JP2007235319A - Noise reduction processing apparatus, method, program, and camera apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain a high noise reduction effect, while leaving a detailed part of an image by individually applying an appropriate noise reduction processing to a moving picture part and a still picture part. <P>SOLUTION: The noise reduction processing apparatus includes an image input section 11 for receiving information; a region partitioning section 50 for partitioning inside of a single frame of a video signal received from the image input section 11 into a moving picture region and a still picture region; and a noise reduction processing section 13 that applies noise reduction processing on the moving picture region and the still picture region by different methods, respectively. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a noise reduction device, a method, a program, and a camera device that reduce noise in an image taken by, for example, a digital camera.

When an analog input image is digitally displayed, it is known that various noises (block noise, mosquito noise, random noise, etc.) of the image accompanying conversion processing occur. As a technique for reducing these noises, conventionally, in-plane noise reduction processing (hereinafter referred to as “in-plane NR processing”) in which noise reduction (NR) processing is performed within one frame, or noise by adding a plurality of frames. Inter-surface noise reduction processing (hereinafter referred to as “inter-surface NR processing”) that performs reduction processing is known.
However, in the in-plane NR process, if the NR process is applied too much in order to reduce the noise of the entire frame as much as possible, there is a problem that the fine part of the image is crushed and the image is deteriorated. On the other hand, if the NR process is weakened so that the fine portions are not deteriorated, the noise reduction effect of the entire image is lowered. Further, the inter-surface NR process has a problem in that when the subject is moving, the moving part is frame-added, resulting in blurring (multiple images) in the moving part.

In recent years, various techniques for solving the above-described problems have been proposed.
For example, Patent Literature 1 discloses a technique for determining a feature of an input image (moving image / still image region in a frame, etc.) and performing noise reduction processing by changing the strength of the NR effect with respect to this feature. Yes.
Further, for example, Patent Document 2 discloses a technique for detecting a motion of a subject in a certain frame image and selecting either in-plane NR processing or inter-plane NR processing to reduce noise.

JP-A-5-091532 JP 2002-010106 A

  However, in the inventions disclosed in Patent Document 1 and Patent Document 2 described above, since NR processing suitable for either a still image region or a moving image region in a frame is performed, noise remains partially. There is a problem that the image becomes blurred due to excessive smoothing.

  The present invention has been made in view of such circumstances, and by performing an appropriate noise reduction process on the moving image portion and the still image portion individually, it remains high while leaving a fine portion of the image. It is an object of the present invention to provide a noise reduction processing device, method, program, and camera device that can obtain a noise reduction effect.

In order to solve the above problems, the present invention employs the following means.
The present invention provides an input means for inputting information, an area classification means for dividing one frame of a video signal input from the input means into a moving image area and a still image area, the moving image area and the still image area, There is provided a noise reduction processing device comprising noise reduction processing means for executing noise reduction processing of different methods.

  According to such a configuration, one frame of the video signal is divided into a moving image region and a still image region, and noise reduction processing of different methods is performed on the moving image region and the still image region. Become. This makes it possible to perform noise processing suitable for moving images in the moving image region and noise processing suitable for still images individually in the still image region, so that high noise is suppressed while suppressing deterioration in details. A reduction effect can be obtained. As a result, the image quality can be improved.

  In the above-described noise reduction processing apparatus, the noise reduction processing unit may execute inter-plane processing for the still image region and in-plane processing for the moving image region.

  According to such a configuration, the inter-plane processing is performed on the still image area and the in-plane processing is performed on the moving image area, thereby preventing blurring (multiple images) and detail deterioration due to frame addition. However, a high noise reduction effect can be obtained.

  In the above-described noise reduction processing apparatus, the noise reduction processing means may perform inter-plane processing and in-plane processing on the still image area.

  According to such a configuration, since the inter-plane processing and the in-plane processing are executed for the still image area, a higher noise reduction effect can be obtained.

  In the above-described noise reduction processing apparatus, the noise reduction processing means may vary the noise reduction level for the moving image area according to the magnitude of a motion vector.

  According to such a configuration, since the noise reduction level is made different depending on the magnitude of the motion vector for the moving image area, the noise reduction level is lowered for the area where the magnitude of the motion vector is small. Thus, it is possible to avoid blurring and deterioration of details. Conversely, for areas where the motion vector is large, increasing the noise reduction level allows blurring and detail deterioration (the moving subject is inherently blurred) and has a high noise reduction effect. Can be obtained. In this way, it is possible to perform more detailed noise reduction processing according to the characteristics of each moving image area.

  In the noise reduction processing apparatus, the region classification unit classifies a region having motion vectors in the same direction and the same size as the pan / tilt region in the frame, and the noise reduction processing unit However, the same noise reduction processing as that in the still image area may be performed on the pan / tilt area.

  For example, when a video signal is captured, functions such as pan and tilt are used, so that a stationary subject may be photographed as if it moves between frames. Even in such a case, an area having the same direction and the same motion vector is divided as a pan / tilt area and handled in the same manner as a still image area, so that appropriate noise reduction processing can be performed. .

  The noise reduction processing device may further include boundary region processing means for detecting a boundary region between the moving image region and the still image region and reducing a difference in noise reduction level in the boundary region.

  According to such a configuration, it is possible to reduce the difference in noise reduction level that occurs in the boundary area between the moving image area and the still image area. The uncomfortable feeling resulting from being applied can be reduced.

  In the noise reduction processing apparatus, the boundary region processing unit includes a discontinuity detection unit that detects a difference in noise reduction level in the boundary region, and a high noise region that is a region having a low noise reduction level in the boundary region. On the other hand, a level difference reduction unit that executes a noise reduction process according to the difference in the noise reduction level detected by the discontinuity detection unit may be provided.

  According to such a configuration, in the boundary area, noise reduction processing is performed on an area with a low noise reduction level so that the noise reduction level matches that of the other areas. The discontinuity occurring at the boundary can be reduced.

  In the above-described noise reduction processing device, the noise reduction processing means may vary the noise reduction level according to a position in the one frame.

