JP2010081463A - Image pickup device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image pickup device for effectively removing the noise of a photographed video. <P>SOLUTION: Orthogonal conversion is performed to an input video signal, and a frequency conversion signal is generated, and the frequency conversion signal is delayed, and a frame or field delay signal is output. Correlativeness between the frequency conversion signal and the frame or field delay signal is searched, and filter processing is performed to the frequency conversion signal, and the processing content of the filter processing is controlled by using the output of the correlativeness generation means. Inversion is performed to the output of the filtering means, and filter processing is performed while adaptively changing a parameter according to the correlativeness between the frequency conversion signal and the frame or field delay signal, and inversion is performed to the processing result, and the processing result is output. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an imaging apparatus applicable to a video recording system having an imaging apparatus and an information processing apparatus.

  Patent Document 1 states that “[Purpose] A frame recursive noise removal apparatus using orthogonal transform achieves a higher S / N improvement effect than before without causing image quality degradation. [Configuration] Frame recursive noise In the removing apparatus, the frame difference signal is decomposed into a plurality of spatial frequency components by the Hadamard transform circuit 105, and then subjected to nonlinear processing by the nonlinear processing circuit 106, synthesized by the Hadamard inverse transform circuit 107, and synthesized on the original time axis. In addition, a motion detection circuit 108 that uses a two-dimensional low-frequency component of the Hadamard transform components is provided to detect a motion component from the frame difference signal, and the coefficient circuit 109 detects the detection result of the motion detection circuit 108. The coefficient is controlled, and the coefficient is multiplied by the output signal of Hadamard inverse transform to obtain a noise component, which is subtracted from the input video signal. That. In the above configuration, can be noise extraction in accordance with the characteristics of the input video signal, it is possible to obtain a high S / N improving effect without causing deterioration of image quality. "Is described as.

JP-A-8-79571

  In recent years, digital video recording systems have become more common due to higher speeds of semiconductor devices, miniaturization of image sensors, development of digital image processing techniques, and the like. This can also be seen from the fact that, for example, consumer video cameras and surveillance camera systems have become familiar. Under such circumstances, there is an increasing demand for higher quality video. Although there are various viewpoints for improving the quality, how to remove noise mixed in the video due to a number of factors is a particularly important issue.

  Noise mixed in video can be broadly divided into random noise whose position and intensity are randomly generated in the image, fixed pattern noise that appears fixedly in the image, and spike noise that appears sporadically. Of these, random noise is a major factor that causes viewers to flicker or feel rough, resulting in a reduction in image quality. As a method for removing this random noise, a frame recursive noise removing apparatus that uses correlation in the time direction of video is generally used.

  Further, as an improved version of this type of noise removal apparatus, a noise removal apparatus that combines Hadamard transform, which is a kind of orthogonal transform, is disclosed.

  An example of a block diagram of the basic configuration of the noise removal apparatus is shown in FIG. The noise removing apparatus includes subtracters 101 and 102, a frame memory 103, a line memory 104, a Hadamard transform circuit 105, a nonlinear process 106, an Hadamard inverse transform circuit 107, a motion detection circuit 108, and a coefficient circuit 109.

  The video signal input to the noise removing device is first subtracted from the frame delay signal stored in the frame memory 103 in the subtractor 101 to generate a differential signal. The generated difference signal holds data for a plurality of lines by the line memory 104, and the held data for the plurality of lines and the difference signal data are supplied to the Hadamard transform circuit 105 in parallel. In the Hadamard transform circuit 105, Hadamard transform processing is performed for each processing unit such as 4 pixels × 2 lines. The eight frequency component signals generated by this calculation are each subjected to nonlinear processing such as limit processing in the nonlinear processing 106. The processing result is subjected to inverse transformation processing in the Hadamard inverse transformation circuit 107. The lowest order signal (DC component) of the eight frequency component signals is also supplied to the motion detection circuit 108. The motion detection circuit 108 determines whether there is motion in the original image before frequency conversion based on the input DC component. The determination result is supplied to the coefficient circuit 109. The coefficient circuit 109 calculates a weighting coefficient based on the determination result, multiplies the result by the inverse transformation processing result, and supplies the result to the subtracter 101. The subtracter 101 generates an output signal by subtracting the output of the coefficient circuit 109 from the input signal.

