KR102207940B1 - Method and system for reducing noise - Google Patents

Abstract

According to an embodiment of the present invention, a channel separation unit for generating image data for each color channel by separating a current frame image for each color channel; A luminance acquisition unit that calculates luminance data from image data for each color channel of a previous frame image; A first filtering unit generating a first filtering value by applying a filter to the luminance data of a previous frame image; A multiplier for calculating a multiplication value for each color channel by multiplying the first filtering value by the image data for each color channel; A second filtering unit generating a second filtering value for each color channel by applying a filter to the multiplication value for each color channel; A gain calculating unit that performs a gain operation on the second filtering value for each color channel; Disclosed is a noise reduction system comprising a.

Description

Image noise removal method and system {METHOD AND SYSTEM FOR REDUCING NOISE}

The present invention relates to a method and system for removing image noise, and more particularly, to a method and system for removing noise by analyzing temporal and spatial characteristics of an image.

The video surveillance system is a system that enables real-time monitoring and post-retrieval by installing cameras in various places where monitoring is required and transmitting images acquired from the cameras to a monitor or storing them in a storage device. Surveillance cameras, which are mainly used in video surveillance systems, need to be operated for 24 hours, so they must generate images of a certain quality or higher even in low-light environments such as dark places or at night. However, there is a high possibility that noise may occur in an image in a low-light environment.

Therefore, techniques for obtaining a surveillance image with reduced noise in a low-light environment are being developed. Representative examples include a 3D filtering technique using an image structure tensor and a non-local mean-based method.

3D filtering using a structure tensor calculates a gradient between adjacent pixels for each pixel of an image, and generates a structure tensor from the generated structure tensor, eigenvector and eigenvalue. ) Is calculated. From this, the covariance matrix, scaling matrix, rotation matrix, etc. of the 3D Gaussian distribution are calculated to finally generate a Gaussian 3D filtering kernel optimized for the corresponding pixel. Such a method has a problem in that the amount of computation required when generating and applying a kernel is enormous, and noise is not smoothly removed, making it difficult to commercialize.

In addition, the non-regional average-based method uses the property that the characteristics of the local region are found in other parts of the image. In the non-regional average-based method, first, for a block of a specific size, a block similar to the block is searched in an image, and the found blocks are gathered together to perform noise removal filtering suitable for characteristics. After that, the filtered blocks are placed back to their original positions. Although the non-regional average-based method has good image noise removal performance, there is a problem in that similar blocks for a specific block must be searched in the entire image, and there is a problem that block artifacts occur.

An object of the present invention is to provide a method for removing noise with a small amount of computation by analyzing temporal and spatial characteristics of a noise signal.

In addition, it is another object to provide a noise removal method that is possible even in a low-light environment.

According to an embodiment of the present invention, a channel separation unit for generating image data for each color channel by separating a current frame image for each color channel; A luminance acquisition unit that calculates luminance data from image data for each color channel of a previous frame image; A first filtering unit generating a first filtering value by applying a filter to the luminance data of a previous frame image; A multiplier for calculating a multiplication value for each color channel by multiplying the first filtering value by the image data for each color channel; A second filtering unit generating a second filtering value for each color channel by applying a filter to the multiplication value for each color channel; A gain calculating unit that performs a gain operation on the second filtering value for each color channel; Disclosed is a noise reduction system comprising a.

In the present invention, the luminance obtaining unit may calculate luminance data and temporarily store it in a memory, and the multiplication unit may obtain the first filtering value from the memory.

In the present invention, the current frame image and the previous frame image may be images captured in a low-light environment.

In the present invention, the image data for each color channel may form red channel image data, green channel image data, and blue channel image data.

In the present invention, the first filtering unit or the second filtering unit may apply a low pass filter.

In the present invention, the gain operation may be a combination of an increase/decrease multiplication operation.

In the present invention, the gain calculator may generate an output image from which noise is removed by combining the image for each color channel generated by performing a gain operation on the second filtering value for each color channel.

Another embodiment of the present invention is a channel separation step of separating a current frame image for each color channel to generate image data for each color channel; A multiplication step of calculating a multiplication value for each color channel by multiplying the image data for each color channel by a first filtering value obtained by applying a filter to the luminance data of a previous frame image; A second filtering step of generating a second filtering value for each color channel by applying a filter to the multiplication value for each color channel; A gain calculating step of performing a gain operation on the second filtering value for each color channel; Disclosed is a noise removal method comprising a.