  According to such a configuration, since the noise reduction level is varied based on the position in one frame, a suitable noise reduction process that suppresses deterioration of details is performed on an important region that collects the line of sight in the frame. It becomes possible.

  In the above noise reduction processing apparatus, the noise reduction processing means has the lowest noise reduction level with respect to the still image region in the central portion and in the vicinity of the central portion in the one frame, and the moving image region in the central portion and in the vicinity of the central portion, The noise reduction level may be gradually increased in the order of the outer peripheral portion and the still image region near the outer peripheral portion, and the outer peripheral portion and the moving image region near the outer peripheral portion.

  According to such a configuration, the noise reduction level is set to be low for the central portion where the user's line of sight is collected most in one frame and the vicinity thereof, so that the image with reduced blur and details is reduced and noise is reduced. Can be provided.

  In the above-described noise reduction processing apparatus, the noise reduction processing means may vary the noise reduction level according to the luminance in the one frame.

  According to such a configuration, since the noise reduction level is made different based on the luminance within one frame, it is preferable that high noise reduction is applied to a region having luminance that makes it easy for humans to recognize noise within the frame. Noise reduction processing can be performed.

  In the above-described noise reduction processing apparatus, the noise reduction processing means may execute a noise reduction process having a higher noise reduction level than other areas on the intermediate luminance area where the luminance is at an intermediate level.

  According to such a configuration, the noise reduction effect can be enhanced by increasing the noise reduction level in an area where the luminance is at an intermediate level in one frame and is easily visually recognized with respect to other areas.

  The noise reduction processing device includes a frequency detection unit that detects a frequency component of the moving image region information, and the noise reduction processing unit has a noise in the high frequency component region that is lower than a noise reduction level in the low frequency component region. The reduction level may be lowered.

  According to such a configuration, since the noise reduction level is lowered for the high frequency component region, it is possible to reduce the deterioration of image details, and for the low frequency component region, the noise is reduced. By increasing the reduction level, the noise reduction effect can be enhanced.

  The noise reduction processing apparatus may further include a threshold adjusting unit that adjusts a threshold used when the region dividing unit divides the frame into the moving image region and the still image region.

  According to such a configuration, it is possible to adjust a threshold value used as a reference when dividing into a still image region and a moving image region, so that an image with a desired image quality can be created.

  The noise reduction processing device may include a mode setting unit capable of setting a moving image mode that emphasizes moving images and a still image mode that emphasizes still images, and the threshold adjustment unit is configured to be the mode setting unit. The threshold value may be adjusted according to the set mode.

  According to such a configuration, it is possible to automatically set a threshold value suitable for each mode by selecting a moving image mode in which a moving image is emphasized or a still image mode in which a still image is emphasized by the user. It becomes possible.

  The present invention relates to an input process in which information is input, an area separation process in which one frame of the input video signal is divided into a moving image area and a still image area, and a method different in the moving image area and the still image area. There is provided a noise reduction processing method including a noise reduction processing process for performing noise reduction processing.

  The present invention relates to an input step in which information is input, a region separation step in which one frame of the input video signal is divided into a moving image region and a still image region, and a method different in the moving image region and the still image region. A noise reduction program for causing a computer to execute a noise reduction processing step for executing noise reduction processing is provided.

  The present invention includes an optical unit that forms an image of a subject, an imaging unit that photoelectrically converts the subject image formed by the optical unit, an input unit that receives a video signal photoelectrically converted by the imaging unit, and the input An area dividing means for dividing one frame of the video signal output from the means into a moving picture area and a still picture area, and a noise reduction processing means for executing noise reduction processing of a different method between the moving picture area and the still picture area A camera device is provided.

  According to such a configuration, the subject image formed by the optical unit is photoelectrically converted by the imaging unit and input to the input unit as a video signal of an electrical signal. This video signal is transferred from the input means to the area sorting means, where it is divided into a moving picture area and a still picture area within one frame. Further, the noise reduction processing means executes different noise reduction processing on the moving image region and the still image region, and the image data after the noise reduction processing is output. As a result, noise processing suitable for moving images can be performed on the moving image region, and noise processing suitable for still images can be performed individually on the still image region, so that the image quality can be improved. .

  The camera apparatus may include a camera shake correction unit that corrects image blurring that occurs during an exposure period of the imaging unit, and the area classification unit is configured to output the video signal corrected by the camera shake correction unit. One frame may be divided into the moving image region and the still image region.

  According to such a configuration, camera shake correction for correcting image blur at the time of shooting is performed on the video signal output from the input unit or the like, and the video signal after the camera shake correction is output to the region sorting unit. Therefore, the subsequent processing can be performed based on a clean image that is not blurred, and the image quality can be improved.

  The noise reduction processing apparatus of the present invention is particularly suitable for being mounted on products such as a stationary camera for broadcasting, an ENG camera, a consumer handy camera, and a digital camera. Further, it is also used in a video signal correction program (CG program) that handles moving images, an image editing device, and the like.

  According to the present invention, since appropriate noise reduction processing is individually performed on the moving image portion and the still image portion, it is possible to obtain a high noise reduction effect while leaving a fine portion of the image.

Hereinafter, embodiments of a noise reduction processing apparatus of the present invention will be described with reference to the drawings.
[First Embodiment]
FIG. 1 is a block diagram schematically showing a schematic configuration of a noise reduction processing apparatus according to the first embodiment of the present invention.
As shown in FIG. 1, the noise reduction processing apparatus according to the present embodiment includes an image input unit (input unit) 11 to which a video signal is input as information, and one frame of the video signal input from the image input unit 11. Reduction for executing noise reduction processing of different methods for the area dividing unit (area dividing means) 50 for dividing the image into the moving image area and the still image area, and the moving image area and the still image area divided by the area dividing unit 50, respectively. And a processing unit (noise reduction processing means) 13.

