1-kn) for noise removal are obtained based on the control signal efcnt to remove noise from the input signal.

  Therefore, the motion and noise components of the video signal included in the difference between frames are separated, the amount of noise included in the frame of the video signal is obtained, and the coefficient of the noise removal filter corresponding to the amount of noise in the video signal is generated can do. Therefore, the amount of noise removal can be increased for video signals that contain a lot of noise components, and the amount of noise removal can be reduced for video signals and moving parts with less noise components to reduce tailing and afterimages. Furthermore, the amount of noise removal can be set to 0 for video signals and moving parts that do not contain noise components, and tailing and afterimages can be eliminated. Accordingly, an optimum noise removal effect can be obtained for a video signal having any amount of noise, and a good noise removal effect can be obtained by reducing adverse effects such as tailing and afterimage.

  The noise removal means 60 includes a noise removal filter using an input video signal and a signal one frame before. However, the number of taps of the noise removal filter, that is, the number of input signal and delay signal frames used is as follows. The number of frames may be three or more. If a plurality of frame memories 11 are used and filter coefficients can be generated, the same configuration can be achieved. If the number of frames is increased, the calculation is performed over more frames, and the noise removal effect can be further increased.

  Furthermore, although each component of the noise removal apparatus 2 has been described as being configured by hardware, each component may be configured to be realized by software processing in program control.

Embodiment 3 FIG.
In the noise removal apparatus 1 of the first embodiment, the motion of the video signal is detected from the frame difference, and the motion difference detection means 50 in the noise amount detection means 30 and the motion detection means 23 in the noise removal means 20 are separately provided. However, it is also possible to obtain a frame difference by a common means and perform motion detection.

  FIG. 12 is a block diagram showing the configuration of the noise removal device of the third embodiment (that is, a device that can implement the noise removal method of the third embodiment), and configures a cyclic noise removal device. Yes. 12, the same or corresponding components as those shown in FIG. 1 are denoted by the same reference numerals.

  In FIG. 12, the noise removal apparatus 3 according to the third embodiment uses the motion difference detection means 50 in the noise detection means 30 and the motion detection means 23 in the noise removal apparatus 1 (see FIG. 1) according to the first embodiment as a video signal. This is provided with a noise amount detecting means 70 for detecting a noise amount from the frame difference Diff from the subtracting means 21 and the motion detection result at the motion detecting means 51. . Configurations and operations other than those relating to the motion detection means 51 and the noise amount detection means 70 are the same as those shown in the first embodiment, and a detailed description thereof will be omitted.

  The noise removal apparatus 3 according to the third embodiment includes a motion detection unit 51 that detects the motion of the video signal based on the frame difference obtained by the subtraction unit 21, and noise that detects the amount of the noise component included in the input video signal. And a quantity detecting means 70.

  The frame memory 11 delays the noise-removed signal Do0 by one frame to generate the previous frame signal Im1, and the subtracting means 21 obtains the frame difference Diff between the current frame signal Di0 and the previous frame signal Im1, and the amplitude limit The means 22 and the timing signal generating means 12 for generating the timing signal based on the synchronization signal SY are the same as those shown in the first embodiment, and detailed description thereof is omitted.

  The motion detection means 51 detects a motion from the inter-frame difference Diff between the input frame signal Di0 and the signal Im1 of the previous frame, and obtains a motion degree signal mds indicating the degree of motion and a motion difference signal mv indicating the motion difference. An absolute value calculation means 511 for calculating an absolute value Dma of the difference between frames, a motion sensitivity conversion means 512 for detecting a motion by nonlinearly converting the difference absolute value Dma between frames, and a difference such as isolated point removal. It is provided with motion conversion means 513 for converting the difference signal subjected to nonlinear conversion as a value indicating the motion difference and the degree of motion by determining whether the motion by the signal is noise or not.

The noise amount detection means 70 obtains the noise amount in the video signal obtained by removing the difference corresponding to “motion” from the frame difference in units of frames. The absolute value calculation means 32 and the difference for nonlinearly converting the difference absolute value. Sensitivity conversion means 73, frame noise extraction means 79, and noise amount conversion means 38 are provided.
The frame noise extraction means 79 extracts only a difference having a noise component in one frame, and outputs the accumulated value of the difference for each frame as a noise difference accumulated value as a frame noise extraction result. Means 74, noise difference accumulation calculation means 36, and frame average accumulation value calculation means 37 are provided. Among the above, the configuration of the absolute value calculation means 32, the noise difference accumulation calculation means 36, the frame average accumulation value calculation means 37, and the noise amount conversion means 38, and the noise difference accumulation value as values indicating the noise amount in the video signal. Since the operation for conversion and output as the amount of noise is the same as in the first embodiment, the description thereof is omitted.
The same operation as the operation of detecting the difference between one frame by the frame difference detection unit 33 described with reference to FIG. 1 with respect to the first embodiment is the same as the subtraction unit 21, the absolute value calculation unit 32, and the difference sensitivity. This is performed in combination with the conversion means 73.

  First, the configuration of the motion detection means 51 will be described. The frame difference Diff from the subtracting means 21 is input to the motion detecting means 51. The motion detection means 51 extracts a difference corresponding to motion from the frame difference Diff, outputs a motion difference signal mv indicating motion, and outputs a motion degree signal mds indicating the degree of motion.

  Note that the motion difference signal mv and the motion degree signal mds output from the motion detection means 51 can be set to be the same.

  The frame difference Diff from the subtracting means 21 is input to the absolute value calculating means 511 in the motion detecting means 51. The absolute value calculating means 511 calculates the absolute value of the frame difference Diff and outputs the frame difference absolute value Dma.

  The difference absolute value Dma from the absolute value calculation means 511 is input to the motion sensitivity conversion means 512. The motion sensitivity conversion means 512 performs non-linear conversion for detecting motion from the difference absolute value Dma, for example, subtracts the offset value from the difference absolute value Dma, multiplies it by a predetermined sensitivity magnification, and outputs the calculation result to a predetermined value. (For example, this processing can be performed in the same manner as described for the difference absolute value Dabs with reference to FIG. 3), and is limited to within the range (for example, the range from 0 to the upper limit DMUL). The result is sent to the motion conversion means 513 as a motion difference signal Dm that is a difference indicating motion.

  The motion difference signal Dm from the motion sensitivity conversion unit 512 is input to the motion conversion unit 513. The motion conversion means 513 performs a majority decision on the difference signal and the value size within a predetermined peripheral pixel range with respect to the difference signal Dm at the target pixel, thereby removing the isolated point so that the isolated value disappears. The difference signal at the pixel is corrected by performing removal processing and processing such as determining noise from the value of the high frequency component of the difference signal Dm and adjusting the difference signal Dm. To compensate for misdetection of movement. Then, the corrected differential signal is used as a motion differential signal mv, and the corrected differential signal is converted into a motion degree signal mds indicating the degree of motion set in, for example, (MD1 + 1) stage (for example, MD1 = 8). To do.

  The motion difference signal mv is a value indicating the motion in the frame difference excluding the noise component mixed with the motion component. When the motion difference signal Dm is detected to include a large amount of the motion component, the value of the motion difference signal mv When the difference is detected not to include many components due to motion but to include many noise components, the value of the motion difference signal mv becomes small and is a still image portion or the difference is motion Is detected, and the motion difference signal mv is zero (mv = 0, complete still pixel).

  Further, when the difference signal after correction is larger than a predetermined value and the frame difference is detected to be due to motion, the motion degree signal mds is set to the maximum value MD1 (complete motion pixel), and the corrected difference signal is If it is small and the still image portion or the difference is not caused by motion but is detected as a noise component, it is set to zero (complete still pixel) and changes from 0 to MD1 (for example, 0) according to the change in the value of the difference signal To 8) is converted into a motion degree signal mds indicating the degree of motion.

  The setting of the movement degree signal mds has been described as nine stages from 0 to 8 indicating the degree of movement, but is not limited to this, and any value indicating the degree of movement may be used in nine stages or more. The following is also acceptable. Similarly to the motion level signal mds, the motion difference signal mv can be converted to a value that changes according to a change in the value of the difference signal, and can be set to be the same.

  The motion difference signal mv from the motion conversion unit 513 is output to the motion subtraction unit 74 in the frame noise extraction unit 39 in the noise amount detection unit 70, and the motion degree signal mds is output to the cyclic coefficient generation unit 24.

  Next, the configuration within the noise amount detection means 70 will be described with reference to FIG. The absolute value calculation means 32 in the noise amount detection means 70 calculates the absolute value of the frame difference Diff from the subtraction means 21 and outputs the frame difference absolute value Dabs. The frame difference Diff and the difference absolute value Dabs include a “motion” component and a “noise” component in the video signal. When the difference absolute value Dabs is 0, the frame difference Diff or the difference absolute value Dabs is completely stationary or noise. The difference absolute value Dabs becomes large when there is a motion or noise component in a portion that does not contain a component.

The difference absolute value Dabs from the absolute value calculating means 32 is input to the difference sensitivity converting means 73. In the difference sensitivity conversion means 73, for example, the offset value is subtracted from the difference absolute value Dabs, multiplied by a predetermined sensitivity magnification, and the calculation result is limited within a predetermined range (for example, a range from 0 to the upper limit value dN). (This process can be performed in the same way as described with respect to the difference absolute value Dabs with reference to FIG. 2), and the result is used as a difference signal ndfs as a motion subtraction unit in the frame noise extraction unit 79. Output to 74.
This difference signal ndfs includes a motion component and a noise component.
Instead of the differential signal ndfs subjected to nonlinear transformation, the absolute difference value Dabs before nonlinear transformation may be supplied to the frame noise extracting means 79 as it is.