  By such an operation, the noise component is extracted in the non-linear processing 106, and the noise can be removed by subtracting it from the input signal. Further, the motion detection by the motion detection circuit 108 enables noise removal at a ratio suitable for the portion of the input signal that is moving and the portion that is stationary.

  In the frame cyclic noise removing apparatus and the improved frame cyclic noise removing apparatus as disclosed in Patent Document 1, random noise is removed by adding the data of the previous frame to the data of the current frame. However, an afterimage such as a double copy remains particularly in a place where there is a movement in the image. In order to eliminate such an adverse effect, a moving part and a stationary part in the image are discriminated, and the moving part is reduced in the ratio of adding the data of the previous frame (hereinafter referred to as a circulation coefficient) to reduce the afterimage. Control is performed to prevent random noise by increasing the circulation coefficient in places where there is no movement. At that time, as shown in Patent Document 1, it is possible to improve the discrimination efficiency by devising such as using Hadamard transform to discriminate between a moving part and a stationary part. However, no consideration has been given to the point that it is difficult to discriminate a moving part with high accuracy.

  In particular, in a digital image imaging system, it is necessary to perform processing by dividing the three primary colors of light (R (red), G (green), and B (blue)), but by separately processing RGB, No consideration was given to the possibility of generating false colors or color noise that should not exist.

  Further, in a consumer digital image pickup system, it is common to perform edge enhancement processing to add high-frequency components lost by signal processing after the noise removal step, and it is necessary to provide a circuit for this purpose separately. No consideration has been given to an increase in processing load in software mounting and an increase in circuit scale and power consumption in hardware mounting such as LSI.

  An object of the present invention is to provide an imaging apparatus capable of effectively removing noise of a captured video.

  In the imaging apparatus according to the present invention, a conversion unit that performs orthogonal transformation on an input video signal to the imaging system to generate a frequency conversion signal, and a delay that delays the frequency conversion signal and outputs a frame or field delay signal A correlation degree generating means for obtaining a correlation between the frequency converted signal and the frame or field delay signal, a filtering means for performing a filtering process on the frequency converted signal, and an output of the correlation degree generating means. Control means for controlling the processing contents of the filter processing and inverse conversion means for performing inverse conversion on the output of the filtering means are provided. Then, the frequency conversion signal is subjected to a filter process while adaptively changing parameters according to the correlation with the frame or field delay signal, and the process result is inversely converted and output.

  ADVANTAGE OF THE INVENTION According to this invention, the imaging device which can remove effectively the noise of the image | photographed image | video can be provided.

  A description will be given below with reference to the drawings.

  A first embodiment of the present invention is shown in FIG. FIG. 1 shows an image pickup apparatus according to the present invention in the form of a video camera. The video camera of the present embodiment includes an optical system 101, an image sensor 102, a demosaicing processing unit 103, frequency conversion units 104 to 106, noise reduction processing units 107 to 109, a frame memory 110, frequency inverse conversion units 111 to 113, and luminance. A signal generation unit 114, a color difference signal generation unit 115, an encoding unit 116, and a recording medium 117 are provided. Hereinafter, the operation of the video camera of this embodiment will be described in order.

  Light emitted from a subject not shown is first condensed by the optical system 101. Although details regarding the optical system 101 are not shown, the optical system 101 includes a plurality of lenses, a motor that moves the lens for focusing and magnification adjustment, a control system that controls the motor, and the like.

  The light condensed by the optical system 101 is converted into an electric signal by the image sensor 102. The imaging device 102 is driven by an imaging unit such as a CCD or CMOS that is actually exposed, an AD conversion circuit that samples an analog electrical signal and converts it into a digital electrical signal, an AGC circuit that performs gain adjustment according to the amount of ambient light, and an imaging device For example, a timing generator for supplying timing signals and a control circuit for controlling them.

  The video signal converted into a digital electric signal by the image sensor 102 is processed by the demosaicing processing unit 103 and outputs an R component signal, a G component signal, and a B component signal. For example, when the imaging unit used in the imaging device 102 is a Bayer array type image sensor, the R component signal, the G component signal, and the B component signal are not collected at all sampling positions. In such a case, the demosaicing processing unit 103 generates and outputs a signal of a component that is insufficient at the sampling position by appropriate processing such as pixel interpolation from surrounding pixels. Note that depending on the type of the imaging unit to be used, there are cases where the R component signal, the G component signal, and the B component signal are all available at the time of output from the imaging element 102. In such a case, the demosaicing processing unit 103 does not have to exist.