  The region segmentation unit 50 detects a motion vector in the video signal from the image input unit 11, classifies one frame into a still image region and a moving image region based on the motion vector, and uses the segmentation result as region information. The motion detection unit 14 that outputs the image signal and the region separation that separates the video signal from the image input unit 11 into a still image region and a moving image region based on the region information from the motion detection unit 14 and outputs the image signal to the noise reduction unit 13 Part 12 is provided. The motion detection unit 14, for example, blocks an image of one frame composed of the video signal from the image input unit 11 with an appropriate pixel size, and performs a region search within the image using this block as a detection unit, thereby determining the motion position. presume. Further, the motion detection unit 14 divides one frame into a moving image region and a still image region according to determination methods 1 to 4 described later.

  As shown in FIGS. 1 and 2, the noise reduction processing unit 13 includes a still image region NR processing unit 17 that performs NR processing on a still image region, and a moving image region that performs NR processing on a moving image region. The NR processing unit 18, the still image region signal after NR processing output from the still image region NR processing unit 17, and the moving image region signal after NR processing output from the moving image region NR processing unit 18 are combined into one frame. And an image synthesizing unit 19 for generating the video signal. As shown in FIG. 2, the still image region NR processing unit 17 of the noise reduction processing unit 13 performs inter-surface NR processing on the still image region signal. The moving image region NR processing unit 18 performs in-plane NR processing on the moving image region signal using the NR coefficient determined by the moving image region NR coefficient determining unit 15 described later.

  In addition, the noise reduction processing apparatus according to the present embodiment includes a moving image region NR coefficient determination unit 15 that determines an NR coefficient to be used for NR processing of a moving image region based on a motion vector from the motion detection unit 14, and a noise reduction processing unit 13. Is provided with an image output unit 16 for outputting a video signal after NR processing. Here, the NR coefficient is a parameter related to the strength of the noise reduction effect (NR effect), and the higher the NR coefficient, the higher the noise reduction level.

Next, a method for determining a still image region and a moving image region in the motion detection unit 14 described above will be described in detail with reference to the drawings.
[Judgment method 1]
In the determination method 1, as illustrated in FIG. 21, the motion detection unit 14 divides one frame into a plurality of blocks having the same size (in this embodiment, nine blocks A to I). At this time, preset values are used for the sizes and shapes of the blocks A to I. Subsequently, in each of the blocks A to I, it is determined whether each pixel is a still image pixel or a moving image pixel. If the proportion of still image pixels in each block is larger than the proportion occupied by moving image pixels, the block is stopped. If the ratio of moving image pixels is large, the block is determined as the moving image area. As described above, in the determination method 1, whether a still image region or a moving image region is determined for each block obtained by dividing one screen.

[Judgment method 2]
In the determination method 2, as shown in FIG. 22, in one frame, a frame center portion (corresponding to the position of the block E in FIG. 21) in which the line of sight is most easily collected is divided into fine blocks compared to other regions. At this time, preset values are used for the size and shape of the block. For the determination of the still image region or the moving image region, the same method as the determination method 1 described above is used. As described above, in the determination method 2, instead of equally dividing one frame, blocks are subdivided for a region that attracts attention, so that the accuracy of noise reduction processing can be improved.

[Determination method 3]
In the determination method 3, one frame is divided into a plurality of blocks based on a predetermined condition. Subsequently, in each block, a still image pixel and a moving image pixel are discriminated, and the block is further subdivided so that the proportion of the still image pixel in the block is equal to or greater than a predetermined threshold.
Hereinafter, the determination method 3 will be described with reference to FIG.

  First, in step SB1 in FIG. 23, one frame is divided into a plurality of blocks based on predetermined conditions. Here, the method of dividing into a plurality of blocks may be divided into equal-sized blocks as in the above-described separation method 1, or the size of the block may be determined depending on whether or not the region is a region of interest. It is also possible to change the shape. The size of the block to be divided here is conditional on being larger than a threshold value Th1 determined in step SB6 described later. Thus, when one frame is divided into a plurality of blocks, the following processing from step SB2 to step SB8 is executed for each block.

First, in step SB2, it is determined whether each pixel constituting the block is a still image pixel or a moving image pixel by comparing the block with the corresponding block in the next successive frame. Further, the ratio R of the number of pixels of the still image to the total number of pixels constituting the block is calculated.
Subsequently, in step SB3, if the ratio R of the still image pixels in the block is larger than the predetermined threshold Th2, the process proceeds to step SB4, where the block is recognized as a still image region and is set as the next block. The same determination process is performed for this.

On the other hand, when the ratio R occupied by the still image pixels is equal to or smaller than the predetermined threshold Th2 in step SB3, the process proceeds to step SB5, and the block is further divided (hereinafter, this divided block is referred to as “subdivision block”). "). Subsequently, in step SB6, it is determined whether or not the size of each subdivided block is larger than a preset threshold value Th1. As a result, when it is determined that the size of the subdivided block is equal to or smaller than the threshold value Th1, the process proceeds to step SB7, and the subdivided blocks subdivided in step SB5 are recombined. Subsequently, in step SB8, the recombined block is recognized as a moving image area, and the same determination process is performed for the next block.
On the other hand, when it is determined in step SB6 that the size of the subdivision block is larger than the threshold value Th1, the process returns to step SB2, and the above-described processing after step SB2 is repeatedly executed.
As described above, according to the determination method 3, it is recognized that the sub-block area is equal to or less than the threshold value Th1, and thus the moving image region is recognized, or the ratio of the still image pixels in the sub-block is the threshold value Th2. As a result, the subdivision of the blocks is repeatedly performed until it is recognized as the still image area. As a result, as shown in FIG. 24, the blocks divided by the determination method have different shapes and sizes.

[Determination method 4]
In the determination method 4, one frame is divided into a plurality of blocks based on a predetermined condition. Subsequently, in each block, a still image pixel and a moving image pixel are discriminated, and the block is subdivided into a still image region including only a still image pixel and a moving image region including only a moving image pixel. Thereby, according to the method, the blocks are subdivided in units of pixels, the still image area is composed of only still image pixels, and the moving image area is composed of only moving image pixels. The moving image area may include not only moving image pixels but also still pixels, without being limited to this example.

