JP5235807B2 - Noise removal device - Google Patents

Noise removal device Download PDF

Info

Publication number
JP5235807B2
JP5235807B2 JP2009172270A JP2009172270A JP5235807B2 JP 5235807 B2 JP5235807 B2 JP 5235807B2 JP 2009172270 A JP2009172270 A JP 2009172270A JP 2009172270 A JP2009172270 A JP 2009172270A JP 5235807 B2 JP5235807 B2 JP 5235807B2
Authority
JP
Japan
Prior art keywords
motion
amount
attenuation coefficient
signal
coefficient
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
JP2009172270A
Other languages
Japanese (ja)
Other versions
JP2011029826A5 (en
JP2011029826A (en
Inventor
善隆 豊田
孝一 山下
俊 伊藤
徹也 久野
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Mitsubishi Electric Corp
Original Assignee
Mitsubishi Electric Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Mitsubishi Electric Corp filed Critical Mitsubishi Electric Corp
Priority to JP2009172270A priority Critical patent/JP5235807B2/en
Publication of JP2011029826A publication Critical patent/JP2011029826A/en
Publication of JP2011029826A5 publication Critical patent/JP2011029826A5/ja
Application granted granted Critical
Publication of JP5235807B2 publication Critical patent/JP5235807B2/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Description

本発明は、映像信号のノイズを除去するノイズ除去装置に関するものであり、特に、画像中の動きを検出し、動き量に応じて適応的にノイズを除去するノイズ除去装置に関する。   The present invention relates to a noise removing device that removes noise from a video signal, and more particularly, to a noise removing device that detects motion in an image and adaptively removes noise according to the amount of motion.

従来から、映像信号のノイズを除去するため、フレーム巡回型のノイズ低減装置が広く用いられている。   Conventionally, in order to remove noise from a video signal, a frame cyclic noise reduction apparatus has been widely used.

従来のノイズ除去装置は、入力映像信号と1画面前の映像信号との差分信号を求め、これに減衰係数(「巡回係数」とも呼び得る)を乗算した信号をノイズ成分として、入力映像信号から差し引くことによってノイズ成分を除去するものである。1画面ごとに順次入力される入力映像信号に対して、上記の処理を繰り返すことで、数画面にわたってノイズ成分を時間方向に平滑化することになり、ノイズを除去することができる。   A conventional noise removal apparatus obtains a difference signal between an input video signal and a video signal one screen before, and uses a signal obtained by multiplying the difference signal by an attenuation coefficient (also referred to as a “cyclic coefficient”) as a noise component from the input video signal. The noise component is removed by subtraction. By repeating the above processing for input video signals that are sequentially input for each screen, noise components are smoothed in the time direction over several screens, and noise can be removed.

減衰係数は0より大きく1より小さい範囲で変化させ、入力映像信号からノイズ成分として差し引く差分信号の割合を決定する。すなわち、減衰係数が大きいほどノイズ除去効果は高くなる。しかし、差分信号にはノイズ成分だけでなく動き成分も含まれるため、減衰係数を大きくするとノイズ成分とともに動き成分まで時間方向に平滑化され、残像が発生するというトレードオフの関係がある。   The attenuation coefficient is changed in a range larger than 0 and smaller than 1, and the ratio of the difference signal to be subtracted as a noise component from the input video signal is determined. That is, the larger the attenuation coefficient, the higher the noise removal effect. However, since the difference signal includes not only the noise component but also the motion component, there is a trade-off relationship that when the attenuation coefficient is increased, the motion component is smoothed in the time direction together with the noise component, and an afterimage is generated.

ここで、残像とは、過去の画面の映像信号が数画面にわたって重ね合わせられることにより、画像中の動きのある部分がボケたり、尾を引いたように見える現象のことである。出力映像信号には、n画面前の映像信号が減衰係数のn乗の割合で重ね合わせられることになるため、減衰係数が大きくなるほど残像は大きくなる。   Here, an afterimage is a phenomenon in which a moving part in an image appears blurred or has a tail when video signals of past screens are superimposed over several screens. Since the video signal before n screens is superimposed on the output video signal at the rate of the nth power of the attenuation coefficient, the afterimage increases as the attenuation coefficient increases.

そこで、従来のノイズ除去装置では、差分信号の大きさに応じて減衰係数を制御することで、ノイズ除去効果を維持しつつ、残像の発生をできるだけ抑えるようにしている。すなわち、差分信号が小さい場合はノイズである確率が高いと判断して減衰係数を大きくし、差分信号が大きい場合は動きである確率が高いと判断して減衰係数を小さくするように制御する。減衰係数Kは、例えば、差分の絶対値の増加とともに、直線的に減少するように制御することができる。   Therefore, in the conventional noise removal device, the attenuation coefficient is controlled according to the magnitude of the difference signal, so that the afterimage generation is suppressed as much as possible while maintaining the noise removal effect. That is, when the difference signal is small, it is determined that the probability of noise is high and the attenuation coefficient is increased, and when the difference signal is large, control is performed so that the probability of motion is high and the attenuation coefficient is decreased. The attenuation coefficient K can be controlled to decrease linearly with an increase in the absolute value of the difference, for example.

減衰係数Kは差分信号の絶対値(|D|とする)の関数として、次式(1)により求められる(特許文献1参照)。

Figure 0005235807
The attenuation coefficient K is obtained by the following equation (1) as a function of the absolute value (referred to as | D |) of the difference signal (see Patent Document 1).
Figure 0005235807

しかしながら、画面間の単純な差分信号のみに基づいて減衰係数を決定する上記のフレーム巡回型ノイズ除去装置では、入力映像信号のノイズの振幅がきわめて大きい場合、ノイズ成分と動き成分の分離が困難となり有効に働かないことがある。例えば、ノイズ成分を動き成分として検出してしまい(誤検出)、動きのない部分でS/Nの改善度が低下したり、動き成分がノイズに埋もれて検出されず(検出漏れ)、動きのある部分で残像が発生しやすくなったりする。   However, in the frame cyclic noise removal device that determines the attenuation coefficient based only on the simple difference signal between the screens, it is difficult to separate the noise component from the motion component when the noise amplitude of the input video signal is extremely large. It may not work effectively. For example, the noise component is detected as a motion component (false detection), the improvement of S / N is reduced in a portion where there is no motion, or the motion component is buried in noise and is not detected (detection omission). An afterimage is likely to occur at a certain part.

こうした問題を解決するため、動き検出手段をさらに設け、入力映像信号と1画面前の映像信号から高精度に動画像の動きを検出して、その動き検出の結果に応じて減衰係数を適応制御するようにしたものがある(特許文献2参照)。   In order to solve these problems, a motion detection means is further provided to detect the motion of the moving image with high accuracy from the input video signal and the video signal of the previous screen, and adaptively control the attenuation coefficient according to the motion detection result. There is something which is made to do (refer patent document 2).

特許文献2に記載のノイズ除去装置では、入力映像信号と1画面前の映像信号との差分信号に対して、非線形回路において、例えば式(1)により求められた係数Kを乗算した信号(出力信号)を出力する。ここで、動き検出回路において検出された動き判定結果に応じて式(1)における係数Kの最大値Kmaxを変化させることで、動きのない部分では係数Kを大きくし、動きのある部分では係数Kを小さくする制御を行っている。さらに、動き検出回路は、入力映像信号と1画面前の映像信号との単純な差分信号だけでなく、動き検出の対象となる画素を中心とした領域の画素の差分信号を利用して動き検出を行うことで、動きの誤検出や検出漏れを減らし、残像の発生を抑えた効果的なノイズ除去を可能にしている。   In the noise elimination device described in Patent Document 2, a signal (output) obtained by multiplying a difference signal between an input video signal and a video signal one screen before by a coefficient K obtained by, for example, Equation (1) in a non-linear circuit. Signal). Here, by changing the maximum value Kmax of the coefficient K in the equation (1) according to the motion determination result detected by the motion detection circuit, the coefficient K is increased in a portion where there is no motion, and the coefficient is determined in a portion where there is motion. Control to reduce K is performed. Furthermore, the motion detection circuit detects motion using not only a simple difference signal between the input video signal and the video signal of the previous screen but also a difference signal of a pixel in a region centering on a pixel to be subjected to motion detection. As a result, it is possible to reduce false detection of motion and detection omission, and to effectively remove noise with reduced afterimage.

特開2002−33942号公報(段落0035、図2)JP 2002-33942 A (paragraph 0035, FIG. 2) 特開平4−280171号公報(段落0032〜0054、図1、図8)JP-A-4-280171 (paragraphs 0032 to 0054, FIGS. 1 and 8)

しかしながら、特許文献2に記載の従来のノイズ除去装置では、動き検出回路による動きの判定結果を3値としており、各判定結果に対する減衰係数Kが大きく異なると、その境界部にS/N改善度が異なることによる段差が生じる場合がある。   However, in the conventional noise removal device described in Patent Document 2, the determination result of the motion by the motion detection circuit is ternary, and if the attenuation coefficient K for each determination result is greatly different, the S / N improvement degree is at the boundary. There may be a difference in level due to the difference.

また、特許文献2に記載の従来のノイズ除去装置では、非線形回路は動き検出回路の動き判定結果に応じて入出力特性を変えているものの、減衰係数Kの最大値を変えるのみであり、差分信号と減衰係数Kの関係については式(1)と同様の関係を用いた例が示されているのみである。   Further, in the conventional noise elimination device described in Patent Document 2, the nonlinear circuit changes the input / output characteristics according to the motion determination result of the motion detection circuit, but only changes the maximum value of the attenuation coefficient K. As for the relationship between the signal and the attenuation coefficient K, only an example using the same relationship as the equation (1) is shown.

式(1)に示した関係により減衰係数Kを決定する従来のノイズ除去装置では、画面間差分信号|D|が一定の閾値以上のときは減衰係数Kがゼロとなる特性であるため、動きの誤検出が発生した場合、画面間差分信号|D|が閾値Thを超えた画素についてはS/Nが改善されない。また、画面間差分信号|D|が一定の閾値Th以下のときは減衰係数Kが急激に高くなる特性であるため、動きの検出漏れが発生した場合、画面間差分信号が閾値Thを下回った画素については残像が大きくなりやすい。よって、式(1)に示した特性では、ノイズによる動きの誤検出や検出漏れが起きうる場合に、ノイズ除去効果が十分に得られない、あるいは残像を十分に抑えることができないという問題があった。   In the conventional noise removing apparatus that determines the attenuation coefficient K based on the relationship shown in the equation (1), the attenuation coefficient K is zero when the inter-screen difference signal | D | When a false detection occurs, the S / N ratio is not improved for pixels for which the inter-screen difference signal | D | exceeds the threshold Th. In addition, when the inter-screen difference signal | D | is equal to or less than a certain threshold value Th, the attenuation coefficient K has a characteristic of rapidly increasing. Therefore, when a motion detection failure occurs, the inter-screen difference signal falls below the threshold value Th. The afterimage tends to be large for pixels. Therefore, the characteristic shown in the equation (1) has a problem that a noise removal effect cannot be sufficiently obtained or an afterimage cannot be sufficiently suppressed when a false detection of motion or detection omission due to noise can occur. It was.