In the present invention, a luminance obtaining step of calculating luminance data from image data for each color channel of a previous frame image; A first filtering step of generating a first filtering value by applying a filter to the luminance data of a previous frame image; It may additionally include.

In the present invention, in the luminance obtaining step, luminance data may be calculated and temporarily stored in a memory, and the multiplying step may obtain the first filtering value from a memory.

In the present invention, the current frame image and the previous frame image may be images captured in a low-light environment.

In the present invention, the image data for each color channel may form red channel image data, green channel image data, and blue channel image data.

In the present invention, a low pass filter may be applied in the first filtering step or the second filtering step.

In the present invention, the gain operation may be a combination of an increase/decrease multiplication operation.

In the present invention, in the gain calculation step, an output image from which noise is removed may be generated by combining images for each color channel generated by performing a gain operation on the second filtering value for each color channel.

According to the present invention, a noise removal method capable of enhancing luminance data of a low-illuminance image can be disclosed while removing noise with a small computational amount by analyzing temporal and spatial characteristics of a noise signal.

According to the present invention, a noise removal method in which an absolute computation amount is reduced, a non-flattening trace by a filtering kernel does not appear, and a block artifact by block processing is reduced can be disclosed.

1 is a diagram showing the configuration of a noise removal system according to an embodiment of the present invention.
2 is a diagram showing an internal configuration of a noise removal server according to an embodiment of the present invention.
3 is a diagram illustrating that a noise removal server receives an input image and generates an output image from which noise is removed according to an embodiment of the present invention.
4 is a block diagram illustrating an operation of a noise removal server according to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION The detailed description of the present invention described below refers to the accompanying drawings, which illustrate specific embodiments in which the present invention may be practiced. These embodiments are described in detail sufficient to enable those skilled in the art to practice the present invention. It is to be understood that the various embodiments of the present invention are different from each other but need not be mutually exclusive. For example, specific shapes, structures, and characteristics described herein may be changed from one embodiment to another and implemented without departing from the spirit and scope of the present invention. In addition, it should be understood that the positions or arrangements of individual elements in each embodiment may be changed without departing from the spirit and scope of the present invention. Therefore, the detailed description to be described below is not made in a limiting sense, and the scope of the present invention should be taken as encompassing the scope claimed by the claims of the claims and all scopes equivalent thereto. Like reference numerals in the drawings indicate the same or similar elements over several aspects.

Hereinafter, various embodiments of the present invention will be described in detail with reference to the accompanying drawings in order to enable those of ordinary skill in the art to easily implement the present invention.

1 is a diagram showing the configuration of a noise removal system according to an embodiment of the present invention.

Referring to FIG. 1, the noise removal server 100 of the present invention includes a noise removal server 100, a plurality of network cameras 200, and a network 300.

Referring to FIG. 1, a plurality of cameras 200 exchange data with a noise removal server 100 through a network 300. Specifically, the plurality of cameras 200 communicate with the noise removal server 100 through a communication channel (DCOM), and transmit live view video data to the noise removal server 100 through an image data channel (DIMA). . Of course, only one surveillance camera rather than a plurality of surveillance cameras may communicate with the noise removal server 100, and one surveillance camera or a plurality of surveillance cameras may communicate with a plurality of noise removal servers 100. Of course, it can be transformed.

At this time, the plurality of surveillance cameras 200 are surveillance cameras and may transmit image data to the noise removal server 100 even in a low-light environment. In the case of a surveillance camera, since a 24-hour surveillance video is required, a surveillance camera may be installed in a low-light environment, such as at night or in a dark place, depending on time. When a plurality of surveillance cameras 200 are installed, it is necessary to remove noise by transmitting a low-illuminance image to the noise removal server 100.

Here, the network 300 forming the communication channel DCOM and the image data channel DIMA may be any means capable of transmitting and receiving data or commands by wire or wirelessly. For example, the network 300 may connect the plurality of cameras 200 and the noise removal server 100 by wire through a cable, and the plurality of cameras 200 and noise may be wirelessly connected by using a wireless LAN or the like. It is also possible to connect the removal server 100.