  As described above, according to the determination method 1 or 2, it is possible to very easily divide one frame into a moving image region and a still image region. On the other hand, according to the determination method 3 or the determination method 4, since it can be divided into a moving image region and a still image region at a very fine level, it is possible to perform NR processing with very high accuracy. Thereby, the image quality after NR processing can be improved. In particular, if a moving image pixel is included in the still image region, there is a problem that blur (double image) occurs when the inter-surface NR process is performed. Since it is composed only of still image pixels, the above-mentioned problems such as occurrence of blur can be solved.

  Note that the determination method for dividing a frame into a moving image region and a still image region is not limited to the above-described determination method, and various methods can be used. Also, in the motion vector detection method by the motion detection unit, a known technique can be appropriately selected and employed. For example, a method of estimating a motion from the level of a light / dark gradation signal when an image is binarized may be employed.

Next, the operation of the noise reduction processing apparatus according to this embodiment will be described with reference to FIGS.
First, the image input unit 11 receives a video signal from an image sensor or the like, and outputs the video signal to the region separation unit 12 and the motion detection unit 14. The motion detection unit 14 divides one frame in the input video signal into a still image region and a moving image region, and outputs the division result to the region separation unit 12 as region information. Further, the motion detection unit 14 calculates a motion vector of the subject in the moving image area and outputs the motion vector to the moving image region NR coefficient determination unit 15.
The moving image region NR coefficient determination unit 15 determines an NR coefficient based on the motion vector input from the motion detection unit 14 and outputs the determined NR coefficient to the moving image region NR processing unit 18. At this time, the moving image region NR coefficient determination unit 15 sets the NR coefficient to a larger value as the size of the motion vector is larger. Thereby, the noise reduction level is set higher as the motion vector is larger.

On the other hand, when the region information is input from the motion detection unit 14, the region separation unit 12 separates one frame image in the video signal from the image input unit 11 into a still image region and a moving image region based on the region information. The still image region signal is output to the still image region NR processing unit 17 and the moving image region signal is output to the moving image region NR processing unit 18.
The still image area NR processing unit 17 has a predetermined NR coefficient in advance, and using this NR coefficient, as shown in FIG. NR processing is executed and a still image region signal after processing is output. As shown in FIG. 2, the moving image region NR processing unit 18 performs in-plane NR processing on the moving image region signal received from the region separating unit 12 using the NR coefficient received from the moving image region NR coefficient determining unit 15. Then, the processed moving image area signal is output. The image synthesizing unit 19 synthesizes the still image region signal from the still image region NR processing unit 17 and the moving image region signal from the moving image region NR processing unit 18 to create a one-frame video signal. The image is output to the image output unit 16.

As described above, according to the noise reduction processing apparatus according to the present embodiment, inter-surface NR processing is performed on a still image region, so that there is no blur (multiple image) due to addition and deterioration of image details. Can be realized. Further, since the in-plane NR processing is performed using the NR coefficient corresponding to the size of the motion vector of the subject even in the moving image area, it is possible to reduce the deterioration of details in the image after the NR processing.
As described above, according to the present embodiment, NR processing suitable for each of the still image region and the moving image region in the frame can be performed, so that a high NR effect can be provided while leaving a fine portion of the image. be able to.

  In the above-described embodiment, the moving image region NR coefficient determining unit 15 that automatically determines the NR coefficient used in the moving image region NR processing unit 18 is provided, but instead, the user sets the NR coefficient. It is good also as providing the NR coefficient manual setting part for setting arbitrarily. Alternatively, an NR adjustment unit that allows the user to adjust the NR coefficient determined by the moving image region NR coefficient determination unit 15 may be further provided.

[Second Embodiment]
Next, a noise reduction processing apparatus according to a second embodiment of the present invention will be described with reference to the drawings. The second embodiment is basically the same as the first embodiment except that the motion detection unit 14 includes a pan / tilt region detection unit 20 as shown in FIG. Hereinafter, the noise reduction processing apparatus according to the present embodiment will not be described with respect to the points common to the first embodiment, and only different points will be described. Note that the same components as those in the first embodiment described above are denoted by the same reference numerals, and detailed description thereof is omitted.

  FIG. 3 is a block diagram illustrating a configuration example of the noise reduction processing apparatus according to the present embodiment. As shown in this figure, the motion detector 14 is provided with a pan / tilt area detector 20. The pan / tilt area detection unit 20 estimates the motion direction (x, y) of the detected motion vector and the absolute value of the movement amount, and the estimated motion vector has the same direction and the same size. In such a case, it is determined that the motion vector is generated by performing pan / tilt shooting in the camera device or the like that has captured the video signal, and this region (that is, the motion vector is in the same direction and the same direction). All areas having a size) are output to the area separation unit 12 as pan / tilt area information.

When the pan / tilt region information is input from the motion detection unit 14, the region separation unit 12 converts one frame image in the video signal from the image input unit 11 based on the pan / tilt region information from the motion detection unit 13. The pan / tilt area and the moving image area (areas other than the pan / tilt area correspond) are separated, and the pan / tilt area signal is transferred to the still image area NR processing unit 17 and the moving image area signal is transferred to the moving image area NR processing unit 18. Output each.
Thereby, the inter-surface NR processing is executed for the pan / tilt region signal by the still image region NR processing unit 17, while the in-plane NR processing is executed for the moving image region signal by the moving image region NR processing unit 18. It will be.
Here, in the still image region NR processing unit 17, inter-surface NR processing (frame addition) is performed on regions of the same pattern between frames. For the determination of the same picture area, for example, a binary integral value in the area when searching for a motion vector in the block is used.
In this way, the video signals after being subjected to NR processing are combined by the image combining unit 19 and output to the image output unit 16.

  As described above, according to the noise reduction processing device according to the present embodiment, the movement of the image due to pan / tilt of the camera device is regarded as a still image region, and the inter-surface NR (frame addition) is performed at the same portion between frames. ) Processing is performed, so that a high NR effect can be realized while leaving a fine portion of the image.