本発明は上述した課題を解決するためになされたもので、入力映像信号の動きを検出し、動きのある部分から動きのない部分まで滑らかなS/N改善を図るとともに、かつ、入力映像信号のノイズの振幅が大きく、動きの誤検出や検出漏れが発生しうる場合にも、動きのない部分では十分にS/Nを改善し、動きのある部分では残像を十分に小さく抑えることができるノイズ除去装置を提供することを目的とする。   The present invention has been made in order to solve the above-described problems, and detects the motion of an input video signal to achieve a smooth S / N improvement from a moving portion to a non-moving portion, and the input video signal. Even when the amplitude of noise is large and erroneous detection or omission of motion may occur, the S / N can be sufficiently improved in a portion where there is no motion, and the afterimage can be suppressed sufficiently small in a portion where there is motion. It aims at providing a noise removal apparatus.

本発明のノイズ除去装置は、
1画面の映像信号と、前記1画面の映像信号の1画面前の映像信号との差を求める第1の減算手段と、
前記1画面の映像信号と、前記1画面の映像信号の1画面前の映像信号から1画面内の画像の動き量を検出する動き検出手段と、
前記動き量の大きさに応じて可変の減衰係数を生成する係数生成手段と、
前記第1の減算手段の出力信号に対して、前記減衰係数を乗ずる乗算手段と、
前記減衰係数を乗じた出力信号を、前記1画像面の映像信号から差し引く第2の減算手段とを具備し、
前記係数生成手段において、前記減衰係数は、前記動き検出手段によって検出された動き量が大きくなるほど小さくなるように定められ、かつ、
前記動き量の変化量に対する前記減衰係数の変化量の比の絶対値は、前記動き量の一部または全ての区間において、前記動き量が大きくなるほど小さくなるように定められ
前記係数生成手段は、
前記動き量をMとし、
前記減衰係数をKとし、
予め定められた前記減衰係数の最大値をKmaxとし、
予め定められた前記減衰係数の最小値をKminとし、
予め定められた静止画領域と判断する前記動き量の閾値をAとし、
予め定められた動画領域と判断する前記動き量の閾値をBとしたとき、次式

Figure 0005235807
の関係を満たす前記減衰係数を生成す
ことを特徴とする。 The noise removal apparatus of the present invention is
First subtracting means for obtaining a difference between a video signal of one screen and a video signal of one screen before the one-screen video signal;
Motion detection means for detecting a motion amount of an image in one screen from the video signal of the one screen and a video signal of one screen before the video signal of the one screen;
Coefficient generating means for generating a variable attenuation coefficient according to the magnitude of the amount of motion;
Multiplying means for multiplying the output signal of the first subtracting means by the attenuation coefficient;
A second subtracting means for subtracting the output signal multiplied by the attenuation coefficient from the video signal of the one image plane;
In the coefficient generation means, the attenuation coefficient is determined so as to decrease as the amount of motion detected by the motion detection means increases, and
The absolute value of the ratio of the change amount of the attenuation coefficient to the change amount of the motion amount is determined so as to decrease as the motion amount increases in a part or all of the motion amount ,
The coefficient generating means includes
The amount of movement is M,
The attenuation coefficient is K,
The maximum value of the predetermined attenuation coefficient is Kmax,
A predetermined minimum value of the attenuation coefficient is defined as Kmin,
A threshold value of the amount of motion that is determined as a predetermined still image area is A,
When the threshold value of the amount of motion to be determined as a predetermined moving image area is B, the following formula
Figure 0005235807
Characterized in that that generates the attenuation coefficient that satisfies the relationship.

本発明によれば、動き検出手段において動き量を求め、係数生成手段において動き量が大きくなるほど小さくなる減衰係数を生成することにより、動きのある部分から動きのない部分まで滑らかなS/N改善を図ることができる。   According to the present invention, the motion detection unit obtains the amount of motion, and the coefficient generation unit generates an attenuation coefficient that decreases as the amount of motion increases, so that a smooth S / N improvement from a portion with motion to a portion without motion is achieved. Can be achieved.

また、本発明によれば、係数生成手段において、動き量が小さい場合は減衰係数の変化量が大きく、動き量が大きい場合は減衰係数の変化量が小さくなる特性により減衰係数を生成することにより、入力映像信号のノイズの振幅が大きく、動きの誤検出や検出漏れが発生しうる場合にも、動きのない部分では十分にS/Nを改善し、動きのある部分では残像を十分に小さく抑えることができる。   Further, according to the present invention, the coefficient generating means generates the attenuation coefficient with a characteristic that the amount of change in the attenuation coefficient is large when the amount of motion is small and the amount of change in the attenuation coefficient is small when the amount of motion is large. Even when the amplitude of the noise of the input video signal is large and erroneous detection or omission of motion may occur, the S / N is sufficiently improved in a portion where there is no motion, and the afterimage is sufficiently small in a portion where there is motion. Can be suppressed.

本発明の実施の形態1のノイズ除去装置の全体的構成を示すブロック図である。It is a block diagram which shows the whole structure of the noise removal apparatus of Embodiment 1 of this invention. 図1の動き検出手段6の内部構成を示すブロック図である。It is a block diagram which shows the internal structure of the motion detection means 6 of FIG. 図1の係数生成手段2の入出力特性の一例を示す図である。It is a figure which shows an example of the input-output characteristic of the coefficient production | generation means 2 of FIG. (A)及び(B)は、係数生成手段2の入出力特性の例を、従来技術における係数生成手段の入出力特性とともに示す図である。(A) And (B) is a figure which shows the example of the input / output characteristic of the coefficient production | generation means 2 with the input / output characteristic of the coefficient production | generation means in a prior art. 係数Kと時定数Tの関係を表す図である。It is a figure showing the relationship between the coefficient K and the time constant T. 係数生成手段2の入出力特性の他の例を示す図である。It is a figure which shows the other example of the input-output characteristic of the coefficient production | generation means 2. 本発明の実施の形態2のノイズ除去装置の全体的構成を示すブロック図である。It is a block diagram which shows the whole structure of the noise removal apparatus of Embodiment 2 of this invention. (A)及び(B)は、輝度信号用係数生成手段302A及び色差信号用係数生成手段302Bの入出力特性を示す図である。(A) And (B) is a figure which shows the input-output characteristic of the coefficient generation means 302A for luminance signals, and the coefficient generation means 302B for color difference signals.

実施の形態1.
図1は本発明の実施の形態1のノイズ除去装置の全体的構成を示すブロック図である。図示のノイズ除去装置は、減算手段1、係数生成手段2、乗算手段3、減算手段4、画面遅延手段5、及び動き検出手段6を有する。
Embodiment 1 FIG.
FIG. 1 is a block diagram showing the overall configuration of the noise removing apparatus according to Embodiment 1 of the present invention. The illustrated noise removing apparatus includes a subtracting unit 1, a coefficient generating unit 2, a multiplying unit 3, a subtracting unit 4, a screen delay unit 5, and a motion detecting unit 6.

入力映像信号Dinは、輝度信号又は色差信号、あるいは3原色信号R,G,Bなどのコンポーネント信号である。減算手段1は、入力映像信号Dinと、画面遅延手段5から出力される1画面前の映像信号D5の差分を求め、画面間差分信号D1を出力する。ここで上記差分は、画素毎に求める。即ち、各画素についての入力映像信号(入力映像信号の各画素の画素値)Dinと同じ画素についての1画面前の映像信号(1画面前の映像信号の当該画素の画素値)D5との差分を求める。   The input video signal Din is a component signal such as a luminance signal, a color difference signal, or three primary color signals R, G, and B. The subtracting means 1 obtains the difference between the input video signal Din and the video signal D5 one screen before output from the screen delay means 5, and outputs an inter-screen difference signal D1. Here, the difference is obtained for each pixel. That is, the difference between the input video signal (pixel value of each pixel of the input video signal) Din for each pixel and the video signal of the previous screen for the same pixel (pixel value of the pixel of the video signal of the previous screen) D5 Ask for.

係数生成手段2は、動き検出手段6により後述のようにして求められた動き量Mに応じて、乗算手段3で乗算する減衰係数Kを生成する。減衰係数Kは各画素について求められる。係数生成手段2における動き量Mに対する減衰係数Kの定め方の特性については後述する。   The coefficient generating means 2 generates an attenuation coefficient K that is multiplied by the multiplying means 3 in accordance with the motion amount M obtained by the motion detecting means 6 as described later. The attenuation coefficient K is obtained for each pixel. The characteristic of how to determine the attenuation coefficient K with respect to the motion amount M in the coefficient generation means 2 will be described later.

乗算手段3は、係数生成手段2により生成された減衰係数Kを、差分信号D1に対して乗算して信号D3を出力する。   The multiplication unit 3 multiplies the difference signal D1 by the attenuation coefficient K generated by the coefficient generation unit 2 and outputs a signal D3.

減算手段4は、入力映像信号Dinから信号D3を減算して、信号D4を出力する。信号D4は、ノイズ除去された出力映像信号Doutとして出力される。   The subtracting means 4 subtracts the signal D3 from the input video signal Din and outputs a signal D4. The signal D4 is output as an output video signal Dout from which noise has been removed.

画面遅延手段5は、ノイズ除去された信号D4に対し、1画面分の遅延処理を行って、次の画面における減算手段1、動き検出手段6での処理で使用される1画面遅延映像信号D5を出力する。   The screen delay means 5 performs a delay process for one screen on the signal D4 from which noise has been removed, and a one-screen delayed video signal D5 used in the processing in the subtraction means 1 and the motion detection means 6 in the next screen. Is output.

動き検出手段6は、入力映像信号Dinと1画面遅延映像信号D5とから、動き量Mを求めて出力する。本発明のノイズ除去装置による動き検出手段6は、入力映像信号Dinと1画面遅延映像信号D5の単純な差分信号ではなく、これらの信号に空間的なフィルタを作用させることによってノイズを低減した信号の差分信号を求めることによって、ノイズの影響を受けにくい動き検出を行う。動き検出手段6の構成例を図2に示す。   The motion detection means 6 obtains and outputs a motion amount M from the input video signal Din and the one-screen delayed video signal D5. The motion detection means 6 by the noise removing apparatus of the present invention is not a simple difference signal between the input video signal Din and the one-screen delayed video signal D5, but a signal in which noise is reduced by applying a spatial filter to these signals. By detecting the difference signal, motion detection that is less susceptible to noise is performed. A configuration example of the motion detection means 6 is shown in FIG.

図2に示した動き検出手段6は、周辺画素平均値算出手段61A、および61B、減算手段62、絶対値算出手段63を有する。
周辺画素平均値算出手段(第1の周辺画素平均値算出手段)61Aは、入力映像信号Dinについて、動き検出の対象となる画素を、その中心又は該中心の近傍に含む領域内の画素、例えば、N×N画素(Nは2以上の整数)から成る領域の画素の信号の平均値を求めて信号D61Aとして出力する。
The motion detection means 6 shown in FIG. 2 includes peripheral pixel average value calculation means 61A and 61B, a subtraction means 62, and an absolute value calculation means 63.
Peripheral pixel average value calculating means (first peripheral pixel average value calculating means) 61A is a pixel in a region that includes a pixel that is a target of motion detection in the center or in the vicinity of the center of the input video signal Din, for example, , N × N pixels (N is an integer equal to or greater than 2), the average value of the signals of the pixels in the region is obtained and output as a signal D61A.