The noise removal server 100 may be used without limitation, such as a general computing device and a server device. For example, a control system for a surveillance camera may be used as the noise removal server 100. The noise removal server 100 may store real-time video from a plurality of cameras 200 as needed.

The noise removal server 100 serves to remove noise from images acquired from the plurality of surveillance cameras 200. In particular, when the image acquired from the plurality of surveillance cameras 200 is a low-illuminance image, the noise removal server 100 may remove noise with higher quality.

In general, a low-illuminance image has more severe noise due to a problem of the camera itself than an image acquired in a general environment. In addition, since the luminance is low in a low-illumination image, it is difficult to distinguish a subject, so the user tends to feel the noise more severely. When the original image with noise of t frames acquired from the plurality of surveillance cameras 200 is I(t), the noise is N(t), and the noise-removed image is dnI(t), the original image I (t) can be expressed as Equation 1 below.

[Equation 1]

I(t) = dnI(t) + N(t)

According to an embodiment of the present invention, a noise removal system obtains a noise image I(t) and performs temporal and spatial filtering to obtain a dnI(t) image from which noise is removed. In more detail, a low-pass filter is used to perform spatial filtering, and luminance data of a previous frame image is referred to to perform temporal filtering.

Noise included in an image can be interpreted as having statistical characteristics of a Gaussian distribution of a specific mean and standard deviation within one frame. That is, since noise has a Gaussian distribution on a spatial axis, a low pass filter that makes good use of this characteristic can remove noise existing in an image to some extent. Since the noise of the next frame image can be said to have a statistical distribution similar to that of the previous frame image, noise existing in the image can also be removed by using a low-pass filter.

As described above, the noise signal included in each frame image can be statistically modeled to some extent. However, when looking at a specific pixel in terms of a temporal axis, there is a greater probability that noise existing in the previous frame does not exist at the same position in the next frame. In other words, noise can be interpreted as moving position when viewed from the perspective of the time axis. Therefore, when comparing the noise between frames, it is unlikely that noise continues to exist at the same pixel position in the previous frame and the next frame. That is, it is very likely that the noise that appears in the current frame for a specific pixel does not exist in the previous frame or its intensity has changed considerably.

Therefore, the noise removal system according to an embodiment of the present invention removes noise by using low-pass filtering spatially, but by referring to the luminance data of the previous frame image temporally, the noise existing in the image in the low-light environment is effectively removed. Can be removed.

2 is a diagram showing the internal configuration of the noise removal server 100 according to an embodiment of the present invention.

2, the noise removal server 100 includes a channel separation unit 111, a luminance acquisition unit 112, a multiplication unit 113, a first filtering unit 114, a second filtering unit 115, and gain It includes an operation unit 116, a memory 120 and a control unit 130.

The control unit 130 of the noise removal server 100 includes a channel separation unit 111, a luminance acquisition unit 112, a multiplication unit 113, a first filtering unit 114, a second filtering unit 115, and a gain calculating unit. The function of controlling data flow between 116 and the memory 120 may be performed. That is, the control unit 130 according to the present invention controls the data flow from/to the outside of the noise removal server 100 or the data flow between each component of the noise removal server 100, so that the channel separation unit 111, The luminance acquisition unit 112, the multiplication unit 113, the first filtering unit 114, the second filtering unit 115, the gain calculating unit 116, and the memory 120 can be controlled to perform their own functions, respectively. have.

3 is a diagram illustrating that the noise removal server 100 receives an input image and generates an output image from which noise is removed according to an embodiment of the present invention.

4 is a block diagram showing the operation of the noise removal server 100 according to an embodiment of the present invention.

Hereinafter, a process of removing noise from an input image by the noise removal server 100 will be described with reference to FIGS. 2 to 4.

First, referring to FIGS. 3 and 4, the noise removal server receives I(t) as an input image (INPUT).

Next, the channel separation unit 111 of FIG. 2 serves to separate the image into three channels, that is, RGB channels, red, green, and blue. The channel separation unit 111 performs RGB conversion on all sequentially incoming images.

RGB is a method of expressing colors using the three primary colors of light. Colors are mixed using three types of light sources (RED, GREEN, and BLUE), and the brighter the colors are mixed. Red, green, and blue channels may all be referred to as color channels. The channel separating unit 111 may analyze the image to generate a red channel, a green channel, and a blue channel, and store them in the memory 120.