  To the frame noise extraction means 79 in the noise amount detection means 70, the difference signal ndfs from the difference sensitivity conversion means 73, the motion difference signal mv from the motion detection means 51, and the vertical synchronization pulse signal from the timing signal generation means 12. The frp and the frame effective area signal erct are input. The frame noise extraction unit 79 includes a motion subtraction unit 74, a noise difference accumulation calculation unit 36, and a frame average accumulation value calculation unit 37. First, the motion difference signal mv is removed from the difference signal ndfs which is a frame difference, and the noise is extracted. By obtaining the difference signal nmd of only the component and accumulating the value of the difference signal nmd of only the noise component in the area of one frame, the extent to which the noise component exists in one frame is extracted. A noise difference cumulative value nmct corresponding to the degree of the noise component is obtained. Then, the noise difference accumulated value nmct is obtained for a plurality of frames, an average noise difference accumulated value frnm that is an average value of the noise difference accumulated values in the plurality of frames is calculated, and output as a frame noise extraction result.

  To the motion subtracting means 74, the difference signal ndfs from the difference sensitivity converting means 73 and the motion difference signal mv from the motion converting means 513 in the motion detecting means 51 are input. The motion subtraction unit 74 subtracts the value of the motion difference signal mv corresponding to the motion from the difference signal ndfs to obtain the difference signal nmd with only the noise component. In addition, when the value obtained by subtracting the value of the motion difference signal mv corresponding to the motion from the difference signal ndfs is a negative (−) value, and the motion difference is larger than the difference signal ndfs, there is no noise component, Alternatively, the differential signal nmd = 0 is output as a part where noise detection is invalid.

  The difference signal ndfs includes a difference due to noise and a difference due to motion in the video signal. By subtracting the motion difference signal mv indicating motion from the difference signal ndfs detected from the frame difference, the motion becomes noise. In the case where it is detected, the difference value can be reduced to correct the moving part to the value of the noise component, and the difference signal nmd as the noise component can be obtained by removing the movement included in the difference signal ndfs.

  Hereinafter, the configuration and operation of the noise difference accumulation calculation unit 36, the frame average accumulation value calculation unit 37, and the noise amount conversion unit 38 in the noise amount detection unit 70 are the same as those in the first embodiment, and the difference signal nmd containing only the noise component. The noise difference cumulative value nmct obtained by extracting the noise component in one frame is obtained, the average noise difference cumulative value frnm is calculated for each frame, and the noise amount flnd in the frame of the video signal is obtained from the average noise difference cumulative value frnm. Then, the control signal efcnt for controlling the effect of noise removal is generated. Then, according to the control signal efcnt obtained from the noise amount flnd, a cyclic coefficient Km for noise is obtained, and noise removal of the input signal is performed.

  As described above, according to the noise removal apparatus 3 of the third embodiment, the subtraction unit 21 obtains a frame difference, the motion detection unit 51 performs motion detection, and the noise amount detection unit 70 corresponds to “motion”. The difference signal nmd of only the noise component is obtained by excluding the motion difference signal mv that is the difference, the noise amount flnd in the frame of the video signal is obtained, and the control signal efcnt for controlling the effect of noise removal is generated, and this control is performed. A cyclic coefficient Km is controlled based on the signal efcnt to obtain a noise cyclic amount for noise removal.

  Therefore, the amount of noise circulation can be increased for video signals containing a lot of noise components, and the amount of noise circulation is reduced for video signals and moving parts with less noise components to reduce tailing and afterimages. In addition, for a video signal or a moving part that does not contain a noise component, the amount of noise circulation can be set to 0 to eliminate tailing and afterimage. Accordingly, an optimum noise removal effect can be obtained for a video signal having any amount of noise, and a good noise removal effect can be obtained by reducing adverse effects such as tailing and afterimage.

  Further, since the motion detection means 51 is configured to generate and output a motion difference signal mv to the noise amount detection means 70 and a motion degree signal mds to be sent to the cyclic coefficient generation means 24 by noise removal processing. The calculation of the frame difference and the motion detection process can be executed by a common means, and the circuit scale can be reduced.

  In addition, although each component of the noise removal apparatus 3 is described as being configured by hardware, each component may be configured to be realized by software processing in program control.

Embodiment 4 FIG.
The noise removal apparatus 3 according to the third embodiment is the same as the noise removal apparatus 1 according to the first embodiment, but uses the same means for obtaining the frame difference. As shown in FIG. In the non-cyclic type noise removing apparatus, as described with respect to the third embodiment, the means for obtaining the frame difference can be shared.

  FIG. 13 is a block diagram showing a configuration of the non-cyclic type noise removal apparatus according to the fourth embodiment (that is, an apparatus that can implement the noise removal method according to the fourth embodiment). In FIG. 13, the same or corresponding components as those shown in FIGS. 1, 9, and 12 are denoted by the same reference numerals.

  In FIG. 13, the noise removal apparatus 4 according to the fourth embodiment includes an amplitude limiting means 22, a cyclic coefficient generation means 24, a multiplication means 25, and an arithmetic means 26 in the noise removal apparatus 3 according to the third embodiment (see FIG. 12). Instead of the configuration of the noise removal processing by the noise, the noise by the coefficient multiplication means 61 and 62, the filter coefficient generation means 64, and the addition means 65 in the noise removal means 60 in the noise removal apparatus 2 (see FIG. 9) of the second embodiment. A configuration using a removal filter is used. The configuration and operation are the same as those described above with reference to FIGS. 12 and 9 except that the frame memory 11 delays the input signal Di0 by one frame. In the case of the non-recursive type, the calculation is performed using the frame correlation between the input frame and the previous frame.

  According to the noise removal apparatus 4 of the fourth embodiment, the subtraction unit 21 calculates the frame difference Diff, the motion detection unit 51 performs motion detection, and generates a motion difference signal mv representing a difference corresponding to “motion”. Then, in the noise amount detection means 70, the motion difference signal mv is subtracted to obtain the noise amount flnd from the difference signal nmd of only the noise component, and further, the control signal efcnt is generated from the noise amount flnd, and this control signal efcnt is generated. Based on this, filter coefficients kn, (1-kn) for noise removal are obtained, and noise removal of the input signal is performed.

  Therefore, the motion and noise components of the video signal included in the difference between frames are separated, the amount of noise included in the frame of the video signal is obtained, and the coefficient of the noise removal filter corresponding to the amount of noise in the video signal is generated can do. Therefore, the amount of noise removal can be increased for video signals that contain a lot of noise components, and the amount of noise removal can be reduced for video signals and moving parts with less noise components to reduce tailing and afterimages. Furthermore, the amount of noise removal can be set to 0 for video signals and moving parts that do not contain noise components, and tailing and afterimages can be eliminated. Accordingly, an optimum noise removal effect can be obtained for a video signal having any amount of noise, and a good noise removal effect can be obtained by reducing adverse effects such as tailing and afterimage.

  Further, since the motion detection means 51 is configured to generate and output a motion difference signal mv to the noise amount detection means 70 and a motion degree signal mds to be sent to the filter coefficient generation means 64, the calculation of the frame difference is performed. The motion detection process can be executed by a common means, and the circuit scale can be reduced.

  In addition, although each component of the noise removal apparatus 4 is described as being configured by hardware, each component may be configured to be realized by software processing in program control.

Embodiment 5. FIG.
In the noise amount detection means 30 in the noise removal apparatus 1 of the first embodiment, the motion subtraction means 35 in the frame noise extraction means 39 obtains the difference signal nmd of only the noise component, and the noise difference accumulation calculation means 36 obtains the difference signal. By accumulating the ndf value to obtain the intra-area noise difference accumulated value nmct of one frame, only the difference having a noise component in the frame is extracted and output as the noise difference accumulated value nmct, whereby the noise of the video signal is obtained. Although it is configured to obtain the quantity flnd, it is determined whether or not the pixel includes a noise component (whether the signal of each pixel includes a noise component), and the pixel accumulated value in the frame It is also possible to extract only a difference having a noise component at to obtain a noise amount.

  FIG. 14 is a block diagram showing a configuration of the noise removal apparatus according to the fifth embodiment (that is, an apparatus that can carry out the noise removal method according to the fifth embodiment), and constitutes a cyclic noise removal apparatus. Yes. In FIG. 14, the same or corresponding components as those in FIG. 1 are denoted by the same reference numerals.

  In FIG. 14, the noise removal device 5 of the fifth embodiment uses a noise amount detection unit 80 instead of the noise amount detection unit 30 in the configuration of the noise removal device 1 (see FIG. 1) of the first embodiment. It is determined whether it is a pixel with motion or a pixel containing noise components, and the amount of noise is obtained from the accumulated pixel value (pixel count value) in the frame to obtain the amount of noise in the frame unit of the video signal It is a thing. Configurations and operations other than those relating to the noise amount detection means 80 are the same as those shown in the first embodiment, and a detailed description thereof will be omitted.