  The signal of each color component output from the demosaicing processing unit 103 is input to another frequency conversion unit 104 to 106, and frequency conversion is performed for each pixel unit (for example, 4 pixels × 4 pixels, 8 pixels × 8 pixels, etc.). Is done. The frequency transform is, for example, discrete cosine transform (DCT), Hadamard transform, wavelet transform, or the like.

  The frequency-converted color component signals are supplied to noise reduction processing units 107 to 109 provided for the respective colors. The noise reduction processing unit is also supplied with a color component signal after frequency conversion of a past frame already stored in the frame memory 110. Specific contents of the noise reduction processing in the noise reduction processing units 107 to 109 will be described later.

  Each color component signal subjected to the noise reduction processing is stored in the frame memory 110, and is supplied to the frequency inverse transform units 111 to 113 in parallel with this. The frequency inverse transform unit 113 performs an inverse operation of the frequency transform process performed in the frequency transform units 104 to 106 to restore the color component signal converted to the frequency domain to the original spatial domain signal.

  Each color component signal restored to the spatial region is supplied to the luminance signal generation unit 114 and the color difference signal generation unit 115. The luminance signal generation unit 114 applies a calculation method such as matrix calculation to the supplied signal of each color component to generate a luminance signal. The color difference signal generation unit 115 similarly generates a color difference signal.

  The generated luminance signal and color difference signal are supplied to the encoding unit 115. In the encoding unit 115, for example, H.264 is applied to the input luminance signal and color difference signal. Compression encoding processing such as H.264 or MPEG-2 is performed. Also, a multiplex process is performed on the compressed audio signal generated by the audio collecting unit, the audio signal preprocessing unit, and the audio signal encoding unit (not shown) and the compressed video signal generated by the compression encoding process. It is recorded on the recording medium 117 as compressed video / audio data.

  Processing of the noise reduction processing units 107 to 109 will be described. As an example, the processing in the noise reduction processing unit 107 will be described in detail, but the internal configuration and operation are the same in the noise reduction processing units 108 and 109 as well.

  The noise reduction processing unit 107 includes a subtractor 201, an enhancement coefficient calculation unit 202, a core coefficient calculation unit 203, a coring processing group 204, an enhancement processing group 205, a circulation coefficient calculation unit 206, a multiplier 207, and an adder 208. Configured.

  The R component signal input from the frequency conversion unit 104 is first subtracted from the R component signal 110 stored in the frame memory and delayed by one frame in the subtractor 201. The generated difference signal is supplied to the enhancement coefficient calculation unit 202, the core coefficient calculation unit 203, and the circulation coefficient generation unit 206, and is necessary for the circulation coefficient, coring processing, and enhancement processing necessary for noise removal processing described later. To calculate various parameters. The parameters can be set individually for each frequency component.

  First, the circulation coefficient generation unit 206 generates a circulation coefficient according to the absolute value of the low-frequency component in the subtraction processing result. The generated circulation coefficient is multiplied by the subtraction processing result by the multiplier 207 and then added to the R component signal input from the frequency conversion unit 104. Since random noise in a still image is evenly distributed over all frequency components and there is no correlation between frames, this processing can reduce random noise. In addition, by performing frequency conversion, for example, the edge portion of a moving object can be detected more sensitively, so that the moving portion and the stationary portion can be distinguished more accurately than in the past. .

  The coring process group 204 includes a plurality of coring processes. The coring process is performed individually for each frequency component. FIG. 6 shows the relationship between input and output in the coring process. In the coring processing unit, for example, mapping from an input signal to an output signal is performed based on the following four parameters. (1) Coring positive influence range (point a in FIG. 6), (2) Coring negative influence range (point b in FIG. 6), (3) Coring positive inclination (figure 6) 6 (c / a) in FIG. 6 and (4) inclination on the negative side of the coring (d / b in FIG. 6).

  By applying frequency conversion to the input image, it is possible to separate the input image into a high frequency component (random noise component, subject edge component) and a low frequency component (such as a portion that is not the subject edge). By taking the difference between frames for the frequency-converted video signal, the movement of the subject and the random noise can be separated efficiently.