3 and 4, image data of red, green, and blue channels in a t frame generated by the channel separation unit 111 from an input image I(t) are R(t), G(t), respectively. It can be expressed as B(t).

When an image captured in a low-light environment is analyzed, there is a case where the ratio of the signal intensity of the red, green, and blue channels is significantly different. Therefore, when comparing the current red, green, and blue channels with the red, green, and blue channels of the previous frame to remove noise, the channel characteristics for each color are reflected as noise characteristics, so noise cannot be properly removed. In an embodiment of the present invention, in order to compensate for these disadvantages, a luminance data value is calculated from image data for each red, green, and blue channel of a previous frame, and this luminance data is used as an image of the red, green, and blue channels of the current frame. It is used as a standard for comparison with data.

To this end, the luminance acquisition unit 112 of FIG. 2 serves to calculate luminance data of an image. The luminance acquisition unit 112 may temporarily store luminance data of the calculated image in the memory 120. In the noise removal method according to an embodiment of the present invention, the luminance acquisition unit 112 calculates the luminance of the previous frame image and uses it in the noise removal method of the current image. If the luminance data value of the previous frame is used to determine the current value of a specific pixel, the difference in noise removal performance caused by the difference in brightness of each channel can be overcome, and noise removal reflecting the noise characteristics between frames is possible. It is possible.

The luminance acquisition unit 112 may generate luminance data based on image data of red, green, and blue channels generated by the channel separation unit 111. When the luminance data in the t frame generated by the luminance acquisition unit 112 is Y(t), Y(t) may be calculated as in [Equation 1] below.

[Equation 1]

Y(t) = 0.299*R(t) + 0.587*G(t) + 0.114*B(t)

3 and 4, it can be seen that the luminance acquisition unit 122 generates luminance data of the input image I(t) as Y(t).

Meanwhile, assuming that the frame t is the current frame, the luminance data of the frame t-1, which is the previous frame, can be obtained as Y(t-1) through a calculation method such as [Equation 1] above.

Next, the first filtering unit 114 of FIG. 2 performs low-pass filtering on the luminance data, and the multiplier 113 serves to multiply the filtered luminance data and image data for each color channel.

In more detail, the first filtering unit 114 may apply a low pass filter to the luminance data of the previous frame. The first filtering unit 114 may obtain the luminance data of the previous frame by reading from the memory. As described above, the luminance data of the previous frame may be expressed as Y(t-1), and the first filtering value obtained by applying the low-pass filter to the luminance data of the previous frame may be expressed as FY(t-1). F() means a low-pass filtering operation.

3 and 4, the luminance data Y(t) of the current frame is passed through the low-pass filter LPF by the first filtering unit 114 to generate FY(t), and the generated FY(t ) Is stored in the memory 120. The first filtering value FY(t) stored in the memory is used to remove noise of the next frame image. 3 and 4, the first filtering value FY(t-1) of the previous frame image stored in the memory is used to remove noise from the current frame image.

Next, the multiplication unit 113 of FIG. 2 includes FY(t-1), which is the first filtering value of the previous frame generated by the first filtering unit 114, and R(t) and G, which are image data for each color channel. It multiplies (t) and B(t) respectively. The value calculated by the multiplier 113 is referred to as a multiplication value for each color channel, and multiplication values for each color channel for each of red, green, and blue are generated. In more detail, referring to FIGS. 3 and 4, the red channel image data and the first filtering value are multiplied to generate the red channel multiplication value R(t)*FY(t-1), and the green channel image data and the first filtering value The green channel multiplication value G(t)*FY(t-1) is generated by multiplying the value, and the blue channel multiplication value B(t)*FY(t-1) is multiplied by the blue channel image data and the first filtering value. Is created.

Next, the second filtering unit 115 of FIG. 2 serves to calculate a second filtering value for each color channel by performing low-pass filtering on the multiplication value for each color channel. In more detail, referring to FIGS. 3 and 4, a low-pass filter (LPF) is applied to the red channel multiplication value, R(t)*FY(t-1), and the red channel second filtering value, F(R(t)). *FY(t-1)) is generated. In addition, a low-pass filter is applied to the green channel multiplication value G(t)*FY(t-1) to generate the green channel second filtering value F(G(t)*FY(t-1)). In addition, a low-pass filter is applied to the blue channel multiplication value B(t)*FY(t-1) to generate the blue channel second filtering value F(B(t)*FY(t-1)).