The noise amount detection means 80 in the fifth embodiment is similar to the noise amount detection means 30 in the video signal in which the difference corresponding to “motion” is removed from the difference between the input frame signal and the signal one frame before. The amount of noise is obtained for each frame, and the frame difference detection means 33 and the motion difference detection means 50 are the same as those shown in FIG.
The noise amount detection unit 80 includes a frame noise extraction unit 89 and a noise amount conversion unit 87 in addition to the frame difference detection unit 33 and the motion difference detection unit 50.

The frame noise extraction unit 89 extracts only a difference having a noise component in one frame, obtains a cumulative value (count value) of pixels having only the noise component in one frame as a noise pixel cumulative value, and generates noise. The average pixel accumulated value is calculated and output as a frame noise extraction result.
The noise amount conversion means 87 converts the averaged noise pixel accumulated value as a value indicating the amount of noise in the video signal, and outputs it as a noise amount.

The frame noise extraction unit 89 includes a difference comparison unit 81, a difference pixel accumulation calculation unit 82, a motion difference comparison unit 83, a motion pixel accumulation calculation unit 84, a motion pixel subtraction unit 85, and a frame average noise pixel number calculation unit. 86.
The difference comparison means 81 binarizes the difference value ndf.

The difference pixel accumulation calculation means 82 accumulates (counts) the number of pixels determined to include a difference from the binarized difference npx in the effective area period in one frame, and accumulates the pixel accumulated value in one frame ( (Count value) is output as the difference pixel accumulated value npcnt.
The motion difference comparison unit 83 binarizes the motion difference signal mv from the motion detection unit 50.

The motion pixel accumulation calculation means 84 accumulates (counts) the number of pixels that become motion from the binarized motion difference signal mpx in the effective area period within one frame, and accumulates the pixel accumulated value (count value) within one frame. ) As a moving pixel cumulative value mvcnt.
The moving pixel subtracting unit 85 subtracts the moving pixel accumulated value mvcnt from the difference pixel accumulated value npcnt, and outputs the subtraction result as a noise pixel accumulated value nsb in one frame unit.

  The frame average noise pixel number calculating means 86 calculates the average frnsp of the noise pixel accumulated values nsb for a plurality of frames.

  In the frame memory 11 which is a frame delay means, the signal Do0 after noise removal is delayed by one frame, and the current frame signal Di0 and the previous frame signal Im1 from the frame memory 11 are compared with the difference calculation means 31 in the noise amount detection means 80. To the noise removing means 20.

  The timing signal generation unit 12 generates a timing signal based on the synchronization signal SY, and the noise removal unit 20 obtains a motion level signal from the frame difference Diff, and the motion level signal mds and the noise amount flnd by the noise removal processing control unit 40. The cyclic coefficient for the noise component is obtained in accordance with the control signal efcnt obtained from the above, and noise removal processing is performed from the difference and the cyclic coefficient. These configurations are the same as those shown in the first embodiment, and detailed description thereof is omitted.

  In FIG. 14, the current frame signal Di0 and the previous frame signal Im1 are input to the frame difference detection unit 33 in the noise amount detection unit 80, and the difference value between one frame between the previous frame signal Im1 and the current frame signal Di0. Is detected, and a difference between one frame is detected. As described above, the frame difference detection unit 33 includes the difference calculation unit 31, the absolute value calculation unit 32 that calculates the absolute value of the interframe difference, and the difference sensitivity conversion unit 34 that nonlinearly converts the difference absolute value. Thus, the difference calculation means 31 calculates the frame difference Dn between the current frame signal Di0 and the previous frame signal Im1.

  The frame difference Dn includes a “motion” component and a “noise” component in the video signal. The absolute value calculation means 32 receives the frame difference Dn from the difference calculation means 31, and the absolute value calculation means 32 calculates the absolute value of the frame difference Dn and outputs the difference absolute value Dabs of the frame difference.

The difference sensitivity conversion means 34 subtracts the offset value from the difference absolute value Dabs, multiplies it by a predetermined sensitivity magnification, and limits the calculation result within a predetermined range (for example, a range from 0 to the upper limit value dN). (This process can be performed in the same way as described for the absolute difference value Dabs with reference to FIG. 2), and the result is output as a differential signal ndf.
This difference signal ndf includes a motion component and a noise component.

  The frame difference detection means 33 outputs both the one-frame difference absolute value Dabs and the difference signal ndf obtained as a result of nonlinear transformation of the difference absolute value Dabs as a one-frame difference detection result, and the difference absolute value Dabs is detected as a motion difference. The difference signal ndf is supplied to the means 50 and to the frame noise extraction means 89. Instead of the difference signal ndf, the difference absolute value Dabs before the nonlinear conversion by the difference sensitivity conversion unit 34 may be supplied to the frame noise extraction unit 89 as it is.

  The motion difference detection means 50 in the noise amount detection means 80 receives the frame difference absolute value Dabs from the absolute value calculation means 32, and the motion difference detection means 50 indicates the motion in the frame difference from the difference absolute value Dabs. A motion difference signal mv is output. The operation so far is the same as the configuration and operation of the noise amount detection means 30 of the first embodiment.

  To the frame noise extraction means 89 in the noise amount detection means 80, the difference signal ndf from the difference sensitivity conversion means 34, the motion difference signal mv from the motion difference value generation means 503 in the motion difference detection means 50, and a timing signal The vertical synchronization pulse signal frp and the frame effective area signal erct from the generation means 12 are input.

  The frame noise extraction unit 89 determines the number of pixels (hereinafter referred to as difference pixels) having a difference that becomes a noise component or a motion component within an area of one frame from the difference signal ndf that is a frame difference, and the motion difference signal mv. To obtain the number of pixels that move within an area of one frame (hereinafter referred to as “moving pixels”), remove this number of moving pixels from the number of difference pixels, and make one frame unit of a pixel having a difference only in noise components To obtain a cumulative value (hereinafter referred to as a noise pixel cumulative value), and to extract the degree of presence of a noise component in one frame, and to obtain a value corresponding to the degree of the noise component in one frame (noise pixel cumulative The value nsb) is obtained. Then, the noise pixel accumulated value nsb is obtained for a plurality of frames, an average value (average noise pixel accumulated value frnp) of noise pixel accumulated values in a plurality of frames is calculated, and output as a frame noise extraction result.

  The difference signal ndf from the difference sensitivity converter 34 is input to the difference comparator 81 in the frame noise extractor 89. The difference comparison unit 81 compares the difference signal ndf from the difference sensitivity conversion unit 34 with a predetermined threshold, and determines whether each pixel is a pixel (difference pixel) having a difference that becomes a noise component or a motion component. Judgment is made, and the binarized value npx is output as a result. For comparison, for example, if the value of the difference signal ndf is larger than a predetermined threshold value, it is determined as a difference pixel. If the difference signal is smaller than the predetermined threshold value, the pixel has no noise component and no motion component. It is determined that the pixel is not a difference pixel, and a binarized value npx is output. The binarized value npx is, for example, a value “1” (npx = 1) when it is determined as a difference pixel, and a value “0” (npx = 0) when it is not a difference pixel. Note that the binarized value is not limited to the above, and may be any other value as long as it indicates a pixel determined to be a difference pixel by comparison.

  The difference pixel accumulation calculation means 82 includes a binarized value npx indicating whether or not it is a difference pixel from the difference comparison means 81, a vertical synchronization pulse signal frp from the timing signal generation means 12, and a frame effective area signal erct. Is entered. In accordance with the binarized value npx indicating the difference pixel from the difference comparison unit 81, the difference pixel accumulation calculation unit 82 is used in a predetermined area in one frame indicated by the frame effective area signal erct from the timing signal generation unit 12. Then, the number of pixels having a value serving as a difference pixel is accumulated (counted). That is, the case where the value determined as the difference pixel from the difference comparison means 81 is “1” (npx = 1) is accumulated over the area of the frame effective area signal erct, and the difference pixel in the area of one frame. Is output as a value corresponding to the number of difference pixels in one frame (difference accumulated pixel accumulation value npcnt) for each frame unit indicated by the vertical synchronization pulse signal frp.

  The accumulated difference pixel value npcnt, which is the accumulation result, is the number of difference pixels in one frame obtained from the value npx indicating the difference pixel, and is a value indicating the degree of presence of the frame difference in one frame, that is, 1 This value indicates the degree to which a difference that becomes a noise component or a motion component exists in the frame. In the difference pixel accumulated value npcnt, if there is no noise component in one frame, or if it is a still image, the value of the difference pixel accumulated value nmcnt in one frame becomes smaller, and there is more noise component, or motion When the component is included, the value of the difference pixel accumulated value npcnt in one frame becomes large.

  The motion difference comparison unit 83 in the frame noise extraction unit 89 receives the motion difference signal mv from the motion difference detection unit 50. The motion difference comparison unit 83 compares the motion difference signal mv from the motion difference detection unit 50 with a predetermined threshold, determines whether or not each pixel is a pixel (motion pixel) that becomes a motion, and as a result The binarized value mpx is output. In the comparison, for example, if the value of the motion difference signal mv is larger than a predetermined threshold, it is determined as a moving pixel, and if it is smaller than the other predetermined threshold, the pixel is not a pixel having a motion component (that is, still or The pixel having a noise component), that is, it is determined that it is not a motion pixel, and a binarized value mpx is output. The binarized value mpx is, for example, a value “1” (mpx = 1) when it is determined as a moving pixel, and a value “0” (mpx = 0) when it is not a moving pixel. The binarized value is not limited to the above-described value, and may be another value as long as it indicates a pixel determined to be a moving pixel by comparison.