周辺画素平均値算出手段(第2の周辺画素平均値算出手段)61Bは、1画面遅延映像信号D5について、動き検出の対象となる画素を、その中心又は該中心の近傍に含む領域内の画素、例えば、N×N画素から成る領域の画素の信号の平均値を求めて信号D61Bとして出力する。
減算手段62は、信号D61Aと信号D61Bの差を表す差分信号D62を出力する。
絶対値算出手段63は、差分信号D62の絶対値を求めて信号D63を出力する。動き検出手段6は、信号D63を動き量Mとして出力する。
Peripheral pixel average value calculation means (second peripheral pixel average value calculation means) 61B is a pixel in a region that includes a pixel that is a target of motion detection at the center or in the vicinity of the center of the one-screen delayed video signal D5. For example, the average value of the signals of the pixels in the area composed of N × N pixels is obtained and output as the signal D61B.
The subtracting means 62 outputs a difference signal D62 representing the difference between the signal D61A and the signal D61B.
The absolute value calculation means 63 calculates the absolute value of the difference signal D62 and outputs a signal D63. The motion detection means 6 outputs the signal D63 as the motion amount M.

以下、本発明の実施の形態1によるノイズ除去装置の動作について詳しく説明する。
まず、動き検出手段6において、入力映像信号Dinと1画面前の映像信号D5を用いて、各画素の動き量Mを求める。
Hereinafter, the operation of the noise removal apparatus according to Embodiment 1 of the present invention will be described in detail.
First, the motion detection means 6 determines the motion amount M of each pixel using the input video signal Din and the video signal D5 one screen before.

ここで、1画面前の映像信号とは、プログレッシブ方式の映像信号システムの場合、1フレーム前のノイズ除去された出力映像信号D4(Dout)のことである。一方、インターレース方式の映像信号システムの場合、1画面前の映像信号は、1フレーム前(同一フィールド)の出力映像信号D4(Dout)のことであってもよいし、1フィールド前(異なるフィールド)の出力映像信号D4(Dout)のことであってもよい。   Here, in the case of a progressive video signal system, the video signal one screen before is the output video signal D4 (Dout) from which noise has been removed one frame before. On the other hand, in the case of an interlace video signal system, the video signal one screen before may be the output video signal D4 (Dout) one frame before (same field) or one field before (different field). Output video signal D4 (Dout).

動き検出手段6では、まず、周辺画素平均値算出手段61A、61Bにおいて、入力映像信号Dinおよび1画面前の映像信号D5について、動き検出の対象となる画素を、その中心又は該中心の近傍に含む領域内の画素、例えば、N×N画素から成る領域の画素の信号の平均値を求めた信号D61A、D61Bを生成する。次に減算手段62において信号D61Aと信号D61Bの差分D62を計算し、その絶対値D63を絶対値算出手段63により計算して、これを動き量Mとして出力する。減算手段62と絶対値算出手段63とにより、差分絶対値算出手段が構成される。   In the motion detection means 6, first, in the peripheral pixel average value calculation means 61A and 61B, the pixel to be subjected to motion detection is placed at the center or in the vicinity of the input video signal Din and the video signal D5 one screen before. Signals D61A and D61B obtained by calculating the average value of the signals of the pixels in the included region, for example, the pixels in the region composed of N × N pixels, are generated. Next, the subtraction means 62 calculates the difference D62 between the signal D61A and the signal D61B, the absolute value D63 is calculated by the absolute value calculation means 63, and this is output as the movement amount M. The subtraction means 62 and the absolute value calculation means 63 constitute a difference absolute value calculation means.

ここで、動き検出におけるノイズの影響を取り除くため、周辺画素平均値算出手段61A、61Bは、周辺16×16画素や周辺32×32画素など、比較的大きな領域の平均値を求める。この処理により近傍画素間で相関のないノイズ成分は取り除かれ、したがって動き量Mは真の動き成分をより正確に反映したものとなる。   Here, in order to remove the influence of noise in motion detection, the peripheral pixel average value calculating means 61A and 61B calculate the average value of a relatively large region such as the peripheral 16 × 16 pixels or the peripheral 32 × 32 pixels. By this processing, noise components having no correlation between neighboring pixels are removed, and therefore the motion amount M more accurately reflects the true motion component.

次に、係数生成手段2において、動き量Mから、図3に示した関係により減衰係数Kを生成する。   Next, the coefficient generation means 2 generates the attenuation coefficient K from the motion amount M according to the relationship shown in FIG.

図3に示した関係は、減衰係数Kは動き量Mの関数として、次式(2)により表現される。

Figure 0005235807
In the relationship shown in FIG. 3, the attenuation coefficient K is expressed by the following equation (2) as a function of the motion amount M.
Figure 0005235807

ここで、Kは減衰係数、Kmaxは減衰係数の最大値、Kminは減衰係数の最小値、Mは動き量、Aは静止画領域と判断する動き量の閾値、Bは動画領域と判断する動き量の閾値(A<B)である。   Here, K is the attenuation coefficient, Kmax is the maximum value of the attenuation coefficient, Kmin is the minimum value of the attenuation coefficient, M is the amount of motion, A is the threshold of the amount of motion that is determined as a still image region, and B is the motion that is determined as a moving image region. This is the amount threshold (A <B).

動き量Mが閾値A以下である画素は静止画領域内の画素と判断し、減衰係数Kを最大値Kmaxとする。また、動き量Mが閾値B以上である画素は動画領域内の画素と判断し、減衰係数Kを最小値Kminとする。さらに、動き量Mが閾値Aと閾値Bの間にあり、静止画領域とも動画領域とも判断されない領域は中間領域とし、この区間では動き量Mに応じて下に凸で減少する特性に基づいて係数Kを決定する。ここで、下に凸で減少する特性とは、動き量Mが大きくなるほど係数Kが減少し、かつ、動き量が小さい場合は減衰係数の変化量が比較的大きく、動き量が大きい場合は減衰係数の変化量が比較的小さくなる特性のことである。   A pixel whose amount of motion M is equal to or less than the threshold A is determined as a pixel in the still image region, and the attenuation coefficient K is set to the maximum value Kmax. In addition, a pixel having the motion amount M equal to or greater than the threshold value B is determined as a pixel in the moving image region, and the attenuation coefficient K is set to the minimum value Kmin. Further, an area where the motion amount M is between the threshold value A and the threshold value B and is not determined as a still image region or a moving image region is defined as an intermediate region. The coefficient K is determined. Here, the downward and convex characteristic is that the coefficient K decreases as the amount of motion M increases, and the amount of change in the attenuation coefficient is relatively large when the amount of motion is small, and is attenuated when the amount of motion is large. This is a characteristic in which the coefficient change amount is relatively small.

なお、ノイズの振幅がきわめて大きく、ノイズと動きの分離が難しい場合には、動きの誤検出や検出漏れの程度が大きくなり、静止画領域と動画領域を閾値A、Bにて明確に判別することが妥当でない場合もある。このような場合、閾値Aを0、閾値Bを動き量Mの最大値(Mが取り得る値の範囲の上限値)に設定し、全ての区間で下に凸で減少する特性に基づいて係数Kを決定することとしてもよい。   If the amplitude of noise is extremely large and it is difficult to separate noise and motion, the degree of false detection of motion and detection omission becomes large, and the still image region and the moving image region are clearly discriminated by threshold values A and B. It may not be appropriate. In such a case, the threshold A is set to 0, the threshold B is set to the maximum value of the amount of motion M (the upper limit of the range of values that M can take), and the coefficient is based on the characteristic that decreases downwardly in all sections. K may be determined.

係数生成手段2は、動き検出手段6で求められた各画素の動き量Mに応じて、以上で説明した特性により、各画素に対する減衰係数Kを生成する。   The coefficient generation unit 2 generates an attenuation coefficient K for each pixel according to the characteristics described above according to the amount of movement M of each pixel obtained by the motion detection unit 6.

なお、係数生成手段2として、式(2)による演算を実行して減衰係数Kを得てもよいが、予め動き量Mに対応する減衰係数Kをルックアップテーブル(LUT)の形で保持しておくこともできる。このようなLUTを用いた場合には、式(2)の演算を行う必要がなくなるので、係数生成手段2における処理の簡単化を図ることができる。   The coefficient generation means 2 may obtain the attenuation coefficient K by performing the calculation according to the equation (2). However, the attenuation coefficient K corresponding to the motion amount M is stored in the form of a lookup table (LUT) in advance. You can also keep it. When such an LUT is used, it is not necessary to perform the calculation of Expression (2), so that the processing in the coefficient generation means 2 can be simplified.

次に、減算手段1において、入力映像信号Dinの現在の1画面と、1画面前の映像信号D5の差分信号D1を各画素について計算する。即ち、減算は、入力映像信号Dinと1画面前の映像信号D5の同じ位置又は対応する位置の画素について行われる。ここで、1画面前の映像信号が異なるフィールドの映像信号である場合には、空間的に1ライン分ずれた位置に存在する画素を上記「対応する位置の画素」として扱い、これらの差分を求める。   Next, in the subtracting means 1, a difference signal D1 between the current one screen of the input video signal Din and the video signal D5 one screen before is calculated for each pixel. That is, the subtraction is performed on pixels at the same position or corresponding positions of the input video signal Din and the video signal D5 one screen before. Here, when the video signal of the previous screen is a video signal in a different field, the pixel existing at the position shifted by one line spatially is treated as the above-mentioned “pixel at the corresponding position”, and these differences are determined. Ask.

次に、乗算手段3において、係数生成手段2により生成された減衰係数Kを、差分信号D1に対して乗算し、信号D3を生成する。乗算手段3では、差分信号D1の各画素に対して、同じ画素について求められた減衰係数Kの乗算が行われる。   Next, the multiplication unit 3 multiplies the difference signal D1 by the attenuation coefficient K generated by the coefficient generation unit 2 to generate a signal D3. In the multiplication means 3, each pixel of the difference signal D1 is multiplied by the attenuation coefficient K obtained for the same pixel.

次に、減算手段4において、入力映像信号Dinから信号D3を減算し、ノイズ除去された信号D4を生成する。減算は、入力映像信号Dinと信号D3の同じ画素について行われる。さらに、信号D4を、現在の画面のノイズ除去された信号Doutとして出力する。   Next, the subtracting unit 4 subtracts the signal D3 from the input video signal Din to generate a signal D4 from which noise has been removed. The subtraction is performed on the same pixel of the input video signal Din and the signal D3. Further, the signal D4 is output as a signal Dout from which noise is removed from the current screen.

最後に、画面遅延手段5において、ノイズ除去された信号D4に対して、1画面分の遅延処理を行う。画面遅延手段5は、ノイズ除去された信号D4をフレームメモリ等に蓄積し、現在の画面の1画面後の処理が行われるタイミングで、1画面前の映像信号D5として出力される。画面遅延手段5による遅延量は、プログレッシブ方式の場合は1フレーム遅延であり、インターレース方式の場合はフレーム遅延又はフィールド遅延のどちらでもよい。   Finally, the screen delay means 5 performs a delay process for one screen on the signal D4 from which noise has been removed. The screen delay means 5 accumulates the noise-removed signal D4 in a frame memory or the like, and outputs it as a video signal D5 one screen before at the timing when the processing one screen after the current screen is performed. The delay amount by the screen delay means 5 is one frame delay in the case of the progressive method, and may be either a frame delay or a field delay in the case of the interlace method.