[Third Embodiment]
Next, a noise reduction processing apparatus according to the third embodiment of the present invention will be described with reference to the drawings. The third embodiment is basically the same as the first embodiment, but the still image region NR processing unit performs in-plane NR processing on the still image region in addition to the inter-surface NR processing. Is different. The still image region NR processing unit further includes a still image region NR coefficient determination unit that determines an NR coefficient when the in-plane NR processing is performed.
Hereinafter, the noise reduction processing apparatus according to the present embodiment will not be described with respect to the points common to the first embodiment, and only different points will be described. Note that the same components as those in the first embodiment described above are denoted by the same reference numerals, and detailed description thereof is omitted.

  4 and 5 are block diagrams showing a configuration example of the noise reduction processing apparatus according to the present embodiment. As shown in this figure, the noise reduction processing apparatus according to the present embodiment includes a still image region NR coefficient determination unit 21 that determines an NR coefficient used in the still image region NR processing unit 22. Further, as shown in FIG. 5, the still image region NR processing unit 22 performs inter-surface NR processing on the still image region, and performs surface processing on the still image region signal subjected to this inter-surface NR processing. The internal NR process is executed. Here, the determination of the NR coefficient by the still image region NR coefficient determination unit 21 is performed, for example, by the user specifying an arbitrary coefficient.

  As described above, according to the noise reduction processing device according to the present embodiment, the inter-surface NR processing and the in-plane NR processing are performed in the still image region, so the noise reduction processing according to the first embodiment described above. Compared with the apparatus, the NR effect in the still image region can be enhanced. Further, by performing in-plane NR processing, it is possible to obtain a high NR effect even if the number of added frames is small, and thus the number of frame memories can be reduced. For example, when noise suppression for SNR 6 dB improvement is performed by non-cyclic inter-surface NR processing, four frame memories are required because addition for four frames is necessary. However, as in the present embodiment, by using the inter-plane NR and the in-plane NR together, the inter-plane NR (3 dB improvement) + the in-plane NR (3 dB improvement) can be obtained. It is possible to improve 6 dB simply by using. As a result, since the number of frame additions is small, the processing speed is high, and there is an advantage that an image captured at a high frame rate can be handled.

[Fourth Embodiment]
Next, a noise reduction processing apparatus according to the fourth embodiment of the present invention will be described with reference to the drawings. The fourth embodiment is basically the same as the first embodiment, but includes a boundary region processing unit (boundary region processing means) for eliminating a difference in processing between the moving image region and the still image region. It is different.
Hereinafter, the noise reduction processing apparatus according to the present embodiment will not be described with respect to the points common to the first embodiment, and only different points will be described. Note that the same components as those in the first embodiment described above are denoted by the same reference numerals, and detailed description thereof is omitted.

FIG. 6 is a block diagram illustrating a configuration example of the noise reduction processing apparatus according to the present embodiment. As shown in this figure, the noise reduction processing device according to the present embodiment performs boundary region processing for eliminating the processing difference between the moving image region and the still image region in the combined image data output from the image combining unit 19. A boundary region processing unit 25 is provided.
For example, in the first embodiment described above, since different NR processing is performed for the moving image region and the still image region, there is a possibility that a noise reduction level difference occurs between the regions, resulting in an uncomfortable image. . In the present embodiment, in order to reduce this discomfort (discontinuity in image quality), the image gradually becomes smaller so that the difference in noise reduction level is not noticeable at the boundary between the moving image area and the still image area (hereinafter referred to as the boundary area). A boundary region processing unit 25 that performs processing that changes is provided to solve the above problem.

  Specifically, as shown in FIG. 7, the boundary region processing unit 25 detects a boundary region between the moving image region and the still image region, and detects a difference in noise reduction level between the still image region and the moving image region. A discontinuity detection unit (discontinuity detection means) 27 to detect and an in-plane NR processing unit (level difference reduction) for executing NR processing by changing the coefficient of in-plane NR processing for a region with a small noise reduction effect. Means) 28. Here, the boundary area is an area indicated by oblique lines in one frame shown in FIG. 8, for example, a neighboring area including a boundary between a still image area and a moving image area, specifically, a still image area. And an area having a predetermined width around the boundary line between the moving image area and the moving image area.

  Next, the in-plane NR processing unit 28 of the boundary region processing unit 25 will be described with reference to the drawings. For example, when one frame is divided into a plurality of blocks, as shown in FIG. 9, in the adjacent block D and block E, when the block D has less noise reduction effect than the block E, that is, the block D If the block E is noisy, the in-plane NR processing unit 28 performs in-plane NR processing on the block D. Here, there are cases where the block D is a still image area and the block E is a moving picture area, and conversely, the block D may be a moving picture area and the block E may be a still picture area. Here, the in-plane NR processing unit 28 executes the noise reduction process so that the closer to the boundary in the region D, the same the noise reduction level as that of the adjacent block E.

  That is, in the example shown in FIG. 9, in block D, the largest NR coefficient is used for the NR processing in the a region, and gradually smaller NR coefficients are used in the order of the b region, the c region, and the d region. As a result, the noise reduction effect can be made substantially the same in the area close to the boundary, and in the area far from the boundary, the NR processing is not applied or weakened, so that the details of the image can be reduced. It becomes possible to prevent deterioration.

  As described above, according to the noise reduction processing apparatus according to the present embodiment, it is possible to reduce a sense of incongruity near the boundary between the two due to the difference in NR processing between the moving image region and the still image region. Become. Thereby, the quality of an image can be improved.

[Fifth Embodiment]
Next, a noise reduction processing apparatus according to the fifth embodiment of the present invention will be described with reference to the drawings. The fifth embodiment is basically the same as the first embodiment, but differs in that the NR coefficient is changed according to the position in one frame.
Hereinafter, the noise reduction processing apparatus according to the present embodiment will not be described with respect to the points common to the first embodiment, and only different points will be described. Note that the same components as those in the first embodiment described above are denoted by the same reference numerals, and detailed description thereof is omitted.