  The motion pixel accumulation calculation means 84 includes a binarized value mpx indicating whether or not it is a motion pixel from the motion difference comparison means 83, a vertical synchronization pulse signal frp from the timing signal generation means 12, and a frame effective area signal erct. Are entered. The motion pixel accumulating calculation unit 84 corresponds to the binarized value mpx indicating the motion pixel from the motion difference comparison unit 83, and within a predetermined area in one frame indicated by the frame effective area signal erct from the timing signal generation unit 12. , The number of pixels having a value to be a moving pixel is accumulated. That is, the value “1” (mpx = 1) determined to be a motion pixel from the motion difference comparison unit 83 is accumulated over the area of the frame effective area signal erct, and the motion within the area of one frame. The accumulated pixel value mpcnt is output as a value corresponding to the number of moving pixels in one frame (moving pixel accumulated value mvcnt) for each frame unit indicated by the vertical synchronization pulse signal frp.

  The cumulative motion pixel value mvcnt, which is the cumulative result, is the number of motion pixels in one frame obtained from the value mpx indicating motion pixels, and is a value indicating the degree to which there is a difference in motion within one frame. . In the moving pixel cumulative value mvcnt, when there is only a noise component in the frame or when it is a still image, the moving pixel cumulative value mvcnt in one frame is small, and when there is a lot of motion, 1 The moving pixel cumulative value mvcnt in the frame increases.

  The motion pixel subtraction unit 85 receives the difference pixel accumulation value npcnt from the difference pixel accumulation calculation unit 82 and the movement pixel accumulation value mvcnt from the movement pixel accumulation calculation unit 84. The motion pixel subtracting means 85 subtracts a motion pixel cumulative value mvcnt indicating a cumulative value corresponding to the number of pixels having a difference corresponding to motion from the difference pixel cumulative value npcnt, thereby obtaining a pixel having a difference corresponding to “motion”. Excluding the moving pixel accumulated value mvcnt corresponding to the accumulated value, the accumulated value in one frame unit (noise pixel accumulated value nsb) of the pixel having the difference of only the noise component is obtained.

  If the result of subtracting the moving pixel cumulative value mvcnt from the differential pixel cumulative value npcnt is a negative (−) value, and the cumulative value of the moving pixel is larger than the differential pixel cumulative value npcnt, is there a noise component? Alternatively, the noise pixel cumulative value nsb = 0 is output as a part where noise detection is invalid.

  Here, the difference pixel accumulation value npcnt indicates an accumulation value of pixels in which either a difference due to noise or a difference due to motion exists in the video signal. Therefore, the motion is detected as noise by subtracting the motion pixel cumulative value mvcnt indicating the number of motion-only pixels from the difference pixel cumulative value npcnt indicating the presence of the frame difference, and is included in the frame difference, The number of pixels that are determined as difference pixels can be removed from the difference pixel accumulated value npcnt and corrected to a value indicating the number of pixels of the noise component excluding the number of moving pixels. That is, by removing the motion included in the frame difference, the noise pixel cumulative value nsb in the frame as the noise component (noise amount) can be obtained, and the degree to which the noise component exists in one frame can be extracted. The pixel cumulative value nsb is a value indicating the degree to which a noise component exists in one frame.

  The frame noise extraction means 39 in FIG. 1 subtracts the motion difference signal mv from the difference signal ndf to obtain a difference signal nmd indicating a difference value of only the noise component, and a noise difference accumulated value in a predetermined area of one frame. Although the nmct is obtained, the frame noise extraction unit 89 in FIG. 14 obtains the noise pixel accumulated value nsb as a value indicating the number of pixels of the noise component excluding the number of moving pixels in the frame. Therefore, instead of directly subtracting the difference value, the degree to which the noise component is present in one frame is obtained as the accumulated value of the noise pixels, and the degree of presence of the noise component in one frame is more appropriately obtained. be able to.

  The frame average noise pixel number calculating means 86 receives the noise pixel accumulated value nsb output from the motion pixel subtracting means 85 and the vertical synchronization pulse signal frp from the timing signal generating means 12. The frame average noise pixel number calculating means 86 delays the noise pixel accumulated value nsb in accordance with the vertical synchronization pulse signal frp, obtains noise pixel accumulated values for a plurality of frames, and is an average value among noise pixel accumulated values in a plurality of frames. An average noise pixel cumulative value frnp is calculated and output as a frame noise extraction result. This averaging prevents the effects of instantaneous noise changes and video movement changes (such as movements due to scene changes and movements that change suddenly), and the average contained in the input video signal. A noise pixel cumulative value can be obtained. The number of frames for which the average value is calculated is, for example, in the range of 32 frames, and the average value in the 32 frame period is output as the average noise pixel accumulated value frnp. Note that the range of frames to be averaged is not limited to 32 frames, but may be set to 16 frames, 8 frames, or more than 32 frames.

An average noise pixel cumulative value frnp, which is a frame noise extraction result from the frame average noise pixel number calculation unit 86 in the frame noise extraction unit 89, is input to the noise amount conversion unit 87.
The noise amount conversion unit 87 converts the average noise pixel cumulative value frnp, which is an averaged noise pixel cumulative value from the frame average noise pixel number calculation unit 86, as a value indicating the amount of noise in the video signal, and outputs the video signal. Is output as a signal indicating the amount of noise (noise amount flnd).

In the noise amount conversion means 87, as in the noise amount conversion means 38 in FIG. 1, for example, the offset value is subtracted from the average noise pixel accumulated value frnp, multiplied by a predetermined sensitivity magnification, and the result of the operation is calculated within a predetermined range. By limiting to the inside, nonlinear conversion is performed and output as a noise amount flnd.
The average noise pixel accumulated value frnp is a noise pixel accumulated value indicating the degree to which a noise component exists in a frame unit. The larger the average noise pixel accumulated value frnp is, the larger the amount of noise in the video signal is. By increasing the value, it becomes easier to obtain the average noise pixel accumulated value frnp as the noise amount.

FIG. 15 shows an example of the relationship (input / output characteristics) between the average noise pixel accumulated value frnp (shown on the horizontal axis) as an input of the noise amount conversion means and the noise amount flnd (shown on the vertical axis) as an output. Yes. In the illustrated example, when the average noise pixel accumulated value frnp is equal to or larger than the predetermined value frp1, the noise amount flnd is maintained at the upper limit value flmax, and becomes a minute noise component in the range up to the offset value frp0, and the noise amount flnd = 0. Is output as Further, in the range where the average noise pixel accumulated value frnp is between the offset value frpo and the predetermined value flp1, the noise amount flnd increases linearly from 0 to the upper limit value flmax.
The noise amount flnd indicates the amount of noise excluding movement in the video signal.

  By increasing the sensitivity magnification of the noise amount conversion means 87, the average noise pixel cumulative value frnp can be easily detected as the noise amount. On the other hand, by increasing the offset value, a video signal without noise components (or a large movement). , Noise detection is disabled), and the range to be detected is expanded (the upper limit of the range is increased). Therefore, the sensitivity of detecting the noise amount is adjusted by the sensitivity magnification and the offset value.

  The noise amount flnd obtained by the noise amount converting means 87 indicates that the larger the value, the more noise components are included in the video signal, and the closer the value is to 0, the smaller the noise components are. The flnd is sent to the noise removal processing control means 40. Furthermore, since the amount of noise is used in external processing, the amount of noise flnd can be sent to the external block of the noise removing apparatus 1.

  Hereinafter, the configuration and operation of the noise removal processing control unit 40 and the noise removal unit 20 are the same as those shown in the first embodiment, and detailed description thereof will be omitted. Based on the control signal efcnt obtained from flnd, a cyclic coefficient Km for noise is controlled, and noise is removed by a cyclic coefficient of noise corresponding to the amount of noise in the video signal.

  Next, in the frame noise extraction means 89 in the noise removing apparatus 5 of the fifth embodiment, the noise pixel accumulated value frnp is obtained from the difference pixel accumulated value npcnt and the motion pixel accumulated value mvcnt, and is thereby included in the video signal. A specific description will be given focusing on the operation of the noise amount detection means 80 for detecting the noise amount in units of frames.

  FIG. 16 is a flowchart for explaining the operation of the noise amount detecting means 80 in the noise removing apparatus 5. Hereinafter, a description will be given with reference to FIG.

  The video signal Di0 input to the noise removal device 5 and the previous frame signal Im1 after noise removal delayed by one frame by the frame memory 11 are the difference in the frame difference detection means 33 in the noise removal means 20 and the noise amount detection means 80. The difference calculation means 31 is inputted to the calculation means 31 and obtains the frame difference Dn between the current frame signal Di0 and the previous frame signal Im1. Then, the absolute value calculation means 32 calculates the absolute value of the frame difference Dn to obtain the difference absolute value Dabs (step S1).