上記の一連の処理を、1画面毎に順次入力される入力映像信号Dinに対して実行し、ノイズ除去された出力映像信号Doutを順次出力する。
以上が、本発明の実施の形態1によるノイズ除去装置の動作の説明である。
The above-described series of processing is executed for the input video signal Din sequentially input for each screen, and the output video signal Dout from which noise is removed is sequentially output.
The above is description of operation | movement of the noise removal apparatus by Embodiment 1 of this invention.

次に、本発明におけるノイズ除去装置の作用、効果について説明する。
本発明の実施の形態1によるノイズ除去装置は、動き検出手段6において、入力映像信号Dinとノイズ除去された1画面前の映像信号D5の各画素に対して、周辺N×N画素の平均値を求めることによって、近隣画素間で相関のないノイズ成分を取り除いている。この処理により、単純に各画素の画面間差分を求めるだけの場合に比べて、ノイズによる動きの誤検出や検出漏れを減らすことができる。
Next, the operation and effect of the noise removing apparatus according to the present invention will be described.
In the noise removal apparatus according to the first embodiment of the present invention, the motion detection means 6 uses the average value of the peripheral N × N pixels for each pixel of the input video signal Din and the video signal D5 one screen before noise removal. Thus, noise components having no correlation between neighboring pixels are removed. By this processing, it is possible to reduce erroneous detection of motion and omission of detection due to noise, compared to the case of simply obtaining the inter-screen difference of each pixel.

しかしながら、ノイズの振幅がきわめて大きい場合には、上記動き検出手段6を用いても、誤検出や検出漏れを完全に防ぐことはできない。そこで、本発明の実施の形態1によるノイズ除去装置では、係数生成手段2において、動き量Mに応じて下に凸で減少する特性に基づいて減衰係数Kを生成することにより、入力映像信号のノイズの振幅が大きく、動きの誤検出や検出漏れが発生しうる場合にも、動きのない部分では十分にS/Nを改善し、動きのある部分では残像を十分に小さく抑えることができる。   However, if the amplitude of the noise is extremely large, erroneous detection and detection omission cannot be completely prevented even if the motion detection means 6 is used. Therefore, in the noise removal apparatus according to the first embodiment of the present invention, the coefficient generation unit 2 generates the attenuation coefficient K based on the characteristic that decreases in a convex manner according to the amount of motion M, so Even when the amplitude of noise is large and erroneous detection or omission of motion may occur, the S / N can be sufficiently improved in a portion where there is no motion, and the afterimage can be suppressed sufficiently small in a portion where there is motion.

以下、係数生成手段2における特性が従来例に比べて有利である理由について説明する。
図4(A)及び(B)は、従来の直線で表される特性と、本発明の係数生成手段2による特性とを比較するための図である。
Hereinafter, the reason why the characteristics of the coefficient generating means 2 are more advantageous than the conventional example will be described.
4A and 4B are diagrams for comparing the characteristic represented by the conventional straight line with the characteristic of the coefficient generation means 2 of the present invention.

図4(A)は、式(2)で表される係数生成手段2の特性(実線)と、式(1)で表される従来の特性(点線)を比較するための図である。ここで、式(2)における閾値Bと、式(1)における閾値Thとの間には、Th<Bの関係が成り立つものとした。   FIG. 4A is a diagram for comparing the characteristic (solid line) of the coefficient generation means 2 represented by the expression (2) with the conventional characteristic (dotted line) represented by the expression (1). Here, a relationship of Th <B is established between the threshold value B in Expression (2) and the threshold value Th in Expression (1).

図4(A)に示した係数生成手段2の特性(実線)では、動き量が小さい領域(MA1)と、動き量が多い領域(MA3)において、従来の特性(点線)よりも係数が大きく設定され、中間の領域(MA2)では、従来の特性(点線)よりも係数が小さく設定されている。動き量が最大値に近い領域(MA4)では、従来の特性(点線)と同様、係数がゼロに設定されている。即ち、式(2)において、Kmin=0と設定されている。   In the characteristic (solid line) of the coefficient generation means 2 shown in FIG. 4A, the coefficient is larger in the region (MA1) where the amount of motion is small and the region (MA3) where the amount of motion is large than in the conventional characteristic (dotted line). In the middle region (MA2), the coefficient is set smaller than that of the conventional characteristic (dotted line). In the region where the amount of motion is close to the maximum value (MA4), the coefficient is set to zero as in the conventional characteristic (dotted line). That is, in the formula (2), Kmin = 0 is set.

領域MA1では、動き量はほぼゼロに近く、静止画領域である確率がきわめて高い。この領域では、従来の特性(点線)よりも係数Kを大きく設定して巡回量を大きくすることで、ノイズ除去効果を高め、静止画領域のS/Nを改善することができる。   In the area MA1, the amount of motion is almost zero, and the probability of being a still image area is extremely high. In this area, the noise removal effect can be enhanced and the S / N of the still picture area can be improved by setting the coefficient K larger than the conventional characteristic (dotted line) to increase the amount of circulation.

領域MA2では、動き成分がある程度含まれる確率が高く、係数Kを大きく設定しすぎると動きのある部分の残像が目立つようになる。この領域では、従来の特性(点線)よりも低い係数を設定することで、残像の発生を抑えることができる。また、下に凸の特性であるため、領域MA2の境界部では従来の特性(点線)と同じレベルになり、境界部付近でのS/N改善度は従来の特性(点線)のレベルをできるだけ維持できる。   In the area MA2, there is a high probability that a motion component is included to some extent. If the coefficient K is set too large, an afterimage of a portion with motion becomes noticeable. In this region, by setting a coefficient lower than the conventional characteristic (dotted line), the occurrence of afterimage can be suppressed. Further, since the characteristic is convex downward, the boundary of the region MA2 is at the same level as the conventional characteristic (dotted line), and the S / N improvement near the boundary is as high as the level of the conventional characteristic (dotted line). Can be maintained.

領域MA3では、動き量が大きいため動画領域である確率が高いが、ノイズの振幅が大きい場合には、ノイズを動きとして検出することによる誤検出が動き量に与える割合も無視できない。従来の特性(点線)では、動き量が閾値Thを超えると係数がゼロとなってしまい、誤検出の場合にS/N改善度が低下してしまっていた。係数生成手段2の特性(実線)では、動き量が閾値Thを超えてからも係数Kはゼロ以上の値をもつため、誤検出があってもゼロ以上の係数Kを作用させることができる。したがって、動きの誤検出によるS/N改善効果の低下を防止することができる。   In the area MA3, since the amount of motion is large, the probability that it is a moving image area is high. However, when the noise amplitude is large, the rate of false detection by detecting noise as motion cannot be ignored. In the conventional characteristic (dotted line), when the amount of motion exceeds the threshold Th, the coefficient becomes zero, and the S / N improvement is reduced in the case of erroneous detection. In the characteristic (solid line) of the coefficient generating means 2, since the coefficient K has a value of zero or more even after the amount of motion exceeds the threshold Th, the coefficient K of zero or more can be applied even if there is a false detection. Therefore, it is possible to prevent a decrease in the S / N improvement effect due to erroneous detection of motion.

領域MA4では、動き量はほぼ最大値に近く、動画領域である確率がきわめて高い。この領域では、閾値B以上で係数Kをゼロにした特性(実線)により、残像を発生させないようにすることができる。   In the area MA4, the amount of motion is close to the maximum value, and the probability of being a moving image area is extremely high. In this region, afterimages can be prevented from being generated by the characteristic (solid line) that is equal to or greater than the threshold value B and the coefficient K is zero.

以上、図4(A)を用いて説明したように、本発明の実施の形態1によるノイズ除去装置は、動き量に応じた係数を下に凸で減少する特性に基づいて生成することによって、入力映像信号のノイズの振幅が大きく、動きの誤検出や検出漏れが発生しうる場合にも、ノイズ除去効果を損なうことなく、動きのある部分では残像を小さく抑えることができる。   As described above with reference to FIG. 4A, the noise removal apparatus according to the first embodiment of the present invention generates a coefficient corresponding to the amount of motion based on a characteristic that decreases in a convex manner downward, Even when the amplitude of noise of the input video signal is large and erroneous detection or omission of motion may occur, the afterimage can be suppressed small in a portion where there is motion without impairing the noise removal effect.

また、図4(B)は、式(2)で表される係数生成手段2の特性(実線)と、線形な関係により定義される特性(点線L1、L2)を比較するための図である。特性L1は、係数Kの最大値と最小値を係数生成手段2の特性(実線)と一致させたもの、特性L2は、係数Kを動き量Mについて積分した値が、係数生成手段2の特性(実線)のそれと略等しくなるようにしたものをそれぞれ表している。ここで、係数生成手段2の特性(実線)では、式(2)における閾値A、Bによる切り替えは設けず、A=0、B=(動き量Mの最大値)とした。   FIG. 4B is a diagram for comparing the characteristic (solid line) of the coefficient generation means 2 expressed by the expression (2) with the characteristic (dotted lines L1, L2) defined by a linear relationship. . The characteristic L1 is obtained by matching the maximum value and the minimum value of the coefficient K with the characteristic (solid line) of the coefficient generation unit 2, and the characteristic L2 is obtained by integrating the coefficient K with the amount of motion M. Each of them is made to be substantially equal to that of (solid line). Here, in the characteristic (solid line) of the coefficient generation means 2, switching by the threshold values A and B in the equation (2) is not provided, and A = 0 and B = (maximum value of the motion amount M).

特性L1は、係数生成手段2の特性(実線)と比べると係数Kが全範囲において大きいため、S/N改善度が大きい。しかし、動き成分がある程度含まれると考えられる領域(MB2)においても比較的高い係数をもつため、残像が大きくなりやすい。一方、係数生成手段2の特性(実線)では、領域MB2において係数を小さく設定することにより、残像の発生を抑えることができる。かつ、係数Kの最大値Kmax、最小値Kminは同じであることから、動きのない部分でのS/N改善度は同等であることが期待できる。   The characteristic L1 has a large S / N improvement since the coefficient K is larger in the entire range than the characteristic (solid line) of the coefficient generating means 2. However, an afterimage tends to be large because the region (MB2) that is considered to contain some motion component has a relatively high coefficient. On the other hand, in the characteristic (solid line) of the coefficient generation means 2, the occurrence of afterimages can be suppressed by setting the coefficient small in the area MB2. In addition, since the maximum value Kmax and the minimum value Kmin of the coefficient K are the same, it can be expected that the S / N improvement degree in the portion where there is no movement is equivalent.

また、係数Kの総和(動き量Mについて積分した値)が同じになるようにした特性L2と比較すると、係数生成手段2の特性(実線)では、動き量が小さい領域(MB1)と、動き量が大きい領域(MB3)において、特性L2よりも係数を大きく設定でき、中間の領域(MB2)では、特性L2よりも係数を小さく設定できる。   In addition, when compared with the characteristic L2 in which the sum of the coefficients K (the value integrated with respect to the movement amount M) is the same, the characteristic (solid line) of the coefficient generation means 2 has a smaller movement amount (MB1) and movement. In the region (MB3) where the amount is large, the coefficient can be set larger than the characteristic L2, and in the intermediate region (MB2), the coefficient can be set smaller than the characteristic L2.