  FIG. 10 is a block diagram illustrating a configuration example of the noise reduction processing apparatus according to the present embodiment. As shown in this figure, the noise reduction processing apparatus according to the present embodiment receives a video signal from an image input unit 11, detects a position in a frame based on the video signal, and uses this position information as a region separation unit 12. Is further provided.

In such a configuration, the image input unit 11 outputs the video signal to the region separation unit 12, the motion detection unit 14, and the screen position detection unit 30. The motion detection unit 14 divides one frame into a still image region and a moving image region from the input video signal, and outputs the divided region information to the screen position detection unit 30. Further, the motion detection unit 14 calculates a motion vector of the subject in the moving image area and outputs the motion vector to the screen position detection unit 30.
The screen position detection unit 30 detects the screen position within one frame from the video signal, and outputs the screen position information, the motion vector and the region information input from the motion detection unit 14 to the moving image region NR coefficient determination unit 15. . Further, the screen position detection unit 30 outputs the region information and the screen position information to the region separation unit 12.

  Based on the region information from the screen position detection unit 30, the region separation unit 12 separates the video signal from the image input unit 11 into a still image region signal and a moving image region signal, and converts the still image region signal into the still image region NR. In addition to outputting to the processing unit 17, the moving image region signal is output to the moving image region NR processing unit 18. On the other hand, the moving image region NR coefficient determination unit 15 determines an NR coefficient based on the motion vector and the screen position input from the screen position detection unit 30 and outputs them to the moving image region NR processing unit 18.

  Here, according to human visual characteristics, the gaze point in the natural monitoring state is at the center of the screen. In addition, although there is a characteristic of paying attention to a moving object, the image quality discrimination ability tends to be lowered due to the movement component of the object. Therefore, in the present embodiment, the central portion of the screen that collects the line of sight and the vicinity thereof are reduced in detail by reducing the in-plane NR processing for the important region, and conversely, the region of low importance ( Basically, the edge of the screen) enhances the noise effect by applying a stronger in-plane NR process. It should be noted that in the case of a moving image region with motion, it is difficult to understand the image quality deterioration, and therefore the NR coefficient (strength) may be increased.

The relationship between the position in the frame, the motion of the subject, and the NR coefficient, that is, the relationship between the motion vector and the noise reduction level is shown below.
a. Large movement around the frame> b. Medium movement around the frame> c. Still image around frame> d. Large movement in the center and near the center> e. Medium movement in the middle of the frame and near the center> f. Still images near the center of the frame and in the vicinity of the center Here, the symbols a to f correspond to the symbols a to f shown in FIG.
In this way, the noise reduction level is set high by using the largest NR coefficient in the moving image area where the motion vector is large in the peripheral part of the frame, and the smallest NR coefficient in the still picture area near the center of the frame and the center. Is used to set the noise reduction level low.

  As described above, according to the noise reduction processing device according to the present embodiment, noise reduction processing is performed using an optimal NR coefficient depending on the position in the frame. It is possible to reduce the noise reduction level and suppress the deterioration of details, while it is possible to increase the noise reduction level and increase the noise reduction effect for non-important areas where the line of sight is not pleasing.

[Sixth Embodiment]
Next, a noise reduction processing apparatus according to the sixth embodiment of the present invention will be described with reference to the drawings. The sixth embodiment is basically the same as the first embodiment, but differs in that the strength of NR is changed based on human noise detection characteristics.
Hereinafter, the noise reduction processing apparatus according to the present embodiment will not be described with respect to the points common to the first embodiment, and only different points will be described. Note that the same components as those in the first embodiment described above are denoted by the same reference numerals, and detailed description thereof is omitted.

  FIG. 12 is a block diagram illustrating a configuration example of the noise reduction processing device according to the present embodiment. As shown in this figure, the noise reduction processing apparatus according to the present embodiment receives a video signal from an image input unit 11, detects luminance in a frame based on this video signal, and uses this luminance information as a region separation unit 12. Is further provided.

In such a configuration, the image input unit 11 outputs the video signal to the region separation unit 12, the motion detection unit 14, and the luminance detection unit 31. The motion detection unit 14 divides one frame of the input video signal into a still image region and a moving image region, and outputs the division result to the luminance detection unit 31 as region information. Further, the motion detection unit 14 calculates a motion vector of the subject in the moving image area and outputs it to the luminance detection unit 31.
The luminance detection unit 31 detects the luminance in each region in one frame from the video signal, adds this luminance information to the motion vector from the motion detection unit 14, and outputs the information to the moving image region NR coefficient determination unit 15. Further, the luminance detection unit 31 outputs the region information and the luminance information to the region separation unit 12.

  Based on the region information from the screen position detection unit 30, the region separation unit 12 separates the video signal from the image input unit 11 into a still image region signal and a moving image region signal, and converts the still image region signal into the still image region NR. In addition to outputting to the processing unit 17, the moving image region signal is output to the moving image region NR processing unit 18. On the other hand, the moving image region NR coefficient determination unit 15 determines an NR coefficient based on the motion vector and the luminance information input from the luminance detection unit 31 and outputs the determined NR coefficient to the moving image region NR processing unit 18. The moving image region NR coefficient determination unit 15 determines the NR coefficient based on the following relationship.

That is, there are a portion where the noise looks good and a portion where the noise is difficult to see due to the luminance in the screen from the detection characteristics of the human eye. The human noise detection characteristic has a characteristic that a slightly brighter part is conspicuous than a completely dark part, and that it becomes less conspicuous when it becomes brighter.
Therefore, in the present embodiment, the portion of the luminance region where noise can be seen well, that is, in the intermediate luminance region where the luminance is at an intermediate level, the noise reduction level is increased by setting the NR coefficient to a large value so that the noise can be seen. For difficult areas, that is, areas other than the central luminance area, the noise reduction level is lowered by setting the NR coefficient to a small value.