  The subtraction means 21 in the noise removal means 20 also obtains the frame difference Diff between the current frame signal Di0 and the previous frame signal Im1, and the amplitude limiting means 22 determines a predetermined value for the frame difference Diff. Amplitude is limited within, and the amplitude-limited difference value Dfn is output as a noise component. Then, the frame difference Diff is sent to the motion detection means 23, and the motion detection means 23 detects the motion in the video signal, and obtains a motion degree signal mds indicating the degree of motion in the frame difference (step S14).

  The difference absolute value Dabs is nonlinearly converted in the difference sensitivity conversion means 34, and the difference signal ndf after the nonlinear conversion is sent to the difference comparison means 81 in the frame noise extraction means 89 (step S2). This difference signal ndf includes a motion component and a noise component.

In addition, in the motion difference detection means 50 in the noise amount detection means 80, in order to detect a motion from the difference absolute value Dabs, nonlinear conversion is performed, and an erroneous detection of motion is corrected by an isolated point removal process or the like, so that it corresponds to the motion. The motion difference signal mv indicating the motion difference value is output (step S3).
As a result, the motion difference detection means 50 obtains a motion difference signal mv indicating the motion in the frame difference excluding the noise component from the difference absolute value Dabs. The operations in steps S1, S14, S2, and S3 described above are generally the same as the operations in steps S1, S14, S2, and S3 in the first embodiment shown in FIG.

  Next, the difference comparison means 81 in the frame noise extraction means 89 in the noise amount detection means 80 compares the difference signal ndf from the difference sensitivity conversion means 34 with a predetermined threshold value, and for each pixel, a noise component or a motion component. It is determined whether or not the pixel is a difference pixel having a difference, and binarization is performed to detect the difference pixel (step S20). For comparison, for example, if the value of the difference signal ndf is larger than a predetermined threshold value, the pixel is determined to be a difference pixel. In other cases, that is, when the value is equal to or less than the predetermined threshold value, the pixel has neither a noise component nor a motion component. It is determined that the pixel is not a difference pixel.

  In the difference pixel accumulation calculation means 82, the number of pixels having a value to be a difference pixel is transferred within the area of the frame effective area signal erct based on the binarized value npx indicating whether or not it is a difference pixel from the difference comparison means 81. Accumulated and output as a differential accumulated pixel accumulated value npcnt for each frame unit (step 21). The accumulated difference pixel value npcnt, which is the accumulation result, is the number of difference pixels in one frame obtained from the value npx indicating the difference pixel, and is a value indicating the degree of presence of the frame difference in one frame, that is, 1 This value indicates the degree to which a difference that becomes a noise component or a motion component exists in the frame.

  On the other hand, the motion difference comparison means 83 in the frame noise extraction means 89 in the noise amount detection means 80 compares the motion difference signal mv from the motion detection means 50 with a predetermined threshold value, and whether or not each pixel is a motion pixel. By determining whether or not and binarizing, a moving pixel is detected (step S30). In the comparison, for example, if the value of the motion difference signal mv is larger than a predetermined threshold, it is determined as a moving pixel. In other cases, that is, when the value is equal to or smaller than the predetermined threshold, the pixel does not have a motion component. That is, it is determined that the pixel is not a motion pixel.

  The motion pixel accumulation calculation unit 84 uses the binarized value mpx indicating whether or not it is a motion pixel from the motion difference comparison unit 83, and the frame effective area signal erct indicates the number of pixels having a value as a motion pixel. Are accumulated in a predetermined area and output as a moving pixel accumulated value mvcnt for each frame unit (step S31). The cumulative motion pixel value mvcnt, which is the cumulative result, is the number of motion pixels in one frame obtained from the value mpx indicating motion pixels, and is a value indicating the degree to which there is a difference in motion within one frame. .

Next, the motion pixel subtracting unit 85 subtracts the motion pixel accumulated value mvcnt from the difference pixel accumulated value npcnt, thereby obtaining the accumulated value of pixels having a difference only in the noise component as the noise pixel accumulated value nsb (step S22). .
If the result of subtracting the moving pixel cumulative value mvcnt from the differential pixel cumulative value npcnt is a negative (−) value, and the cumulative value of the moving pixel is larger than the differential pixel cumulative value npcnt, is there a noise component? Alternatively, the noise pixel cumulative value nsb = 0 is output as a part where noise detection is invalid.

  The difference pixel accumulated value npcnt indicates the accumulated value of pixels in which either a difference due to noise or a difference due to motion exists in the video signal, and a motion pixel accumulated value indicating the number of pixels of only motion from the difference pixel accumulated value npcnt. By subtracting mvcnt, it is possible to correct to a value indicating the number of pixels of the noise component excluding the number of moving pixels. That is, the noise pixel cumulative value nsb in the frame as a noise component can be obtained excluding the motion included in the frame difference, and the noise pixel cumulative value nsb indicates the degree to which the noise component exists in one frame. Value. Rather than directly subtracting the difference value, the degree to which the noise component exists in one frame is obtained as the accumulated value of the noise pixels, and the degree of presence of the noise component in one frame can be obtained more appropriately. it can.

The frame average noise pixel number calculating means 86 obtains noise pixel cumulative values for a plurality of frames, calculates an average noise pixel cumulative value frnp in the noise pixel cumulative values in a plurality of frames, and outputs the result as a frame noise extraction result (step). S23).
This prevents the effects of instantaneous noise changes and video movement changes (such as movements due to scene changes and sudden changes), and the average noise pixels included in the input video signal. Cumulative values can be obtained.

The average noise pixel cumulative value frndp from the frame average noise pixel number calculation means 86 is nonlinearly converted as a value indicating the noise amount in the video signal by the noise amount conversion means 38 to obtain the noise amount flnd in the frame of the video signal. (Step S7).
The average noise pixel cumulative value frnp is a cumulative value of noise pixels indicating the degree to which a noise component exists in a frame unit. The larger the average noise pixel cumulative value frnp is, the larger the amount of noise in the video signal is. The means 38 adjusts the sensitivity of detection of the noise amount by the sensitivity magnification and the offset value in the conversion.

  Next, the noise amount flnd is sent to the noise removal processing control means 40, and a control signal efcnt for indicating the level of the noise removal effect and controlling the noise removal effect is generated from the noise amount flnd and output ( Step S8).

Since the following operations are the same as those in the first embodiment shown in FIG. 8, detailed description thereof is omitted. However, the cyclic coefficient generation unit 24 in the noise removal unit 20 performs the degree of motion from the motion detection unit 23. A cyclic coefficient Km for noise is generated according to the signal mds and the noise amount flnd (steps S9 and S10). A cyclic coefficient Km obtained according to the motion level signal mds indicating the level of motion and the control signal efcnt indicating the level of the noise removal effect is input to the multiplication unit 25, and the difference value Dfn from the amplitude limiting unit 22 is calculated. The cyclic coefficient Km is multiplied, and the noise cyclic amount Nd is calculated as follows: Nd = Km × Dfn
And the obtained noise circulation amount Nd is sent to the calculating means 26 (step S11).

  In the calculating means 26, the noise circulating amount Nd from the multiplying means 25 is added to the input signal Di0 (step S12), thereby removing the noise in the video signal and obtaining the output signal Do0 after noise removal. (Step S13).

  The cyclic coefficient Km is calculated according to the value of the motion degree signal mds, and further controlled and calculated by a control signal efcnt indicating the level of the noise removal effect. Therefore, when the motion level signal mds indicates movement and when there is no noise component in the video signal, the cyclic coefficient Km = 0 is sent, so the noise cyclic amount Nd in this case is Nd = 0, Noise removal is not performed. Since the amount of noise for each frame included in the video signal is obtained, and the cyclic coefficient Km for noise is obtained according to the signal indicating the amount of noise, the noise cyclic amount for video signals with less noise components The value of Nd becomes small, and the adverse effects of tailing and afterimage can be reduced to make it inconspicuous.

  On the other hand, when it is detected that the degree of motion is low and it is detected that there are many noise components in the video signal, Km = Kmax, so the noise circulation amount Nd other than the motion part is set to the maximum value. This can increase the noise removal effect.

  As described above, according to the noise removal apparatus 5 of the fifth embodiment, the frame noise extraction unit 89 in the noise amount detection unit 80 has a difference pixel having a frame difference and a motion pixel having a difference corresponding to a motion in the frame difference. To determine the pixel accumulated value in the effective area in each frame, and subtract the motion pixel accumulated value mvcnt from the difference pixel accumulated value npcnt, thereby obtaining a noise pixel in a frame unit of a pixel having a difference only in the noise component The accumulated value nsb is obtained, and the noise component in the frame is extracted. Then, an average noise pixel cumulative value frnp of the noise pixel cumulative value nsb is calculated, a noise amount flnd in the frame of the video signal is obtained from the average noise pixel cumulative value frnp, and a control signal efcnt for controlling the effect of noise removal And the cyclic coefficient Km is controlled based on the control signal efcnt to obtain a noise cyclic amount for noise removal.

  Therefore, the amount of noise appropriately included in the frame of the video signal is obtained by separating the motion component and noise component of the video signal included in the difference between frames and obtaining the extent that pixels containing the noise component are present in the frame. And a cyclic coefficient corresponding to the amount of noise in the video signal is obtained. Therefore, the cyclic coefficient can be increased for video signals that contain a lot of noise components, and the amount of noise circulation can be increased for video signals that contain few noise components. It is possible to reduce afterimages and eliminate tailing and afterimages by reducing the noise circulation amount to 0 for video signals and moving parts that do not contain noise components. Therefore, an optimum noise removal effect can be obtained for a video signal having any amount of noise, and adverse effects such as tailing and afterimage can be reduced and a good noise removal effect can be obtained.