領域MB1では、動き量Mはほぼゼロに近く、静止画領域である確率がきわめて高い。この領域では、特性L2よりも係数Kを大きく設定することで、静止画領域のS/Nを改善することができる。   In the area MB1, the motion amount M is almost zero, and the probability of being a still image area is extremely high. In this region, the S / N of the still image region can be improved by setting the coefficient K larger than the characteristic L2.

領域MB2では、動き成分がある程度含まれる確率が高く、係数Kを大きく設定しすぎると動きのある部分の残像が目立つようになる。この領域では、特性L2よりも低い係数を設定することで、残像の発生を抑えることができる。また、下に凸の特性であるため、領域MB2の境界部では従来の特性L2と同じレベルとすることができ、境界部付近でのS/N改善度は従来の特性L2のレベルをできるだけ維持している。   In the area MB2, there is a high probability that a motion component is included to some extent. If the coefficient K is set too large, an afterimage of a portion with motion becomes noticeable. In this region, by setting a coefficient lower than the characteristic L2, occurrence of afterimage can be suppressed. In addition, since it has a downwardly convex characteristic, it can be set to the same level as the conventional characteristic L2 at the boundary portion of the region MB2, and the S / N improvement in the vicinity of the boundary part maintains the level of the conventional characteristic L2 as much as possible. doing.

領域MB3では、動き量が大きいため動画領域である確率が高いが、ノイズの振幅が大きい場合には、ノイズを動きとして検出することによる誤検出が動き量に与える割合も無視できない。特性L2によれば、この領域では係数Kの値は線形に減少するため、動きの誤検出の場合にS/N改善度の低下が大きくなっていた。しかし、係数生成手段2の特性(実線)では、係数Kの最小値に漸近するように緩やかに減少するので、動きがあるところでも一定のノイズ除去を行うことができ、動きの誤検出があってもS/N改善度の低下が少ない。したがって、動きの誤検出によるS/N改善効果の低下を防止することができる。   In the area MB3, since the amount of motion is large, the probability that it is a moving image region is high. However, when the amplitude of noise is large, the ratio of false detection by detecting noise as motion cannot be ignored. According to the characteristic L2, since the value of the coefficient K decreases linearly in this region, the reduction in the S / N improvement is large in the case of erroneous detection of motion. However, in the characteristic (solid line) of the coefficient generation means 2, since it gradually decreases so as to be asymptotic to the minimum value of the coefficient K, it is possible to perform constant noise removal even when there is movement, and there is an erroneous detection of movement. However, there is little decrease in the S / N improvement. Therefore, it is possible to prevent a decrease in the S / N improvement effect due to erroneous detection of motion.

以上、図4(A)及び(B)を参照して説明したように、本発明の実施の形態1によるノイズ除去装置は、係数生成手段2において、動き量Mが小さい領域では係数Kの変化量が比較的大きくなる特性により係数Kを決定することで、動き量Mが小さい領域でのS/Nを効果的に改善し、さらに、動き量Mが大きい領域では係数Kの変化量が比較的小さくなる特性により係数Kを決定することで、動き量Mが大きい領域での係数の低下を防ぎ、動きの誤検出によるS/N改善度の低下を防止することができる。   As described above with reference to FIGS. 4A and 4B, the noise removal apparatus according to Embodiment 1 of the present invention uses the coefficient generator 2 to change the coefficient K in the region where the motion amount M is small. By determining the coefficient K according to the characteristic that the amount is relatively large, the S / N in the region where the motion amount M is small is effectively improved, and further, the change amount of the coefficient K is compared in the region where the motion amount M is large. By determining the coefficient K based on the characteristic that becomes smaller, it is possible to prevent a decrease in the coefficient in a region where the motion amount M is large, and to prevent a decrease in the S / N improvement due to erroneous detection of motion.

また、上記特性により動き量Mを係数Kに変換することで、動きのある部分と動きのない部分の係数Kの値に差をつけ、動き量Mに応じたS/N改善度の強弱にメリハリをつける作用がある。上記特性は係数Kの変化がほとんどの領域において滑らかであるから、S/N改善度について急激な段差ができることなく、違和感のない自然な表示が可能である。   Further, by converting the motion amount M into the coefficient K according to the above characteristics, the difference in the coefficient K between the moving portion and the non-moving portion is differentiated, and the S / N improvement degree according to the motion amount M is increased or decreased. Has the effect of sharpening. In the above characteristics, the change in the coefficient K is smooth in most regions, and therefore, a natural display without a sense of incongruity is possible without making a steep step in the S / N improvement degree.

次に、別の観点から図3および式(2)に示した特性を用いることの意味を述べる。
一般に、フレーム巡回型のノイズ除去装置において、係数Kを作用させたときの残像の減衰の時定数Tは、次式(3)で表される。
T=−1/lnK×1/FR[sec] (3)
ここで、FRは1[sec]あたりの画面数である。
Next, the meaning of using the characteristics shown in FIG. 3 and formula (2) from another viewpoint will be described.
In general, in a frame recursive noise removal apparatus, a time constant T of afterimage attenuation when a coefficient K is applied is expressed by the following equation (3).
T = −1 / lnK × 1 / FR [sec] (3)
Here, FR is the number of screens per 1 [sec].

図5は、FR=30の場合の時定数T[sec]のグラフを示している。時定数Tは、K=1に近い値で急激に増加することがわかる。   FIG. 5 shows a graph of the time constant T [sec] when FR = 30. It can be seen that the time constant T increases rapidly at a value close to K = 1.

上記の残像特性を考慮すると、残像を目立たないようにするためには、K=1に近い値に設定する領域を極力小さくしたほうがよい。そこで、動き量Mが小さい領域では係数Kの変化量が大きくなる特性により係数Kを設定する(即ち、動き量Mの増加に伴い、Kを急激に小さくする)ことは理に適っている。   Considering the above afterimage characteristics, it is better to make the region set to a value close to K = 1 as small as possible in order to make the afterimage unnoticeable. Therefore, it is reasonable to set the coefficient K based on the characteristic that the change amount of the coefficient K is large in the region where the motion amount M is small (that is, K is rapidly decreased as the motion amount M increases).

また、図5の残像特性によれば、略K=0.6以下の係数を設定しても残像の増加は目立たないということを示している。したがって、動き量Mが大きい領域においても、ゼロ以上の係数Kを設定することで、残像を目立たせずにS/Nの改善を図ることができる。したがって、動き量Mが大きい領域では係数Kの変化量が小さくなり、ゼロ以上の最小値に漸近する特性に基づいて係数Kを設定することは理に適っている。   Further, the afterimage characteristics shown in FIG. 5 indicate that the increase in afterimage is not noticeable even when a coefficient of approximately K = 0.6 or less is set. Therefore, even in a region where the amount of motion M is large, by setting a coefficient K equal to or greater than zero, it is possible to improve the S / N without conspicuous afterimages. Therefore, it is reasonable to set the coefficient K based on the characteristic that the change amount of the coefficient K is small in the region where the motion amount M is large and asymptotically approaches a minimum value of zero or more.

以上で説明したように、動き量Mが小さい領域では係数Kの変化量が大きくなる特性により係数Kを決定し、動き量が大きい領域では係数Kの変化量が小さくなる特性により係数Kを決定することで、残像を目立たないようにしつつ、できるだけ大きなS/N改善効果を得ることができる。   As described above, the coefficient K is determined based on the characteristic that the change amount of the coefficient K is large in the region where the motion amount M is small, and the coefficient K is determined based on the characteristic that the change amount of the coefficient K is small in the region where the motion amount is large. By doing so, it is possible to obtain as large an S / N improvement effect as possible while making the afterimage inconspicuous.

以上で説明したように、本発明によるノイズ除去装置では、周辺N×N画素で平均化した画面間の差分を用いて動き量を検出し、動き量に応じた係数を下に凸で減少する特性に基づいて生成することによって、動きのある部分から動きのない部分まで滑らかなS/N改善を図るとともに、かつ、入力映像信号のノイズの振幅が大きく、動きの誤検出や検出漏れが発生しうる場合にも、動きのない部分では十分にS/Nを改善し、動きのある部分では残像を十分に小さく抑えることができる。   As described above, in the noise removal device according to the present invention, the amount of motion is detected using the difference between the screens averaged by the peripheral N × N pixels, and the coefficient corresponding to the amount of motion is convexly decreased. By generating based on the characteristics, smooth S / N improvement is made from the moving part to the non-moving part, and the noise amplitude of the input video signal is large, resulting in erroneous detection and omission of detection. Even in such a case, the S / N can be sufficiently improved in a portion where there is no movement, and the afterimage can be suppressed sufficiently small in a portion where there is movement.

なお、上記効果を期待できるものであれば、図3あるいは式(2)で示した関係のみによらず、同様の特性をもつあらゆる関係により係数Kを求めてもよい。例えば、図6に示したような特性を用いることができる。図6に示した特性による係数Kは、動き量Mが0からm1までの区間では傾きk1で減少し、動き量Mがm1からm2までの区間では傾きk2で減少し、動き量Mがm2からm3までの区間では傾きk3で減少し、動き量Mがm3からの区間では傾きk4で減少し、これらの傾きの大小関係は|k1|≧|k2|≧|k3|≧|k4|となっている。この特性は、下に凸で減少する特性の近似特性であるから、上記で説明したのと同様の効果を得ることができる。また、各区間において線形演算のみで係数Kを求めることができるので、係数生成手段2における演算処理の簡単化を図ることができる。   As long as the above effect can be expected, the coefficient K may be obtained by any relationship having similar characteristics, not only by the relationship shown in FIG. 3 or Expression (2). For example, characteristics as shown in FIG. 6 can be used. The coefficient K according to the characteristics shown in FIG. 6 decreases with a slope k1 in a section where the motion amount M is from 0 to m1, decreases in a slope k2 in a section where the motion amount M is from m1 to m2, and the motion amount M is m2. In the interval from m3 to m3, the amount of motion M decreases in the interval from m3 to the gradient k4, and the magnitude relationship between these gradients is | k1 | ≧ | k2 | ≧ | k3 | ≧ | k4 | It has become. Since this characteristic is an approximate characteristic of a characteristic that is convex downward and decreases, the same effect as described above can be obtained. In addition, since the coefficient K can be obtained only by linear calculation in each section, the calculation processing in the coefficient generation means 2 can be simplified.