The gain calculator 116 of FIG. 2 obtains image data for each color channel from which noise is removed by performing a gain operation on the second filtering value for each color channel. In the gain operation, multiplication, division, addition, and subtraction operations of appropriate constants or variables are performed. When the gain operation is G(), when the gain operation is performed on the second filtering value for each color channel as in the example of FIGS. 3 and 4, G(F(R(t)*FY(t-1))), G(F(G(t)*FY(t-1))) and G(F(B(t)*FY(t-1))) are calculated.

As described above, when the noise removal function of the noise image is expressed as dn(), the noise removal images for each color channel are dnR(t), dnG(t), and dnB(t), respectively. The noise image for each color channel is image data for each color channel from which the gaze calculated by the gain calculating unit 116 is removed, and may be expressed as [Equation 2] below.

[Equation 2]

dnR(t) = G(F(R(t)*FY(t-1)))

dnG(t) = G(F(G(t)*FY(t-1)))

dnB(t) = G(F(B(t)*FY(t-1)))

Accordingly, according to an embodiment of the present invention, dnI(t), which is an output image from which noise is finally removed, may be obtained by combining dnR(t), dnG(t), and dnB(t).

For reference, in the present specification, the gain association is expressed as G(), and gain calculations of various known methods can be used in the present invention. For example, a gain calculation method based on constant multiplication and constant sum, an amplification method based on tone mapping, a gain calculation method based on gamma correction, and the like may be used without a proposal.

5 is a flow chart showing a noise removal method according to an embodiment of the present invention.

Referring to FIG. 5, first, the noise removal server acquires an original noise image of a current frame from a camera as an input image I(t) (S1).

Next, the current frame image is separated for each color channel to generate image data R(t), G(t), and B(t) for each color channel (S2).

Next, the luminance data Y(t-1) calculated from the image data for each color channel of the previous frame is obtained from the memory (S3).

Next, a first filtering value FY(t-1) is generated by applying a low pass filter (LPF) to the luminance data of the previous frame image (S4).

Next, the first filtering value is multiplied by the image data for each color channel, and the multiplication values for each color channel R(t)*FT(t-1), G(t)*FT(t-1), B(t)*FT (t-1) is generated (S5).

Next, the second filtering value F(R(t)*FT(t-1)), F(G(t)*FT(t) for each color channel by applying a low pass filter (LPF) to the multiplication value for each color channel -1)), F(B(t)*FT(t-1)) is generated (S6).

Next, a gain operation is performed on the second filtering value for each color channel to obtain dnR(t), dnG(t), and dnB(t) (S7).

Next, the image dnI(t) from which noise has been removed from dnR(t), dnG(t), and dnB(t) is output (S8).

In the above-described embodiment of the noise removal method according to the present invention, the level of noise in the frame may be reduced by performing spatially low-pass filtering within one frame. In addition, for temporal frames, a luminance signal was used to minimize the difference in intensity and noise between the R, G, and B channels. By multiplying the R, G, B channels of the current frame and the luminance data of the previous frame, The noise level of is reduced and the luminance component is enhanced. According to such an embodiment of the present invention, compared to conventional methods, the amount of computation for noise removal can be drastically reduced.

The embodiments according to the present invention described above may be implemented in the form of program instructions that can be executed through various computer components and recorded in a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded in the computer-readable recording medium may be specially designed and constructed for the present invention or may be known and usable to those skilled in the computer software field. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical recording media such as CD-ROMs and DVDs, and magnetic-optical media such as floptical disks. medium), and a hardware device specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of the program instructions include not only machine language codes such as those produced by a compiler but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware device can be changed to one or more software modules to perform the processing according to the present invention, and vice versa.

The specific implementations described in the present invention are examples and do not limit the scope of the present invention in any way. For brevity of the specification, descriptions of conventional electronic configurations, control systems, software, and other functional aspects of the systems may be omitted. In addition, the connection or connection members of the lines between the components shown in the drawings exemplarily represent functional connections and/or physical or circuit connections, and in an actual device, various functional connections that can be replaced or added Connections, or circuit connections. In addition, if there is no specific mention such as “essential” or “importantly”, it may not be an essential component for the application of the present invention.