The relationship between the luminance and the NR coefficient in the frame, that is, the relationship between the luminance and the noise reduction level is shown below.
Large movements where noise can be seen well> Medium movements where noise can be seen well> Still areas where noise can be seen well> Large movements where noise is difficult to see> Medium movements where noise is difficult to see> Noise Rest of hard-to-see parts

  As described above, according to the noise reduction processing device according to the present embodiment, noise reduction processing is performed according to luminance information. Therefore, highly effective NR processing can be performed on a portion where image noise is conspicuous, and noise is also generated. For inconspicuous portions, NR processing can be performed so that there is less deterioration of details. As described above, both the still image region and the moving image region in the frame can be subjected to appropriate NR processing according to the luminance, so that the noise reduction effect can be improved.

[Seventh Embodiment]
Next, a noise reduction processing apparatus according to the seventh embodiment of the present invention will be described with reference to the drawings. The seventh embodiment is basically the same as the first embodiment, but as shown in FIG. 13, the threshold value used when determining whether the region is a moving image region or a still image region is adjusted. The difference is that a threshold adjustment unit (threshold adjustment means) 32 is provided. Hereinafter, the noise reduction processing apparatus according to the present embodiment will not be described with respect to the points common to the first embodiment, and only different points will be described. Note that the same components as those in the first embodiment described above are denoted by the same reference numerals, and detailed description thereof is omitted.

  FIG. 13 is a block diagram illustrating a configuration example of the noise reduction processing apparatus according to the present embodiment. As shown in this figure, in the noise reduction processing apparatus according to the present embodiment, a video signal is input from the image input unit 11, and a still image region and a moving image region in a frame based on the video signal are designated in advance. A threshold adjustment unit 32 for determining based on the threshold is further provided.

In such a configuration, the image input unit 11 outputs the video signal to the region separation unit 12 and the motion detection unit 14. The motion detection unit 14 divides one frame from the input video signal into a still image region and a moving image region, and outputs the division result to the threshold adjustment unit 32 as region information. Further, the motion detection unit 14 calculates a motion vector of the subject in the moving image area and outputs the motion vector to the threshold adjustment unit 32.
The threshold adjustment unit 32 compares the specified threshold value with the size of the motion vector, re-divides an area where the motion vector is less than the threshold value as a still image area, and an area where the motion vector is greater than the threshold value as a moving image area. The region information is input to the region separation unit. Further, the threshold adjustment unit 32 outputs the motion vector of the moving image region to the moving image region NR coefficient determination unit 15. Here, the threshold value of the threshold adjustment unit 32 is determined by the user.

  As described above, according to the noise reduction processing apparatus according to the present embodiment, the user can adjust the threshold value that is used as a reference when the still image region and the moving image region are classified. As a result, an image having an image quality desired by the user can be generated.

  In the above embodiment, the threshold value may be automatically adjusted by selecting a mode, for example, instead of the user setting the threshold value directly. For example, a moving image emphasis mode and a still image emphasis mode may be provided, and a threshold value suitable for each mode may be automatically set by the user selecting a desired mode among these modes. As a specific mode setting mode, for example, a selection button (switch) for setting a threshold value is provided on a moving image camera or a digital camera equipped with the noise reduction processing device, and the user operates the selection button. Thus, the setting of the threshold value is realized.

[Eighth Embodiment]
Next, a noise reduction processing apparatus according to the eighth embodiment of the present invention will be described with reference to the drawings. Although the eighth embodiment is basically the same as the first embodiment, as shown in FIG. 14, a frequency component detection unit (frequency component detection means) 33 that detects a frequency component of one frame image is provided. It differs in the point to prepare.
Hereinafter, the noise reduction processing apparatus according to the present embodiment will not be described with respect to the points common to the first embodiment, and only different points will be described. Note that the same components as those in the first embodiment described above are denoted by the same reference numerals, and detailed description thereof is omitted.

  FIG. 14 is a block diagram illustrating a configuration example of the noise reduction processing apparatus according to the present embodiment. As shown in this figure, the noise reduction processing apparatus according to the present embodiment receives a video signal from the image input unit 11, detects a frequency component in a frame based on the video signal, and uses this frequency component as a region separation unit. 12 and a frequency component detection unit 33 that outputs to the moving image region NR coefficient determination unit 15.

In such a configuration, the image input unit 11 outputs the video signal to the region separation unit 12, the motion detection unit 14, and the frequency detection unit 33. The motion detection unit 14 divides one frame of the input video signal into a still image region and a moving image region, and outputs the division result to the frequency component detection unit 33 as region information. Further, the motion detection unit 14 calculates a motion vector of the subject in the moving image area and outputs the motion vector to the frequency detection unit 33.
The frequency detection unit 33 detects a frequency component in one frame from the video signal, adds the frequency component to the motion vector from the motion detection unit 14, and outputs it to the moving image region NR coefficient determination unit 15. Further, the frequency component detection unit 33 outputs the region information and the frequency component information to the region separation unit 12. Here, for example, the frequency component detection unit 33 divides one frame image into a plurality of blocks, and detects a frequency component by performing Fourier transform on each block.

Based on the region information from the frequency component detection unit 33, the region separation unit 12 separates the video signal from the image input unit 11 into a still image region signal and a moving image region signal, and converts the still image region signal into the still image region NR. In addition to outputting to the processing unit 17, the moving image region signal is output to the moving image region NR processing unit 18.
The moving image region NR coefficient determination unit 15 determines an NR coefficient based on the motion vector and frequency component information input from the frequency component detection unit 33 and outputs the determined NR coefficient to the moving image region NR processing unit 18. As a result, the NR processing of the moving image region using the NR coefficient in which not only the motion vector but also the frequency component is added is executed. At this time, the moving image region NR coefficient determination unit 15 sets the NR coefficient to be lower as the frequency component is higher.

  As described above, according to the noise reduction processing apparatus according to the present embodiment, the noise reduction level for the high-frequency component region is lowered, so that the NR processing can be performed with less deterioration of image details. As a result, it is possible to perform moderate noise reduction while leaving details by reducing the noise reduction level for the high frequency region, and to enhance the noise reduction effect by increasing the noise reduction level for the low frequency region. Can do.