実施の形態2.
図7は本発明の実施の形態2のノイズ除去装置の全体的構成を示すブロック図である。図示のノイズ除去装置は、輝度信号減算手段301A、輝度信号用係数生成手段302A、輝度信号乗算手段303A、輝度信号減算手段304A、輝度信号画面遅延手段305A、動き検出手段306、色差信号減算手段301B、色差信号用係数生成手段302B、色差信号乗算手段303B、色差信号減算手段304B、及び色差信号画面遅延手段305Bを有する。このうち、輝度信号減算手段301A、輝度信号用係数生成手段302A、輝度信号乗算手段303A、輝度信号減算手段304A、輝度信号画面遅延手段305A、及び動き検出手段306は、入力映像信号の輝度信号Yinに対するノイズ除去装置を構成し、色差信号減算手段301B、色差信号用係数生成手段302B、色差信号乗算手段303B、色差信号減算手段304B、及び色差信号画面遅延手段305Bは、入力映像信号の色差信号Cinに対するノイズ除去装置を構成する。
Embodiment 2. FIG.
FIG. 7 is a block diagram showing the overall configuration of the noise removing apparatus according to the second embodiment of the present invention. The illustrated noise removing apparatus includes a luminance signal subtracting unit 301A, a luminance signal coefficient generating unit 302A, a luminance signal multiplying unit 303A, a luminance signal subtracting unit 304A, a luminance signal screen delay unit 305A, a motion detecting unit 306, and a color difference signal subtracting unit 301B. , Color difference signal coefficient generation means 302B, color difference signal multiplication means 303B, color difference signal subtraction means 304B, and color difference signal screen delay means 305B. Among them, the luminance signal subtraction means 301A, the luminance signal coefficient generation means 302A, the luminance signal multiplication means 303A, the luminance signal subtraction means 304A, the luminance signal screen delay means 305A, and the motion detection means 306 are the luminance signal Yin of the input video signal. The color difference signal subtracting means 301B, the color difference signal coefficient generating means 302B, the color difference signal multiplying means 303B, the color difference signal subtracting means 304B, and the color difference signal screen delay means 305B are included in the color difference signal Cin of the input video signal. Constitutes a noise removal device for

本発明の実施の形態2のノイズ除去装置に対する入力映像信号は、輝度信号と色差信号を含む。なお、他の形態のカラー映像信号が供給される場合、輝度信号及び色差信号への変換を行う手段を付加しても良い。   The input video signal for the noise removal apparatus according to the second embodiment of the present invention includes a luminance signal and a color difference signal. In addition, when a color video signal of another form is supplied, a means for converting to a luminance signal and a color difference signal may be added.

輝度信号減算手段301A、輝度信号用係数生成手段302A、輝度信号乗算手段303A、輝度信号減算手段304A、輝度信号画面遅延手段305A、及び動き検出手段306で構成される入力映像信号の輝度信号Yinに対するノイズ除去装置の構成および動作は、入力映像信号Dinを輝度信号Yinとした本発明の実施の形態1のノイズ除去装置と同じであるのでその説明を省略する。   A luminance signal subtraction unit 301A, a luminance signal coefficient generation unit 302A, a luminance signal multiplication unit 303A, a luminance signal subtraction unit 304A, a luminance signal screen delay unit 305A, and a motion detection unit 306 with respect to the luminance signal Yin of the input video signal Since the configuration and operation of the noise removal apparatus are the same as those of the noise removal apparatus according to the first embodiment of the present invention in which the input video signal Din is the luminance signal Yin, the description thereof is omitted.

一方、入力映像信号の色差信号Cinに対するノイズ除去装置の構成は、色差信号に対する動き検出手段を持たず、輝度信号用の動き検出手段306で求められた動き量Mに基づいて色差信号用の係数を生成する点のみが異なり、その他の構成は同様である。   On the other hand, the configuration of the noise removal apparatus for the color difference signal Cin of the input video signal does not have a motion detection unit for the color difference signal, and the coefficient for the color difference signal based on the motion amount M obtained by the motion detection unit 306 for the luminance signal. Only the point of generating is different, and other configurations are the same.

ここで、色差信号用係数生成手段302Bが出力する色差信号用減衰係数Kcは、輝度信号用係数生成手段302Aが出力する輝度信号用減衰係数Kyより大きくなるように定められる。   Here, the color difference signal attenuation coefficient Kc output from the color difference signal coefficient generation unit 302B is determined to be larger than the luminance signal attenuation coefficient Ky output from the luminance signal coefficient generation unit 302A.

図8(A)及び(B)は、輝度信号用係数生成手段302Aおよび色差信号用係数生成手段302Bによる係数生成の特性の一例を示す図である。図8(A)は輝度信号用係数生成手段302Aの特性、図8(B)は色差信号用係数生成手段302Bの特性を示している。   8A and 8B are diagrams illustrating an example of characteristics of coefficient generation by the luminance signal coefficient generation unit 302A and the color difference signal coefficient generation unit 302B. 8A shows the characteristics of the luminance signal coefficient generating means 302A, and FIG. 8B shows the characteristics of the color difference signal coefficient generating means 302B.

図8(A)に示す輝度信号用の特性は、式(2)において、Kmax=Kymax、Kmin=Kymin、A=0、B=(動き量Mの最大値)としたものである。図8(B)に示す色差信号用の特性は、式(2)において、Kmax=Kcmax、Kmin=Kcmin、A=0、B=(動き量Mの最大値)としたものである。ここで、Kymax≦Kcmax、Kymin<Kcminとする。   The characteristics for the luminance signal shown in FIG. 8A are obtained by setting Kmax = Kymax, Kmin = Kymin, A = 0, and B = (maximum value of the motion amount M) in the equation (2). The characteristics for the color difference signal shown in FIG. 8B are obtained by setting Kmax = Kcmax, Kmin = Kcmin, A = 0, and B = (maximum value of the motion amount M) in the equation (2). Here, Kymax ≦ Kcmax and Kymin <Kcmin.

図8(A)及び(B)に示すように、輝度信号用の減衰係数Kyに比べて、色差信号用の減衰係数Kcが大きくなるように設定されている。   As shown in FIGS. 8A and 8B, the attenuation coefficient Kc for the color difference signal is set to be larger than the attenuation coefficient Ky for the luminance signal.

以下、本発明の実施の形態2によるノイズ除去装置の作用、効果について説明する。
いま、入力映像信号のノイズの振幅がきわめて大きい場合を考える。例えば、EMCCD(Electron Multiplying Charge Coupled Device)を用いた高感度撮像装置で暗所撮影された映像は、通常、色差信号の信号量が少なく、色ノイズの振幅はきわめて大きくなっている。また、色ノイズを抑えるため、クロマサプレス処理等が施される場合もあり、色差信号の信号量はかなり小さくなる。
The operation and effect of the noise removal device according to Embodiment 2 of the present invention will be described below.
Let us consider a case where the noise amplitude of the input video signal is extremely large. For example, an image taken in a dark place with a high-sensitivity imaging device using an EMCD (Electron Multiplexing Charged Coupled Device) usually has a small amount of color difference signal and an extremely large amplitude of color noise. Further, in order to suppress color noise, chroma suppression processing or the like may be performed, and the signal amount of the color difference signal is considerably reduced.

このような映像信号の色差信号が入力された場合、ノイズ成分に対して動き成分自身が小さいため、周辺N×N画素の影響を考慮したとしても、動きを正しく検出するのは困難である。そこで、本実施の形態2のノイズ除去装置では、輝度信号を用いて求めた動き量Mに基づき、色差信号用の減衰係数Kcを生成することで、色ノイズの振幅が大きい映像信号が入力される場合においても、動きを適切に検出し、その動きに応じたノイズ除去を行うことができる。   When such a color difference signal of the video signal is input, since the motion component itself is smaller than the noise component, it is difficult to correctly detect the motion even if the influence of the surrounding N × N pixels is taken into consideration. Therefore, in the noise removal apparatus according to the second embodiment, a video signal having a large amplitude of color noise is input by generating the attenuation coefficient Kc for the color difference signal based on the amount of motion M obtained using the luminance signal. Even in the case where the motion is detected, it is possible to appropriately detect the motion and perform noise removal according to the motion.

また、上記のような入力映像信号の色差信号ではもともと動き成分が弱いため、高い係数を用いて強力にノイズを除去しても残像が目立ちにくい。そこで、本実施の形態2のノイズ除去装置では、輝度信号用の減衰係数Kyに比べて、色差信号用の減衰係数Kcを大きな値に設定することで、色ノイズを効果的に減らし、かつ残像を目立たないようにすることができる。   In addition, since the motion difference is originally weak in the color difference signal of the input video signal as described above, an afterimage is hardly noticeable even if noise is strongly removed using a high coefficient. Therefore, in the noise removal apparatus according to the second embodiment, the color noise is effectively reduced by setting the attenuation coefficient Kc for the color difference signal to a larger value than the attenuation coefficient Ky for the luminance signal, and the afterimage. Can be made inconspicuous.

さらに、本発明による実施の形態2のノイズ除去装置は、輝度信号用の動き検出手段306で求められた動き量Mに基づいて色差信号用の係数を生成する点のみが異なり、その他の構成は同様であるので、本発明による実施の形態1の撮像装置と同様の効果を有する。   Furthermore, the noise removal apparatus according to the second embodiment of the present invention is different only in that a coefficient for color difference signals is generated based on the amount of motion M obtained by the motion detection means 306 for luminance signals. Since it is the same, it has the same effect as the imaging device of Embodiment 1 by this invention.

なお、本実施の形態2では、色差信号用係数生成手段302Bは輝度信号から求めた動き量Mに応じて係数を生成するものとしたが、別途、色差信号から動き量を求める色差信号用の動き検出手段を設け、色差信号から求めた動き量に応じて色差信号用の係数を生成するように構成してもよい。色ノイズの振幅がさほど大きくない場合には、色差信号から動き信号を取り出すことが可能であり、色差信号のノイズを除去する際に、色差信号から検出した動きに応じて係数の制御を行うことで、より正確な係数設定が行える。   In the second embodiment, the color difference signal coefficient generating unit 302B generates a coefficient in accordance with the motion amount M obtained from the luminance signal. However, for the color difference signal for obtaining the motion amount from the color difference signal separately. A motion detection unit may be provided to generate a color difference signal coefficient in accordance with the amount of motion obtained from the color difference signal. When the amplitude of the color noise is not so large, it is possible to extract the motion signal from the color difference signal, and when removing the noise of the color difference signal, the coefficient is controlled according to the motion detected from the color difference signal. Thus, more accurate coefficient setting can be performed.

1 減算手段、 2 係数生成手段、 3 乗算手段、 4 減算手段、 5 画面遅延手段、 6 動き検出手段、 301A 輝度信号減算手段、 302A 輝度信号用係数生成手段、 303A 輝度信号乗算手段、 304A 輝度信号減算手段、 305A 輝度信号画面遅延手段、 306 動き検出手段、 301B 色差信号減算手段、 302B 色差信号用係数生成手段、 303A 色差信号乗算手段、 304B 色差信号減算手段、 305B 色差信号画面遅延手段、 M 動き量。   DESCRIPTION OF SYMBOLS 1 Subtraction means, 2 Coefficient generation means, 3 Multiplication means, 4 Subtraction means, 5 Screen delay means, 6 Motion detection means, 301A Luminance signal subtraction means, 302A Luminance signal coefficient generation means, 303A Luminance signal multiplication means, 304A Luminance signal Subtraction means, 305A luminance signal screen delay means, 306 motion detection means, 301B color difference signal subtraction means, 302B color difference signal coefficient generation means, 303A color difference signal multiplication means, 304B color difference signal subtraction means, 305B color difference signal screen delay means, M motion amount.