  A parameter determination method will be described. First, the influence range (a, c) of coring can be determined by the noise level. Random noise characteristics include a certain distribution within a certain level (uniform distribution and Gaussian distribution), a relatively low level itself, and uniform inclusion in all frequency components. For this reason, the level of noise included in the input image is checked, and a and c are determined accordingly. Further, due to the above-described characteristics, it is usually set so that c = −a.

  The degree of noise removal is determined by the inclination (c / a, d / b) of the coring region. That is, the smaller the inclination (closer to the horizontal), the greater the noise removal effect. In the present invention, the coring parameters can be set individually for each frequency component. However, since random noise components appear uniformly in each frequency component, the same coring is applied to all frequency components. It is common to set the influence range setting and the inclination of the coring area.

  By applying the coring process as described above, the random noise component is suppressed, and the low frequency component and the high frequency component having a large level (that is, the edge component of the object) are output as they are, so that the image is blurred. Thus, random noise can be suppressed.

  The video signal whose random noise component has been reduced by the coring processing group 204 is subsequently input to the enhancement processing group 205. Similar to the coring process, the enhancement process can be performed individually for each frequency component. FIG. 7 shows the relationship between input and output in the enhancement processing. The enhancement process parameters are determined as follows, for example. (1) Influence range on the positive side of enhancement (point e in FIG. 7), (2) Influence range on the negative side of enhancement (point f in FIG. 7), (3) Inclination on the positive side of enhancement (g in FIG. 7) / E), (4) the negative inclination of the coring (h / f in FIG. 7).

  By calculating and setting these for each frequency component, it is possible to enhance the appearance of the image by emphasizing a specific frequency component, which was difficult with conventional filtering. In general, the low-frequency component does not impair the naturalness of the image by setting the enhancement gradient to be small, and the visual characteristics are improved by applying a strong enhancement process to the mid-range component. If the high-frequency components are emphasized too much, the roughness of the image may have a visual adverse effect. Therefore, the enhancement processing should be suppressed.

  As described so far, by performing the coring process and the enhancement process separately for each frequency component, it is possible to remove random noise with extremely high accuracy.

  Further, as an input / output characteristic having both characteristics of coring processing and enhancement processing, there is a characteristic as shown in FIG. By setting the boundary point of coring processing and enhancement processing as parameters and the slope of each part as parameters, it is possible to perform both processing simultaneously, reducing processing time in software processing and reducing circuit scale in hardware processing. It is possible to realize power reduction.

  A second embodiment of the present invention is shown in FIG. FIG. 3 shows an image pickup apparatus according to the present invention implemented in the form of a video camera. The video camera of this embodiment includes an optical system 101, an image sensor 102, a demosaicing processing unit 103, a frequency conversion unit 104, a noise reduction processing unit 107, a frame memory 110, a frequency inverse conversion unit 111, a luminance signal generation unit 114, a color difference. A signal generation unit 115, an encoding unit 116, and a recording medium 117 are included. Hereinafter, the operation of the video camera of this embodiment will be described in order.

  Since the second embodiment of the present invention is almost the same as the first embodiment, only the differences will be described. First, the video signal captured by the optical system 101 and the image sensor 102 is input to the frequency conversion unit 104 with RGB color mixture. The frequency conversion unit 104 performs frequency conversion in the same manner as in the first embodiment. At this time, the frequency conversion is performed using the values of surrounding pixels of the same color. For example, when a Bayer array image sensor is used, the first line is arranged as R component, G component, R component, G component,..., And the second line is G component, B component, G component, B component,. It is arranged like…. The third line has the same arrangement as the first line. Therefore, for example, when frequency conversion is performed for the R component, the frequency is returned by using the center R component pixel and the R component surrounding the pixel (horizontal and vertical adjacent pixels and three pixels apart). Shall be applied. The same applies to the G component and the B component.

  The frequency-converted video signal is subjected to noise reduction processing and enhancement processing in the noise reduction processing unit 107, and is inversely converted in the frequency inverse conversion unit 111 to be restored to a spatial domain signal.

  The restored video signal is separated into an R component, a G component, and a B component by the demosaicing processing unit 103, and converted into a luminance signal and a color difference signal by the luminance signal generation unit 114 and the color difference signal generation unit 115. Thereafter, compression encoding and multiplexing are performed by the encoding unit 116 and stored in the recording medium 117 as data.