  In addition, although each component of the noise removal apparatus 5 has been described as being configured by hardware, each component may be configured to be realized by software processing in program control.

Embodiment 6 FIG.
In the noise removal apparatus 5 of the fifth embodiment, the noise removal means 20 is configured as a frame cyclic noise removal that delays and processes the output signal Do0 after noise removal by one frame. Similarly to the modification according to the second embodiment, the modification to the fifth embodiment can be added. That is, as shown in FIG. 17, the noise removing means 60 of FIG. 9 can be used instead of the noise removing means 20 of FIG.

  FIG. 17 is a block diagram showing the configuration of the non-cyclic noise removal apparatus according to the sixth embodiment (that is, an apparatus that can implement the noise removal method according to the sixth embodiment). In FIG. 17, the same or corresponding components as those in FIGS. 9 and 14 are denoted by the same reference numerals.

  In FIG. 17, the noise removal apparatus 6 according to the sixth embodiment is replaced with the noise removal apparatus 2 according to the second embodiment (FIG. 9) instead of the noise removal means 20 in the noise removal apparatus 5 according to the fifth embodiment (see FIG. 14). In this case, a configuration using a non-recursive noise removal filter by the noise removal means 60 in the reference) is used. The configuration and operation are the same as those described in the second embodiment with reference to FIGS. 9 and 12 except that the frame memory 11 delays the input signal Di0 by one frame. Although omitted, in the case of the acyclic type, the calculation is performed using the frame correlation between two frames of the input frame and the previous frame.

  According to the noise removal apparatus 6 of the sixth embodiment, the frame noise extraction unit 89 in the noise amount detection unit 80 obtains the difference pixel accumulated value npcnt and the motion pixel accumulated value mvcnt, and the motion pixel accumulation from the difference pixel accumulated value npcnt. By subtracting the value mvcnt, a noise pixel cumulative value nsb of a pixel unit having a difference only in the noise component is obtained, and the noise component in the frame is extracted. Then, an average noise pixel accumulated value frnp of the noise pixel accumulated value nsb is calculated, a noise amount flnd in the frame of the video signal is obtained from the average noise pixel accumulated value frnp, and a control signal for controlling the effect of noise removal Efcnt is generated, and the noise removal filter coefficient kn, (1-kn) in the noise removal means 60 is obtained based on the control signal efcnt, and noise removal of the input signal is performed.

  Therefore, the amount of noise appropriately included in the frame of the video signal is obtained by separating the motion component and noise component of the video signal included in the difference between frames and obtaining the extent that pixels containing the noise component are present in the frame. Thus, the coefficient of the noise removal filter corresponding to the amount of noise in the video signal is obtained. Therefore, the amount of noise removal can be increased for video signals that contain a lot of noise components, and the amount of noise removal can be reduced for video signals that contain less noise components to reduce tailing and afterimages. For video signals and moving parts that do not contain noise components, the amount of noise removal can be set to 0 to eliminate tailing and afterimages. Accordingly, an optimum noise removal effect can be obtained for a video signal having any amount of noise, and a good noise removal effect can be obtained by reducing adverse effects such as tailing and afterimage.

  In addition, although each component of the noise removal apparatus 6 has been described as being configured by hardware, each component may be configured to be realized by software processing in program control.

Embodiment 7 FIG.
In the noise removing apparatus 5 of the fifth embodiment, the motion of the video signal is detected from the frame difference, and the motion difference detecting means 50 in the noise amount detecting means 80 and the motion detecting means 23 in the noise removing means 20 are separately provided. However, as described with reference to the third embodiment, it is also possible to obtain a frame difference using a common means and perform motion detection.

  FIG. 18 is a block diagram showing a configuration of a noise removal apparatus according to the seventh embodiment (that is, an apparatus that can carry out the noise removal method according to the seventh embodiment), and constitutes a cyclic noise removal apparatus. Yes. In FIG. 18, the same or corresponding components as those shown in FIGS. 12 and 14 are denoted by the same reference numerals.

  In FIG. 18, the noise removal device 7 of the seventh embodiment is the same as described with respect to the third embodiment, and the motion detection unit 23 and the noise amount detection unit in the noise removal device 5 (see FIG. 14) of the fifth embodiment. The motion difference detecting means 50 in 80 is also used as one motion detecting means 51 for detecting the motion of the video signal, and the amount of noise is detected from the frame difference Diff from the subtracting means 21 and the motion detection result in the motion detecting means 51. A noise amount detecting means 90 is provided. The motion detection means 51 is the same as that in the noise removal apparatus 3 (see FIG. 12) of the third embodiment, and the configuration and operation other than the part related to the noise amount detection means 90 are the fifth embodiment and the third embodiment. The detailed description thereof will be omitted.

  The noise removing apparatus 7 according to the seventh embodiment includes a motion detecting unit 51 that detects the motion of the video signal based on the frame difference obtained by the subtracting unit 21, and a noise that detects the amount of the noise component included in the input video signal. And a quantity detecting means 90.

  The frame memory 11 delays the noise-removed signal Do0 by one frame to generate the previous frame signal Im1, and the subtracting means 21 obtains the frame difference Diff between the current frame signal Di0 and the previous frame signal Im1, and the amplitude limit The means 22 and the timing signal generating means 12 for generating the timing signal based on the synchronization signal SY are the same as those shown in the fifth embodiment, and further, the difference corresponding to the motion is extracted from the frame difference Diff. The motion detection signal 51 for outputting the motion difference signal mv indicating the motion difference value (motion amount) and the motion degree signal mds is the same as that shown in the third embodiment, and detailed description thereof is omitted. To do.

  The noise amount detection means 90 obtains the noise amount in the video signal obtained by removing the difference corresponding to “motion” from the frame difference in units of frames. The absolute value calculation means 32 and the difference for nonlinearly converting the difference absolute value. Sensitivity conversion means 73, frame noise extraction means 89, and noise amount conversion means 87 are provided. These configurations and the operation of converting the noise pixel accumulated value as a value indicating the amount of noise in the video signal and outputting it as the amount of noise are shown in FIG. 14 (Embodiment 5) and FIG. 12 (Embodiment) described above. Since it is the same as 3), its detailed description is omitted.

  According to the noise removal apparatus 7 of the seventh embodiment, the frame difference is obtained by the subtraction means 21, the motion detection means 51 performs motion detection, and the frame noise extraction means 89 in the noise amount detection means 90 performs the difference pixel accumulated value. By obtaining npcnt and the moving pixel cumulative value mvcnt and subtracting the moving pixel cumulative value mvcnt from the difference pixel cumulative value npcnt, a noise pixel cumulative value nsb of a pixel having a difference only in the noise component is obtained. Then, an average noise pixel cumulative value frnp of the noise pixel cumulative value nsb is calculated, a noise amount flnd in the frame of the video signal is obtained from the average noise pixel cumulative value frnp, and a control signal efcnt for controlling the effect of noise removal And the cyclic coefficient Km is controlled based on the control signal efcnt to obtain a noise cyclic amount for noise removal.

  Therefore, the amount of noise appropriately included in the frame of the video signal is obtained by separating the motion component and noise component of the video signal included in the difference between frames and obtaining the extent that pixels containing the noise component are present in the frame. And a cyclic coefficient corresponding to the amount of noise in the video signal is obtained. Therefore, the amount of noise circulation can be increased for video signals containing a lot of noise components, and the amount of noise circulation is reduced for video signals and moving parts with less noise components to reduce tailing and afterimages. In addition, for a video signal or a moving part that does not contain a noise component, the amount of noise circulation can be set to 0 to eliminate tailing and afterimage. Accordingly, an optimum noise removal effect can be obtained for a video signal having any amount of noise, and a good noise removal effect can be obtained by reducing adverse effects such as tailing and afterimage.

  In addition, since the motion detection means 51 is configured to generate and output a motion difference signal mv to the noise amount detection means 90 and a motion degree signal mds to be sent to the cyclic coefficient generation means 24 by noise removal processing. The calculation of the frame difference and the motion detection process can be executed by a common means, and the circuit scale can be reduced.

  In addition, although each component of the noise removal apparatus 7 is described as being configured by hardware, each component may be configured to be realized by software processing in program control.

Embodiment 8 FIG.
The noise removal apparatus 7 according to the seventh embodiment is the same as the noise removal apparatus 5 according to the fifth embodiment, but uses the same means for obtaining the frame difference. As shown in FIG. In the non-cyclic type noise removing apparatus, as described with respect to the sixth embodiment, the means for obtaining the frame difference can be shared.

  FIG. 19 is a block diagram showing a configuration of an acyclic noise removal apparatus according to the eighth embodiment (that is, an apparatus that can implement the noise removal method according to the eighth embodiment). In FIG. 19, the same or corresponding components as those shown in FIGS. 14, 17 and 18 are denoted by the same reference numerals.