In the specification of the present invention (especially in the claims), the use of the term “above” and a similar reference term may correspond to both the singular and the plural. In addition, when a range is described in the present invention, the invention to which an individual value falling within the range is applied (unless otherwise stated), and each individual value constituting the range is described in the detailed description of the invention. Same as Finally, unless explicitly stated or contradictory to the steps constituting the method according to the present invention, the steps may be performed in an appropriate order. The present invention is not necessarily limited according to the order of description of the steps. The use of all examples or illustrative terms (for example, etc.) in the present invention is merely for describing the present invention in detail, and the scope of the present invention is limited by the above examples or illustrative terms unless limited by the claims. It does not become. In addition, those skilled in the art can recognize that various modifications, combinations, and changes may be configured according to design conditions and factors within the scope of the appended claims or their equivalents.

100: noise removal server 111: channel separation unit
112: luminance acquisition unit 113: multiplication unit
114: first filtering unit 115: second filtering unit
116: gain operation unit 120: memory
130: control unit 200: camera
300: network

Claims (15)

A channel separating unit for separating the current frame image for each red, green, and blue channel to generate red channel image data, green channel image data, and blue channel image data;
A luminance acquisition unit that calculates luminance data from image data for each red, green, and blue channel of a previous frame image;
A first filtering unit generating a first filtering value by applying a filter to the luminance data of a previous frame image;
A multiplier for multiplying the image data for each of the red, green, and blue channels by the first filtering value to separate and calculate a red channel multiplication value, a green channel multiplication value, and a blue channel multiplication value;
A second filtering unit for generating a red channel second filtering value, a green channel second filtering value, and a blue channel second filtering value by applying a filter to each of the red channel multiplication value, green channel multiplication value, and blue channel multiplication value. ;
A gain calculating unit that performs a gain operation on the red channel second filtering value, the green channel second filtering value, and the blue channel second filtering value;
Noise reduction system comprising a. The method of claim 1,
The luminance obtaining unit calculates luminance data and temporarily stores it in a memory, and the multiplication unit obtains the first filtering value from a memory. The method of claim 1,
The current frame image and the previous frame image are images captured in a low-light environment. delete The method of claim 1,
A noise removal system for applying a low pass filter to the first filtering unit or the second filtering unit. The method of claim 1,
The noise removal system in which the gain calculation is a combination of an increase/decrease multiplication operation. The method of claim 1,
The gain calculator combines the image for each color channel generated by performing a gain operation on the red channel second filtering value, the green channel second filtering value, and the blue channel second filtering value to generate an output image from which noise is removed. Removal system. A channel separation step of separating the current frame image for each red, green, and blue channel to generate red channel image data, green channel image data, and blue channel image data;
A multiplication step of dividing and calculating a red channel multiplication value, a green channel multiplication value, and a blue channel multiplication value by multiplying each of the red, green, and blue channel image data by a filter applied to the luminance data of the previous frame image ;
A second filtering step of generating a red channel second filtering value, a green channel second filtering value, and a blue channel second filtering value by applying a filter to each of the red channel multiplication value, green channel multiplication value, and blue channel multiplication value. ;
A gain calculating step of performing a gain operation on the red channel second filtering value, the green channel second filtering value, and the blue channel second filtering value;
Noise removal method comprising a. The method of claim 8,
A luminance obtaining step of calculating luminance data from image data for each color channel of a previous frame image;
A first filtering step of generating a first filtering value by applying a filter to the luminance data of a previous frame image;
Noise removal method further comprising. The method of claim 9,
In the luminance obtaining step, luminance data is calculated and temporarily stored in a memory, and in the multiplying step, the first filtering value is obtained from a memory. The method of claim 8,
The current frame image and the previous frame image are images captured in a low-light environment. delete 10. The method of claim 9, wherein the first filtering step or the second filtering step applies a low pass filter. The method of claim 8, wherein the gain operation is a combination of an increase/decrease multiplication operation. The method of claim 8,
In the gain calculation step, a noise-removed output image is generated by combining the image for each color channel generated by performing a gain operation on the red channel second filtering value, the green channel second filtering value, and the blue channel second filtering value. How to remove noise.

Images