Claims (10)

1画面の映像信号と、前記1画面の映像信号の1画面前の映像信号との差を求める第1の減算手段と、
前記1画面の映像信号と、前記1画面の映像信号の1画面前の映像信号から1画面内の画像の動き量を検出する動き検出手段と、
前記動き量の大きさに応じて可変の減衰係数を生成する係数生成手段と、
前記第1の減算手段の出力信号に対して、前記減衰係数を乗ずる乗算手段と、
前記減衰係数を乗じた出力信号を、前記1画像面の映像信号から差し引く第2の減算手段とを具備し、
前記係数生成手段において、前記減衰係数は、前記動き検出手段によって検出された動き量が大きくなるほど小さくなるように定められ、かつ、
前記動き量の変化量に対する前記減衰係数の変化量の比の絶対値は、前記動き量の一部または全ての区間において、前記動き量が大きくなるほど小さくなるように定められ
前記係数生成手段は、
前記動き量をMとし、
前記減衰係数をKとし、
予め定められた前記減衰係数の最大値をKmaxとし、
予め定められた前記減衰係数の最小値をKminとし、
予め定められた静止画領域と判断する前記動き量の閾値をAとし、
予め定められた動画領域と判断する前記動き量の閾値をBとしたとき、次式
Figure 0005235807
の関係を満たす前記減衰係数を生成す
ことを特徴とするノイズ除去装置。
First subtracting means for obtaining a difference between a video signal of one screen and a video signal of one screen before the one-screen video signal;
Motion detection means for detecting a motion amount of an image in one screen from the video signal of the one screen and a video signal of one screen before the video signal of the one screen;
Coefficient generating means for generating a variable attenuation coefficient according to the magnitude of the amount of motion;
Multiplying means for multiplying the output signal of the first subtracting means by the attenuation coefficient;
A second subtracting means for subtracting the output signal multiplied by the attenuation coefficient from the video signal of the one image plane;
In the coefficient generation means, the attenuation coefficient is determined so as to decrease as the amount of motion detected by the motion detection means increases, and
The absolute value of the ratio of the change amount of the attenuation coefficient to the change amount of the motion amount is determined so as to decrease as the motion amount increases in a part or all of the motion amount ,
The coefficient generating means includes
The amount of movement is M,
The attenuation coefficient is K,
The maximum value of the predetermined attenuation coefficient is Kmax,
A predetermined minimum value of the attenuation coefficient is defined as Kmin,
A threshold value of the amount of motion that is determined as a predetermined still image area is A,
When the threshold value of the amount of motion to be determined as a predetermined moving image area is B, the following formula
Figure 0005235807
Noise removal device, characterized in that that generates the attenuation coefficient that satisfies the relationship.
1画面の映像信号と、前記1画面の映像信号の1画面前の映像信号との差を求める第1の減算手段と、
前記1画面の映像信号と、前記1画面の映像信号の1画面前の映像信号から1画面内の画像の動き量を検出する動き検出手段と、
前記動き量の大きさに応じて可変の減衰係数を生成する係数生成手段と、
前記第1の減算手段の出力信号に対して、前記減衰係数を乗ずる乗算手段と、
前記減衰係数を乗じた出力信号を、前記1画像面の映像信号から差し引く第2の減算手段とを具備し、
前記係数生成手段において、前記減衰係数は、前記動き検出手段によって検出された動き量が大きくなるほど小さくなるように定められ、かつ、
前記動き量の変化量に対する前記減衰係数の変化量の比の絶対値は、前記動き量の一部または全ての区間において、前記動き量が大きくなるほど小さくなるように定められ、
前記係数生成手段は、
前記動き量をMとし、
前記減衰係数をKとし、
予め定められた前記減衰係数の最大値をKmaxとし、
予め定められた前記減衰係数の最小値をKminとし、
予め定められた前記動き量の最大値をMmaxとしたとき、次式
Figure 0005235807
の関係を満たす前記減衰係数を生成することを特徴とするノイズ除去装置。
First subtracting means for obtaining a difference between a video signal of one screen and a video signal of one screen before the one-screen video signal;
Motion detection means for detecting a motion amount of an image in one screen from the video signal of the one screen and a video signal of one screen before the video signal of the one screen;
Coefficient generating means for generating a variable attenuation coefficient according to the magnitude of the amount of motion;
Multiplying means for multiplying the output signal of the first subtracting means by the attenuation coefficient;
A second subtracting means for subtracting the output signal multiplied by the attenuation coefficient from the video signal of the one image plane;
In the coefficient generation means, the attenuation coefficient is determined so as to decrease as the amount of motion detected by the motion detection means increases, and
The absolute value of the ratio of the change amount of the attenuation coefficient to the change amount of the motion amount is determined so as to decrease as the motion amount increases in a part or all of the motion amount,
The coefficient generating means includes
The amount of movement is M,
The attenuation coefficient is K,
The maximum value of the predetermined attenuation coefficient is Kmax,
A predetermined minimum value of the attenuation coefficient is defined as Kmin,
When the predetermined maximum amount of motion is Mmax, the following formula
Figure 0005235807
Noise removal device configured to generate the attenuation coefficient that satisfies the relationship you characterized.
前記動き検出手段は、
前記1画面の映像信号について、動き検出の対象となる画素を、その中心又は該中心の近傍に含む領域内の画素の平均値を求める第1の周辺画素平均値算出手段と、
前記1画面の映像信号の1画面前の映像信号について、動き検出の対象となる画素を、その中心又は該中心の近傍に含む領域内の画素の平均値を求める第2の周辺画素平均値算出手段と、
前記第1の周辺画素平均値算出手段の出力と前記第2の周辺画素平均値算出手段の出力の差分を求める第3の減算手段を有する
ことを特徴とする請求項1又は2に記載のノイズ除去装置。
The motion detection means includes
A first peripheral pixel average value calculating means for calculating an average value of pixels in a region including a pixel that is a target of motion detection in the center or in the vicinity of the center of the one-screen video signal;
Second peripheral pixel average value calculation for obtaining an average value of pixels in a region including a pixel that is a target of motion detection at the center or in the vicinity of the center of the video signal one screen before the one-screen video signal Means,
Noise according to claim 1 or 2, characterized in that it has a third subtraction means for obtaining a difference between outputs of said second peripheral pixel average value calculation means of the first peripheral pixel average value calculation means Removal device.
前記動き検出の対象となる画素を、その中心又は該中心の近傍に含む領域が、N×N画素(Nは2以上の整数)から成る領域であることを特徴とする請求項に記載のノイズ除去装置。 The pixel of interest of the movement detection, the area including the vicinity of the center or said center is, N × N pixels (N is an integer of 2 or more) according to claim 3, characterized in that the region consisting of Noise removal device. 輝度信号と色差信号を含む映像信号のノイズを除去するノイズ除去装置であって、
1画面の輝度信号と、前記1画面の輝度信号の1画面前の輝度信号との差を求める第1の輝度信号減算手段と、
前記1画面の輝度信号と、前記1画面の輝度信号の1画面前の輝度信号から1画面内の画像の動き量を検出する動き検出手段と、
前記動き量の大きさに応じて可変の輝度信号用減衰係数を生成する輝度信号用係数生成手段と、
前記第1の輝度信号減算手段の出力信号に対して、前記輝度信号用減衰係数を乗ずる輝度信号乗算手段と、
前記輝度信号用減衰係数を乗じた出力信号を、前記1画像面の輝度信号から差し引く第2の輝度信号減算手段と、
前記1画面の輝度信号と同じ1画面の色差信号と、前記1画面の色差信号の1画面前の色差信号との差を求める第1の色差信号減算手段と、
前記動き検出手段で検出された動き量の大きさに応じて可変の色差信号用減衰係数を生成する色差信号用係数生成手段と、
前記第1の色差信号減算手段の出力信号に対して、前記色差信号用減衰係数を乗ずる色差信号乗算手段と、
前記色差信号用減衰係数を乗じた出力信号を、前記1画像面の色差信号から差し引く第2の色差信号減算手段を具備し、
前記輝度信号用係数生成手段において、前記輝度信号用減衰係数は、前記動き検出手段によって検出された動き量が大きくなるほど小さくなるように定められ、かつ、
前記動き量の変化量に対する前記輝度信号用減衰係数の変化量の比の絶対値は、前記動き量の一部または全ての区間において、前記動き量が大きくなるほど小さくなるように定められ、
前記色差信号用係数生成手段において、前記色差信号用減衰係数は、前記動き検出手段によって検出された動き量が大きくなるほど小さくなるように定められ、かつ、
前記動き量の変化量に対する前記色差信号用減衰係数の変化量の比の絶対値は、前記動き量の一部または全ての区間において、前記動き量が大きくなるほど小さくなるように定められ
前記色差信号用係数生成手段が生成する色差信号用減衰係数は、
前記輝度信号用係数生成手段が同じ動き量に対して生成する輝度信号用減衰係数よりも大きい
ことを特徴とするノイズ除去装置。
A noise removing device for removing noise of a video signal including a luminance signal and a color difference signal,
First luminance signal subtracting means for obtaining a difference between a luminance signal of one screen and a luminance signal of one screen before the luminance signal of the one screen;
Motion detection means for detecting a motion amount of an image in one screen from the luminance signal of the one screen and a luminance signal one screen before the luminance signal of the one screen;
Luminance signal coefficient generation means for generating a variable luminance signal attenuation coefficient according to the magnitude of the amount of motion;
Luminance signal multiplying means for multiplying the output signal of the first luminance signal subtracting means by the luminance signal attenuation coefficient;
A second luminance signal subtracting means for subtracting the output signal multiplied by the luminance signal attenuation coefficient from the luminance signal of the one image plane;
First color difference signal subtracting means for obtaining a difference between a color difference signal of the same screen as the luminance signal of the one screen and a color difference signal of the previous screen of the color difference signal of the one screen;
Color difference signal coefficient generating means for generating a variable color difference signal attenuation coefficient according to the amount of motion detected by the motion detection means;
Chrominance signal multiplication means for multiplying the output signal of the first chrominance signal subtracting means by the chrominance signal attenuation coefficient;
A second color difference signal subtracting means for subtracting an output signal multiplied by the color difference signal attenuation coefficient from the color difference signal of the one image plane;
In the luminance signal coefficient generating means, the luminance signal attenuation coefficient is determined so as to decrease as the amount of motion detected by the motion detecting means increases, and
The absolute value of the ratio of the change amount of the luminance signal attenuation coefficient to the change amount of the motion amount is determined so as to decrease as the motion amount increases in a part or all of the motion amount,
In the color difference signal coefficient generation means, the color difference signal attenuation coefficient is determined so as to decrease as the amount of motion detected by the motion detection means increases, and
The absolute value of the ratio of the change amount of the color difference signal attenuation coefficient to the change amount of the motion amount is determined so as to decrease as the motion amount increases in a part or all of the motion amount ,
The color difference signal attenuation coefficient generated by the color difference signal coefficient generation means is:
The noise removing apparatus, wherein the luminance signal coefficient generating means is larger than a luminance signal attenuation coefficient generated for the same amount of motion .
前記動き検出手段は、
前記1画面の輝度信号について、動き検出の対象となる画素を、その中心又は該中心の近傍に含む領域内の画素の平均値を求める第1の周辺画素平均値算出手段と、
前記1画面の輝度信号の1画面前の映像信号について、動き検出の対象となる画素を、その中心又は該中心の近傍に含む領域内の画素の平均値を求める第2の周辺画素平均値算出手段と、
前記第1の周辺画素平均値算出手段の出力と前記第2の周辺画素平均値算出手段の出力の差分を求める第3の減算手段を有する
ことを特徴とする請求項に記載のノイズ除去装置。
The motion detection means includes
A first peripheral pixel average value calculating means for calculating an average value of pixels in a region including a pixel that is a target of motion detection at the center or in the vicinity of the center of the luminance signal of the one screen;
Second peripheral pixel average value calculation for obtaining an average value of pixels in a region including a pixel as a motion detection target at the center or in the vicinity of the center of the video signal one screen before the luminance signal of the one screen Means,
6. The noise removing apparatus according to claim 5 , further comprising third subtracting means for obtaining a difference between the output of the first peripheral pixel average value calculating means and the output of the second peripheral pixel average value calculating means. .
前記動き検出の対象となる画素を、その中心又は該中心の近傍に含む領域が、N×N画素(Nは2以上の整数)から成る領域であることを特徴とする請求項に記載のノイズ除去装置。 The pixel of interest of the movement detection, the area including the vicinity of the center or said center is, N × N pixels (N is an integer of 2 or more) according to claim 6, characterized in that the region consisting of Noise removal device. 前記輝度信号用係数生成手段は、
前記動き量をMとし、
前記輝度信号用減衰係数をKyとし、
予め定められた前記輝度信号用減衰係数の最大値をKymaxとし、
予め定められた前記輝度信号用減衰係数の最小値をKyminとし、
予め定められた静止画領域と判断する前記動き量の閾値をAとし、
予め定められた動画領域と判断する前記動き量の閾値をBとしたとき、次式
Figure 0005235807
の関係を満たす前記輝度信号用減衰係数を生成し、
前記色差信号用係数生成手段は、
前記動き量をMとし、
前記色差信号用減衰係数をKcとし、
予め定められた前記色差信号用減衰係数の最大値をKcmaxとし、
予め定められた前記色差信号用減衰係数の最小値をKcminとし、
予め定められた静止画領域と判断する前記動き量の閾値をAとし、
予め定められた動画領域と判断する前記動き量の閾値をBとしたとき、次式
Figure 0005235807
の関係を満たす前記色差信号用減衰係数を生成する
ことを特徴とする請求項5乃至7のいずれかに記載のノイズ除去装置。
The luminance signal coefficient generating means includes:
The amount of movement is M,
The luminance signal attenuation coefficient is Ky,
The predetermined maximum value of the luminance signal attenuation coefficient is Kymax,
The predetermined minimum value of the luminance signal attenuation coefficient is Kymin,
A threshold value of the amount of motion that is determined as a predetermined still image area is A,
When the threshold value of the amount of motion to be determined as a predetermined moving image area is B, the following formula
Figure 0005235807
Generating the luminance signal attenuation coefficient satisfying the relationship:
The color difference signal coefficient generating means includes:
The amount of movement is M,
The attenuation coefficient for the color difference signal is Kc,
The predetermined maximum value of the color difference signal attenuation coefficient is Kcmax,
The predetermined minimum value of the color difference signal attenuation coefficient is Kcmin,
A threshold value of the amount of motion that is determined as a predetermined still image area is A,
When the threshold value of the amount of motion to be determined as a predetermined moving image area is B, the following formula
Figure 0005235807
The noise removal apparatus according to claim 5, wherein the color difference signal attenuation coefficient satisfying the relationship is generated.
前記輝度信号用係数生成手段は、
前記動き量をMとし、
前記輝度信号用減衰係数をKyとし、
予め定められた前記輝度信号用減衰係数の最大値をKymaxとし、
予め定められた前記輝度信号用減衰係数の最小値をKyminとし、
予め定められた前記動き量の最大値をMmaxとしたとき、次式
Figure 0005235807
の関係を満たす前記輝度信号用減衰係数を生成し、
前記色差信号用係数生成手段は、
前記動き量をMとし、
前記色差信号用減衰係数をKcとし、
予め定められた前記色差信号用減衰係数の最大値をKcmaxとし、
予め定められた前記色差信号用減衰係数の最小値をKcminとし、
予め定められた前記動き量の最大値をMmaxとしたとき、次式
Figure 0005235807
の関係を満たす前記色差信号用減衰係数を生成する
ことを特徴とする請求項5乃至7のいずれかに記載のノイズ除去装置。
The luminance signal coefficient generating means includes:
The amount of movement is M,
The luminance signal attenuation coefficient is Ky,
The predetermined maximum value of the luminance signal attenuation coefficient is Kymax,
The predetermined minimum value of the luminance signal attenuation coefficient is Kymin,
When the predetermined maximum amount of motion is Mmax, the following formula
Figure 0005235807
Generating the luminance signal attenuation coefficient satisfying the relationship:
The color difference signal coefficient generating means includes:
The amount of movement is M,
The attenuation coefficient for the color difference signal is Kc,
The predetermined maximum value of the color difference signal attenuation coefficient is Kcmax,
The predetermined minimum value of the color difference signal attenuation coefficient is Kcmin,
When the predetermined maximum amount of motion is Mmax, the following formula
Figure 0005235807
The noise removal apparatus according to claim 5, wherein the color difference signal attenuation coefficient satisfying the relationship is generated.
前記輝度信号用減衰係数の最大値Kymax、前記輝度信号用減衰係数の最小値Kymin、前記色差信号用減衰係数の最大値Kcmax、及び前記色差信号用減衰係数の最小値Kcminは、
Kymax≦Kcmax、Kymin<Kcmin
の関係を満たすことを特徴とする請求項に記載のノイズ除去装置。
The luminance signal attenuation coefficient maximum value Kymax, the luminance signal attenuation coefficient minimum value Kymin, the color difference signal attenuation coefficient maximum value Kcmax, and the color difference signal attenuation coefficient minimum value Kcmin are:
Kymax ≦ Kcmax, Kymin <Kcmin
The noise removal device according to claim 9 , wherein the relationship is satisfied.
JP2009172270A 2009-07-23 2009-07-23 Noise removal device Active JP5235807B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2009172270A JP5235807B2 (en) 2009-07-23 2009-07-23 Noise removal device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP2009172270A JP5235807B2 (en) 2009-07-23 2009-07-23 Noise removal device