  In FIG. 19, the noise removing device 8 of the eighth embodiment is based on the amplitude limiting means 22, the cyclic coefficient generating means 24, the multiplying means 25, and the calculating means 26 in the noise removing apparatus 7 (see FIG. 18) of the fifth embodiment. Instead of the configuration of the noise removal processing, noise removal by the coefficient multiplication means 61 and 62, the filter coefficient generation means 64, and the addition means 65 in the noise removal means 60 in the noise removal apparatus 6 (see FIG. 17) of the sixth embodiment. A configuration using a filter is used. The configuration and operation are the same as those described above with reference to FIGS. 18 and 17 except that the frame memory 11 delays the input signal Di0 by one frame. In the case of the non-recursive type, the calculation is performed using the frame correlation between the input frame and the previous frame.

  In the noise removal apparatus 8 according to the eighth embodiment, the subtraction means 21 obtains the frame difference Diff, the motion detection means 51 performs motion detection, and the frame noise extraction means 89 in the noise amount detection means 90 performs the difference pixel accumulated value. By obtaining npcnt and moving pixel cumulative value mvcnt, and subtracting the moving pixel cumulative value mvcnt from the differential pixel cumulative value npcnt, a noise pixel cumulative value nsb of a pixel having a difference only in the noise component is obtained, The noise component is extracted. Then, an average noise pixel cumulative value frnp of the noise pixel cumulative value nsb is calculated, a noise amount flnd in the frame of the video signal is obtained from the average noise pixel cumulative value frnp, and control for controlling the effect of noise removal A signal efcnt is generated, and filter coefficients kn and (1-kn) for noise removal processing are obtained based on the control signal efcnt to remove noise from the input signal.

  Therefore, the amount of noise appropriately included in the frame of the video signal is obtained by separating the motion component and noise component of the video signal included in the difference between frames and obtaining the extent that pixels containing the noise component are present in the frame. And a coefficient of the noise removal filter corresponding to the amount of noise in the video signal is obtained. Therefore, the amount of noise removal can be increased for video signals that contain a lot of noise components, and the amount of noise removal can be reduced for video signals and moving parts with less noise components to reduce tailing and afterimages. Furthermore, the amount of noise removal can be set to 0 for video signals and moving parts that do not contain noise components, and tailing and afterimages can be eliminated. Accordingly, an optimum noise removal effect can be obtained for a video signal having any amount of noise, and a good noise removal effect can be obtained by reducing adverse effects such as tailing and afterimage.

  Further, since the motion detection unit 51 is configured to generate and output a motion difference signal mv to the noise amount detection unit 90 and a motion degree signal mds to be sent to the filter coefficient generation unit 64, the calculation of the frame difference is performed. The motion detection process can be executed by a common means, and the circuit scale can be reduced.

  In addition, although each component of the noise removal device 8 is described as being configured by hardware, each component may be configured to be realized by software processing in program control.

Embodiment 9 FIG.
In the first to eighth embodiments, the noise removal device and the device capable of performing the noise removal method have been described. However, the present invention performs processing by removing noise from the input video signal and displays the image with high image quality. It can also be applied to signal display devices. Hereinafter, as a ninth embodiment, a TV broadcast signal, a video signal input from a recording / playback device such as a DVD or VTR, a TV broadcast receiving device, or the like is processed, and the noise removal device according to any one of the first to eighth embodiments. A video signal display device for displaying a video signal using the above will be described.

  FIG. 20 is a block diagram illustrating an example of a video signal display device according to the ninth embodiment. 20, the video signal display device 9 according to the ninth embodiment includes an input terminal 101, an input signal processing means 102, and the noise removal device 1 according to the first embodiment (or the noise removal device according to the second to eighth embodiments). 2 to 8), and a display processing means 103 and a display means 104 are provided to display a video signal from which noise has been removed. Note that the configuration and operation of the noise removal apparatus are the same as those shown in Embodiment 1 (or 2 to 8), and a detailed description thereof will be omitted.

  The input terminal 101 receives a TV broadcast signal, a signal from a recording / reproducing device such as a DVD or VTR, a TV broadcast receiving device, or the like. A signal input to the input terminal 101 is sent to the input signal processing means 102.

  The input signal processing means 102 converts, for example, an analog signal to a TV broadcast signal, a recording / playback device such as a DVD or VTR, a TV broadcast receiving device, etc. Input signal processing such as signal separation, and processing such as decoding of MPEG data when MPEG data is received. Then, the video signal subjected to the input signal processing is output to the noise removing device 1 (or any one of 2 to 8). Here, the video signal subjected to the input signal processing may be an interlace signal or a progressive signal.

  The noise removal apparatus 1 (or any one of 2 to 8) is configured so that, with respect to the video signal from the input signal processing unit 102, an input frame and a video signal after noise removal one frame before or a video signal one frame before For detecting the frame difference of the image signal, performing motion detection based on the frame difference, extracting only the noise component excluding the motion component from the frame difference, obtaining the noise amount flnd in the frame of the video signal, and controlling the effect of noise removal A control signal efcnt is generated. By using this control signal efcnt, a noise removal coefficient corresponding to the amount of noise in the video signal is obtained, thereby removing noise from the input signal. Then, the video signal after noise removal is output to the display processing means 103. Since the noise removal operation is the same as that described in the first embodiment (or the second to eighth embodiments), detailed description thereof is omitted.

The display processing means 103 receives the video signal from which noise is removed from the noise removing device 1 (or 2 to 8). When the input video signal is an interlace signal, the display processing unit 103 performs processing such as converting the interlace signal into a progressive signal, and performs signal processing for converting into a display signal such as scaling processing. A display signal for display by means 304 is output to display means 104.
The display unit 104 displays an image based on the display signal from the display processing unit 103.

  As described above, according to the video signal display device of the ninth embodiment, since the noise removing device or the noise removing method of the first embodiment (or 2 to 8) is used, the moving part is not tailed. Afterimages can be reduced, and high-quality video based on video signals that have obtained a good noise removal effect can be displayed.

It is a block diagram which shows an example of a structure of the noise removal apparatus 1 of Embodiment 1 of this invention. It is a figure which shows an example of the input-output characteristic of the difference sensitivity conversion means 34 in the noise removal apparatus 1 of Embodiment 1. It is a figure which shows an example of the input-output characteristic of the motion sensitivity conversion means 502 in the noise removal apparatus 1 of Embodiment 1. It is a figure which shows an example of the input-output characteristic of the noise amount conversion means 38 in the noise removal apparatus 1 of Embodiment 1. 3 is a block diagram illustrating an example of a configuration of a cyclic coefficient generation unit 24 in the noise removal apparatus 1 according to Embodiment 1. FIG. It is a figure which shows an example of the input-output characteristic of the cyclic coefficient generation means 24 in the noise removal apparatus 1 of Embodiment 1. It is a block diagram which shows the other structural example (24b) of the cyclic coefficient generation | occurrence | production means in the noise removal apparatus 1 of Embodiment 1. 3 is a flowchart illustrating an operation of the noise removal device 1 according to the first embodiment. 6 is a block diagram illustrating an example of a configuration of a non-cyclic type noise removal apparatus 2 according to Embodiment 2. FIG. It is a block diagram which shows one structural example (64) of the filter coefficient generation means in the noise removal apparatus 2 of Embodiment 2. It is a block diagram which shows the other structural example (64b) of the filter coefficient generation means in the noise removal apparatus 2 of Embodiment 2. 6 is a block diagram illustrating an example of a configuration of a noise removal device 3 according to Embodiment 3. FIG. FIG. 10 is a block diagram illustrating an example of a configuration of a noise removal device 4 according to a fourth embodiment. FIG. 10 is a block diagram illustrating an example of a configuration of a noise removal device 5 according to a fifth embodiment. It is a figure which shows an example of the input-output characteristic of the noise amount conversion means 87 in the noise removal apparatus 5 of Embodiment 5. 10 is a flowchart illustrating an operation of the noise removal device 5 according to the fifth embodiment. FIG. 10 is a block diagram illustrating an example of a configuration of an acyclic noise removal device 6 according to a sixth embodiment. FIG. 10 is a block diagram illustrating an example of a configuration of a noise removal device 7 according to a seventh embodiment. FIG. 20 is a block diagram illustrating an example of a configuration of a noise removal device 8 according to an eighth embodiment. It is a block diagram which shows an example of a structure of the video signal display apparatus of Embodiment 9 of this invention.

Explanation of symbols

1, 2, 3, 4, 5, 6, 7, 8 Noise removing device, 9 Video signal display device, 11 Frame memory, 12 Timing signal generating means, 20,60 Noise removing means, 21 Subtracting means, 22 Amplitude limiting means , 23, 63 motion detection means, 24, 24b cyclic coefficient generation means, 25 multiplication means, 26 calculation means, 61, 62 coefficient multiplication means, 64 filter coefficient generation means, 65 addition means, 30, 70, 80, 90 noise amount Detecting means, 33 frame difference detecting means, 31 difference calculating means, 32 absolute value calculating means, 34, 73 difference sensitivity converting means, 50, 74 motion difference detecting means, 502, 512 motion sensitivity converting means, 503 motion difference value generating means 51 motion detection means, 511 absolute value calculation means, 513 motion conversion means, 39, 79, 89 frame noise extraction means 35 motion subtraction means, 36 noise difference accumulation calculation means, 37 frame average accumulation value calculation means, 38,87 noise amount conversion means, 40 noise removal processing control means, 241 coefficient calculation means, 242 coefficient limit means, 243, 243b slope Setting means, 244, 244b coefficient maximum value setting means, 245 noise amount coefficient control means, 81 difference comparison means, 82 difference pixel accumulation calculation means, 83 motion difference comparison means, 84 motion pixel accumulation calculation means, 85 motion pixel subtraction means, 86 frame average noise pixel number calculating means, 101 input terminal, 102 input signal processing means, 103 display processing means, 104 display means.