  By taking the form of the present invention, it is possible to reduce the processing amount of frequency conversion, noise removal, and inverse frequency conversion while realizing noise removal and enhancement similar to those of the first embodiment.

  Another embodiment of the present embodiment is shown in FIG. FIG. 4 shows an image pickup apparatus according to the present invention realized in the form of a video camera. The video camera of this embodiment includes an optical system 101, an image sensor 102, a demosaicing processing unit 103, frequency conversion units 104 and 105, noise reduction processing units 107 and 108, a frame memory 110, frequency inverse conversion units 111 and 112, luminance A signal generation unit 114, a color difference signal generation unit 115, an encoding unit 116, and a recording medium 117 are provided. Hereinafter, the operation of the video camera of this embodiment will be described in order.

  Again, since there are many common points with the first embodiment, only a brief description will be given, and the detailed operation of each part is the same as in the first embodiment, and is omitted. First, a video signal captured by the optical system 101 and the image sensor 102 is separated into an R component, a G component, and a B component by the demosaicing processing unit 103, and the luminance signal is generated by the luminance signal generation unit 114 and the color difference signal generation unit 115. And converted into a color difference signal. The generated luminance signal and color difference signal are input to the frequency conversion units 104 and 105 and subjected to frequency conversion. Thereafter, noise reduction processing and enhancement processing are performed in the noise reduction processing units 107 and 108, and restored in the frequency inverse transformation units 111 and 112. The luminance signal and color difference signal from which noise has been removed are subjected to compression encoding and multiplexing by the encoding unit 116 and stored as data in the recording medium 117.

  By taking the form of the present invention, it is possible to suppress the generation of false colors and color noise that may occur by processing individually for RGB while realizing noise removal and enhancement similar to those of the first embodiment. Is possible.

  Further, as a development of the above-described embodiment, it is possible to adopt a form in which the noise signal is removed in the frequency domain for the luminance signal and the frame cyclic noise removal that has been used conventionally is performed for the color difference signal. This embodiment is shown in FIG. By adopting this embodiment, it is possible to reduce the frequency conversion processing unit and inverse frequency conversion processing on the color difference signal side, which leads to a reduction in processing amount, circuit scale, and power consumption.

  Although the embodiments of the present invention have been described above, the configurations shown in the first and second inventions are not limited to the first and second embodiments, and can be freely changed within the scope of the invention. Is also possible.

  In the imaging apparatus described above, the conversion unit that performs orthogonal transformation on the input video signal to the imaging system to generate a frequency conversion signal, and the delay unit that delays the frequency conversion signal and outputs a frame or field delay signal Correlation generation means for obtaining the correlation between the frequency conversion signal and the frame or field delay signal, filtering means for performing a filtering process on the frequency conversion signal, and using the output of the correlation generation means Control means for controlling the processing content of the filter processing and inverse conversion means for performing inverse conversion on the output of the filtering means are provided. Then, the frequency conversion signal is subjected to a filter process while adaptively changing parameters according to the correlation with the frame or field delay signal, and the process result is inversely converted and output. As a result, it is possible to discriminate between a moving part and a stationary part with high accuracy, and it is possible to remove random noise using a circulation coefficient according to image characteristics.

  Further, by making the signal processing means a color signal generating means for generating three color signals of red, green, and blue, it becomes possible to remove random noise without generating false colors or color noise. .

  In addition, the signal processing means is a means for generating a luminance signal and a color difference signal, or a means for generating only a luminance signal, thereby reducing the processing load and the circuit scale while discriminating a place with movement with high accuracy. The objective is achieved.

  In addition, a first parameter generation unit that generates a noise removal strength parameter using the correlation as an input, a second parameter generation unit that generates an enhancement processing strength parameter using the correlation as an input, and the noise removal A noise reduction process according to a spatial frequency in the non-linear processing unit is provided by providing a non-linear processing unit that receives the intensity parameter, the intensity parameter of the enhancement process, and the output of the filtering unit as input, and applies a non-linear process to the output of the filtering unit. The enhancement processing can be performed, and the object of performing noise removal and edge enhancement processing without significantly increasing the processing load and circuit scale is achieved.

  It is possible to integrate high-frequency component enhancement processing and random noise processing by frequency conversion, which has the effect of reducing processing load, circuit scale, and power.