Publications (3)

Publication Number Publication Date
JP2011029826A JP2011029826A (en) 2011-02-10
JP2011029826A5 JP2011029826A5 (en) 2012-08-23
JP5235807B2 true JP5235807B2 (en) 2013-07-10

Family

ID=43638080

Family Applications (1)

Application Number Title Priority Date Filing Date
JP2009172270A Active JP5235807B2 (en) 2009-07-23 2009-07-23 Noise removal device

Country Status (1)

Country Link
JP (1) JP5235807B2 (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6024876B2 (en) 2012-03-26 2016-11-16 パナソニックIpマネジメント株式会社 Noise removal device

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2934036B2 (en) * 1991-03-07 1999-08-16 松下電器産業株式会社 Motion detection method and noise reduction device
JPH0647035A (en) * 1992-07-31 1994-02-22 Shimadzu Corp Image processing system
JP3704740B2 (en) * 1994-10-17 2005-10-12 ソニー株式会社 Quantization apparatus and quantization method
DE10109968A1 (en) * 2001-03-01 2002-09-12 Philips Corp Intellectual Pty Device for noise reduction in image signals
EP1515544A4 (en) * 2002-06-25 2008-03-12 Matsushita Electric Ind Co Ltd Motion detection device and noise reduction device using that
JP4625705B2 (en) * 2005-02-24 2011-02-02 Okiセミコンダクタ株式会社 Video signal processing device
JP2007288595A (en) * 2006-04-18 2007-11-01 Pioneer Electronic Corp Frame circulation noise reduction device

Also Published As

Publication number Publication date
JP2011029826A (en) 2011-02-10

Similar Documents

Publication Publication Date Title
JP4670756B2 (en) Noise detection apparatus and method, and noise reduction apparatus and method
JP4273428B2 (en) Image processing apparatus, image processing method, program for image processing method, and recording medium recording program for image processing method
US8218083B2 (en) Noise reducer, noise reducing method, and video signal display apparatus that distinguishes between motion and noise
US7483081B2 (en) Edge compensated feature detector and method thereof
US8571348B2 (en) Apparatus and method for image noise reduction based on weighted pixel difference
KR101125977B1 (en) Image processing apparatus and method, recording medium
JP2006081150A (en) Image processing apparatus and method, recording medium, and program
JP2009130537A (en) Image processing device and image processing method
EP1515544A1 (en) Motion detection device and noise reduction device using that
JP4358283B1 (en) Noise reduction device and noise reduction method
JP2002010106A (en) Device and method for reducing noise and medium storing noise reduction program
JPWO2006025396A1 (en) Image processing apparatus and image processing program
JP4328115B2 (en) Solid-state imaging device
JP3240371B2 (en) Video signal processing device
JP5235807B2 (en) Noise removal device
JP4679501B2 (en) Noise removing device, noise removing method, and video signal display device
JP5147655B2 (en) Video signal processing device and video display device
JPH10262160A (en) Signal-to-noise ratio (s/n)detector and noise reduction device
JP4633437B2 (en) Time recursive color signal noise reduction method and apparatus
JP2010200236A (en) Lateral noise correcting method
JP5559275B2 (en) Image processing apparatus and control method thereof
JP5219771B2 (en) Video processing apparatus and video processing apparatus control method
JP6031993B2 (en) Image processing apparatus and image processing method
JP5111310B2 (en) Image processing method and image processing apparatus
JP2010245998A (en) Frame circulation type noise reduction device, and frame circulation type noise reduction method

Legal Events

Date Code Title Description
A521 Written amendment

Free format text: JAPANESE INTERMEDIATE CODE: A523

Effective date: 20120709

A621 Written request for application examination

Free format text: JAPANESE INTERMEDIATE CODE: A621

Effective date: 20120709

A977 Report on retrieval

Free format text: JAPANESE INTERMEDIATE CODE: A971007

Effective date: 20121130

A131 Notification of reasons for refusal

Free format text: JAPANESE INTERMEDIATE CODE: A131

Effective date: 20121218

A521 Written amendment

Free format text: JAPANESE INTERMEDIATE CODE: A523

Effective date: 20130207

TRDD Decision of grant or rejection written
A01 Written decision to grant a patent or to grant a registration (utility model)

Free format text: JAPANESE INTERMEDIATE CODE: A01

Effective date: 20130226

A61 First payment of annual fees (during grant procedure)

Free format text: JAPANESE INTERMEDIATE CODE: A61

Effective date: 20130326

R150 Certificate of patent or registration of utility model

Free format text: JAPANESE INTERMEDIATE CODE: R150

Ref document number: 5235807

Country of ref document: JP

Free format text: JAPANESE INTERMEDIATE CODE: R150

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20160405

Year of fee payment: 3

R250 Receipt of annual fees

Free format text: JAPANESE INTERMEDIATE CODE: R250

R250 Receipt of annual fees

Free format text: JAPANESE INTERMEDIATE CODE: R250

R250 Receipt of annual fees

Free format text: JAPANESE INTERMEDIATE CODE: R250

R250 Receipt of annual fees

Free format text: JAPANESE INTERMEDIATE CODE: R250

R250 Receipt of annual fees

Free format text: JAPANESE INTERMEDIATE CODE: R250