Claims (12)

A three-dimensional noise removing apparatus that removes a noise component having no correlation between frames from a correlation between frames of a video signal,
A frame delay means for delaying an input video signal or a video signal from which noise, which is an output of a noise removal device, has been removed;
A frame difference detection means for inputting a video signal of an input frame and a video signal delayed by the frame delay means, obtaining a signal difference between the frames, and outputting a frame difference;
Motion difference detection means for detecting a motion of a video signal included in the frame difference from the frame difference and outputting a motion difference signal indicating a difference corresponding to the motion;
A frame indicating the degree of the noise component included in the frame by extracting only the noise component in the video signal included in the frame difference by removing the motion difference in the frame difference from the frame difference and the motion difference signal. Frame noise extraction means for outputting the noise extraction result in frame units;
A noise amount converting means for converting the frame noise extraction result into a value indicating a noise amount in a frame of the video signal, detecting the noise amount in the video signal, and outputting;
Noise removal processing control means for generating a control signal for controlling the strength of noise removal according to the amount of noise;
A noise removal apparatus comprising: noise removal means for performing noise removal processing by changing a coefficient for determining noise removal strength according to the control signal. The frame difference detection means includes:
A difference calculating means for inputting a video signal of the current frame and a video signal delayed by the frame delay means, and calculating and outputting a difference between signals of the frames;
An absolute value calculating means for obtaining an absolute value of the difference from the difference calculating means and outputting the absolute value of the difference;
Differential sensitivity conversion means for performing nonlinear conversion by multiplying and subtracting a preset value with respect to the output of the absolute value calculation means, and outputting a differential signal subjected to nonlinear conversion, and output from the absolute value means 2. The noise removing device according to claim 1, wherein the difference absolute value and the difference signal output from the difference sensitivity conversion means are output as the frame difference. The motion difference detecting means is
Motion sensitivity conversion means for performing non-linear conversion by multiplying and subtracting a preset value with respect to the frame difference from the frame difference detection means, and outputting a non-linearly converted difference signal;
Motion difference value generation means for generating a signal indicating a difference value of motion by correcting the value of the difference signal by the difference signal from the motion sensitivity conversion means,
2. The noise removing apparatus according to claim 1, wherein the signal generated by the motion difference value generating means is output as a motion difference signal indicating a value corresponding to a motion included in a frame difference. The frame noise extraction means includes:
Motion subtraction means for subtracting the motion difference signal from the motion difference detection means from the frame difference from the frame difference detection means and outputting as a noise difference signal excluding the motion in the frame difference;
The value indicated by the noise difference signal from the motion subtracting means is accumulated for pixels in a predetermined area or all areas in the target frame of the video signal, and the accumulated result for each frame is output as a noise difference accumulated value. Noise difference accumulation calculation means;
Frame average cumulative value calculating means for obtaining the noise difference cumulative value from the noise difference cumulative calculating means for a plurality of frames, calculating an average value of the noise difference cumulative values in a plurality of frames, and outputting the average noise difference cumulative value; With
4. The average noise difference cumulative value output from the frame average cumulative value calculating means is output as a frame noise extraction result indicating a noise component included in the frame. The noise removal apparatus of crab. The frame noise extraction means includes:
A difference comparison unit that compares the frame difference from the frame difference detection unit with a predetermined value, determines whether each pixel has a difference component, and outputs a signal indicating the difference pixel based on the determination result;
Based on the signal indicating the difference pixel from the difference comparison means, the number of pixels determined to have a difference component is accumulated for pixels in a predetermined region or all regions in the target frame of the video signal, Difference pixel accumulation calculation means for outputting the difference pixel accumulation result as a difference pixel accumulation value;
A motion that compares the motion difference signal from the motion difference detection means with a predetermined value, determines whether each pixel has a difference component corresponding to motion, and outputs a signal indicating the motion pixel based on the determination result Difference comparison means;
Based on the signal indicating the motion pixel from the motion difference comparison means, the number of pixels determined to have motion is accumulated for pixels in a predetermined region or all regions in the target frame of the video signal, and A moving pixel accumulation calculating means for outputting a moving pixel accumulated result as a moving pixel accumulated value;
The number of pixels having a noise-only difference obtained by subtracting the motion pixel cumulative value from the motion pixel cumulative calculation hand from the difference pixel cumulative value from the difference pixel cumulative calculation means, and subtracting the motion difference in the frame difference from the subtraction result. Motion pixel subtraction means for outputting as a noise pixel cumulative value indicating,
Frame average noise pixel number calculating means for obtaining a plurality of frames of noise pixel accumulated values from the moving pixel subtracting means, calculating an average value of noise pixel accumulated values in a plurality of frames, and outputting as an average noise pixel accumulated value; Prepared,
4. The average noise pixel cumulative value output from the frame average noise pixel number calculating means is output as a frame noise extraction result indicating a noise component included in a frame. The noise removal apparatus in any one. The noise amount converting means is
The frame noise extraction result from the frame noise extraction means is input,
The frame noise extraction result is multiplied and subtracted by a preset value, and nonlinear calculation is performed by limiting the calculation result to a predetermined value. The nonlinearly converted value is used as the amount of noise in the video signal. 2. The noise removing device according to claim 1, wherein the noise removing device detects and outputs the detected noise. The noise removal processing control means includes:
Based on the amount of noise from the noise amount conversion means, it changes according to the magnitude of the noise amount, and outputs a signal indicating the level of the strength of the noise removal effect as a control signal,
The noise removing means is
Motion detection means for detecting the motion of the video signal included in the frame difference and outputting a motion motion degree signal indicating the degree of the motion;
The amount of noise in the video signal indicated by the control signal is determined in accordance with the control signal from the noise removal processing control means, while obtaining a noise removal coefficient that changes according to the movement degree signal and determines the strength of noise removal. In order to increase the noise removal effect, the coefficient generation means for controlling the coefficient is provided,
The noise removing apparatus according to claim 1, wherein noise removal is performed using a coefficient adjusted according to the control signal. The coefficient generating means in the noise removing means is
The coefficient takes a value in the range from 0 to 1,
When the amount of noise in the video signal represented by the control signal is at the maximum level, the coefficient is “1”.
The noise removal apparatus according to claim 7, wherein the coefficient is set to "0" when the amount of noise in the video signal represented by the control signal is at a minimum level. The coefficient generating means in the noise removing means is
A slope setting means for setting a value for setting a slope for changing a noise removal coefficient;
Coefficient maximum value setting means for setting the maximum value of the coefficient;
Coefficient calculating means for obtaining and outputting a coefficient based on a motion degree signal from the motion detecting means and a value from the inclination setting means;
Coefficient limiting means for outputting a coefficient whose value is limited to the maximum value from the coefficient maximum value setting means with respect to the coefficient from the coefficient calculating means,
Multiplying the coefficient from the coefficient limiting means by a predetermined magnification that changes in response to the control signal by the control signal from the noise removal processing control means, and limiting the calculation result within a predetermined range. The noise removal apparatus according to claim 7 or 8, further comprising: a noise amount coefficient limiting unit that performs conversion according to (1) and outputs the converted coefficient. The coefficient generating means in the noise removing means is
A slope setting means for setting a value for setting a slope that changes according to the control signal and changes a noise removal coefficient in accordance with a control signal from the noise removal processing control means;
Coefficient maximum value setting means for changing the control signal from the noise removal processing control means according to the control signal and setting the maximum value of the coefficient;
Coefficient calculating means for obtaining and outputting a coefficient based on a motion degree signal from the motion detecting means and a value from the inclination setting means;
9. A coefficient limiting means for outputting a coefficient whose value is limited to a maximum value from the coefficient maximum value setting means with respect to the coefficient from the coefficient calculating means. The noise removal apparatus as described. A three-dimensional noise removal method for removing a noise component having no correlation between frames from a correlation between frames of a video signal,
A frame delay step for delaying the input video signal or the video signal after noise removal in units of frames;
A frame difference detection step of obtaining a difference between signals between the video signal of the input frame and the frame-delayed video signal and outputting the frame difference;
A motion difference detection step of detecting a motion of a video signal included in the frame difference from the frame difference and outputting a motion difference signal indicating a difference corresponding to the motion;
A frame indicating the degree of the noise component included in the frame by extracting only the noise component in the video signal included in the frame difference by removing the motion difference in the frame difference from the frame difference and the motion difference signal. A frame noise extraction step for outputting the noise extraction result in units of frames;
Converting the frame noise extraction result into a value indicating the amount of noise in the frame of the video signal, detecting it as a noise amount in the video signal, and outputting a noise amount conversion step;
A noise removal processing control step for generating a control signal for controlling the strength of noise removal according to the amount of noise,
A noise removal method comprising: a noise removal step of performing noise removal processing by changing a coefficient for determining noise removal strength according to the control signal based on the control signal. A noise removing device according to any one of claims 1 to 10,
Display means for displaying video;
A video signal display device, comprising: a display processing unit that causes the display unit to display a video based on the video signal from which noise has been output output from the noise removal device.

Images