It is a figure showing the 1st Embodiment of this invention. It is a figure showing the internal structure of the noise reduction process part. It is a figure showing the 2nd Embodiment of this invention. It is a figure showing the 2nd another embodiment of this invention. It is a figure showing the 2nd another embodiment of this invention. It is a figure showing the input-output characteristic in a coring process part. It is a figure showing the input-output characteristic in an enhancement process part. It is a figure showing the example of the input-output characteristic which has the characteristic of both a coring process and an enhancement process.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Optical system 102 Image pick-up element 103 Demosaicing process part 104-106 Frequency conversion part 107-109 Noise reduction process part 110 Frame memory 110-113 Frequency reverse conversion part 114 Luminance signal generation part 115 Color difference signal generation part 116 Encoding part 117 Recording Medium

Claims (7)

Conversion means for performing orthogonal transformation on the input video signal to generate a frequency conversion signal;
Delay means for delaying the frequency conversion signal and outputting a frame or field delay signal;
Correlation degree generating means for obtaining a correlation between the frequency converted signal and the frame or field delay signal;
Filtering means for filtering the frequency converted signal;
Control means for controlling the processing content of the filter processing using the output of the correlation degree generation means;
Inverse transformation means for performing inverse transformation on the output of the filtering means;
Have
An imaging apparatus, wherein the frequency conversion signal is subjected to filter processing according to the correlation with the frame delay signal, and the processing result is inversely converted and output. An image sensor that converts light into an electrical signal;
An optical system that focuses and images the image sensor;
An electric signal output from the image sensor, and a signal processing means for performing signal processing and outputting one or a plurality of processing results;
With
The signal processing means performs orthogonal transformation on the output of the signal processing means to generate a frequency converted signal, delay means for delaying the frequency converted signal and outputting a frame or field delay signal, Correlation degree generating means for obtaining the correlation between the frequency converted signal and the frame or field delay signal, filtering means for performing filter processing on the frequency converted signal, and filtering using the output of the correlation degree generating means Control means for controlling the processing content of, and inverse conversion means for performing an inverse transformation on the output of the filtering means,
An imaging apparatus, wherein the frequency conversion signal is subjected to filter processing according to the correlation with the frame delay signal, and the processing result is inversely converted and output. The imaging device according to claim 2,
The image pickup apparatus, wherein the signal processing means is a color signal generation means for generating three color signals of red, green, and blue. The imaging device according to claim 2,
The image pickup apparatus, wherein the signal processing means is means for generating a luminance signal and a color difference signal. The imaging device according to claim 2,
The image processing apparatus according to claim 1, wherein the signal processing means is means for generating a signal in which any one of three color signals of red, green, and blue is alternately arranged. An image sensor that converts light into an electrical signal;
An optical system that focuses and images the image sensor;
Signal processing means for receiving an electrical signal output from the image sensor and generating a luminance signal and a color difference signal;
Conversion means for performing orthogonal transform on the luminance signal to generate a frequency converted signal;
Delay means for delaying the frequency conversion signal and outputting a frame or field delay signal;
Correlation degree generating means for obtaining a correlation between the frequency converted signal and the frame or field delay signal;
Filtering means for filtering the frequency converted signal;
Control means for controlling the processing content of the filter processing using the output of the correlation degree generation means;
Inverse transformation means for performing inverse transformation on the output of the filtering means;
Second delay means for delaying the color difference signal and outputting a frame or field delay signal;
Second filtering means for performing a second filtering process on the output of the second delay means;
Second control means for controlling the processing content of the second filter processing using the output of the correlation degree generation means;
Have
An image pickup apparatus, wherein the frequency conversion signal and the color difference signal are subjected to filter processing in accordance with the correlation with the frame delay signal, and the result of processing of the frequency conversion signal is inversely converted and output. . The imaging device according to any one of claims 1 to 6,
First parameter generating means for generating a noise removal strength parameter using the correlation as an input;
Second parameter generation means for generating an intensity parameter for enhancement processing using the correlation as an input;
A non-linear processing means for receiving the noise removal strength parameter, the enhancement processing strength parameter and the output of the filtering means, and applying a non-linear processing to the output of the filtering means;
And a non-linear processing means for performing noise reduction processing and enhancement processing according to a spatial frequency.

Images