[Ninth Embodiment]
Next, a noise reduction processing apparatus according to the ninth embodiment of the present invention will be described with reference to the drawings. The ninth embodiment is basically the same as the first embodiment, but differs in that a camera shake correction unit (camera shake correction unit) 35 that performs camera shake correction is provided as shown in FIG.
According to such a configuration, the video signal from the image input unit 11 is input to the camera shake correction unit 35, so that the influence of the camera shake is eliminated, and the video signal after the camera shake correction is converted into the region separation unit 12 and the motion detection unit. 14 is output. As a result, the above-described noise reduction processing is performed on the video signal after the camera shake correction.
As described above, according to the noise reduction processing apparatus according to the present embodiment, motion detection is performed based on the video signal subjected to camera shake correction, so that there is no erroneous recognition of subject motion and motion due to camera shake, thereby reducing noise. Can be increased.

[Tenth embodiment]
Next, a noise reduction processing apparatus according to the tenth embodiment of the present invention will be described with reference to the drawings. The tenth embodiment includes a configuration in which the components of the noise reduction processing devices according to the first to ninth embodiments described above are combined.
Hereinafter, the noise reduction processing apparatus according to the present embodiment will be described with reference to FIG.
As shown in FIG. 16, the video signal output from the image input unit 11 to the camera shake correction unit 35 is subjected to camera shake correction by the camera shake correction unit 35, and the corrected video signal is converted into the region separation unit 12 and the motion detection unit. 14, and input to the screen position detection unit 30, the luminance detection unit 31, and the frequency component detection unit 33.

  The motion detection unit 14 divides one frame of the input video signal into a still image region and a moving image region door, and outputs the division result to the threshold adjustment unit 32 as region information. Further, the motion detection unit 14 calculates a motion vector of the subject in the moving image area, and outputs this motion vector to the threshold adjustment unit 32. The threshold adjustment unit 32 compares the threshold specified by the user with the motion vector detected by the motion detection unit 14, sets a region where the motion vector is equal to or greater than the threshold as a moving image region, and sets a region where the motion vector is within the threshold as a still image. As a region, the region is repartitioned. The threshold adjustment unit 32 outputs the motion vector input from the motion detection unit 14 and the result of the re-segmentation to the screen position detection unit 30 and the moving image region NR coefficient determination unit 15 as region information.

  The screen position detection unit 30 detects the screen position in one frame from the video signal from the camera shake correction unit 35 and adds this screen position information to the motion vector and area information transferred from the threshold adjustment unit 32 to obtain the luminance. Output to the detector 31. The luminance detection unit 31 detects luminance in each region in one frame from the video signal from the camera shake correction unit 35, and this luminance information and information from the screen position detection unit 30 are a motion vector, screen position information, The region information is output to the frequency component detection unit 33, and the motion vector, the screen position information, and the luminance information are output to the moving image region NR coefficient determination unit 15. The frequency component detection unit 33 detects a frequency component in one frame from the video signal from the camera shake correction unit 35, and the motion vector, screen position information, luminance information, which is information from the frequency component information and the luminance detection unit 31, The region information is output to the region separation unit 12 and the motion vector, screen position information, luminance information, and frequency component information are output to the moving image region NR coefficient determination unit 15. The moving image region NR coefficient determination unit 15 determines an NR coefficient based on the motion vector, screen position information, luminance information, and frequency component information, and outputs the determined NR coefficient to the moving image region NR processing unit 18. Specifically, the moving image region NR coefficient determination unit 15 sets the NR coefficient by executing the determination processing in steps SA5 to SA19 in the flowcharts illustrated in FIGS.

  The region separation unit 12 separates the video signal input from the camera shake correction unit 35 into a still image region signal and a moving image region signal based on the region information from the frequency component detection unit 33, and converts the still image region signal into a still image. In addition to outputting to the region NR processing unit 22, the moving image region signal is output to the moving image region NR processing unit 18.

As illustrated in FIG. 17, the still image region NR processing unit 22 performs inter-surface NR processing on the still image region signal input from the region separation unit 12, and then further performs a still image region NR coefficient determination unit 21. The in-plane NR process is executed using the NR coefficient input from. Here, the NR coefficient input from the still image region NR coefficient determination unit 21 is determined based on a motion vector, screen position information, luminance information, and frequency component information. More specifically, the still image region NR coefficient determination unit 15 determines the NR coefficient by executing the determination process of steps SA20 to SA34 in the flowcharts shown in FIGS.
The still image region NR processing unit 22 outputs the still image region signal after the NR processing to the image composition unit 19.
On the other hand, the moving image region NR processing unit 18 performs in-plane NR processing on the moving image region signal input from the region separating unit 12 using the NR coefficient input from the moving image region NR coefficient determining unit 15, and NR The processed moving image area signal is output to the image composition unit 19.

  The image composition unit 19 creates a composite image signal of one frame by combining the still image region signal and the moving image region signal, and outputs this composite image signal to the boundary region processing unit 25 (see FIG. 16). The boundary region processing unit 25 performs NR processing on the boundary region between the still image region and the moving image region, thereby reducing the uncomfortable feeling in the boundary region, and outputs the processed image signal to the image output unit 16.

  As described above, according to the noise reduction processing device according to the present embodiment, appropriate NR processing can be performed on a still image region and a moving image region in a frame. A high NR effect can be provided while remaining. Thereby, it is possible to provide a high-quality image.

  In the above-described embodiment, the case where all the components are provided has been described, but these components can be appropriately selected and adopted. Moreover, it is good also as a structure which a user can select suitably which component is used to implement | achieve noise reduction processing. This can take the form which provides selection parts, such as a switch, in the device which mounts the noise reduction processing device, for example.

  In the above-described embodiment, hardware processing is assumed as the noise reduction processing device, but the present invention is not limited to such a configuration. For example, a configuration in which processing is performed separately by software is possible. In this case, the noise reduction processing apparatus includes a main storage device such as a CPU and a RAM, and a computer-readable recording medium on which a program for realizing all or part of the above processing is recorded. Then, the CPU reads out the program recorded in the storage medium and executes information processing / calculation processing, thereby realizing the same processing as the above-described noise reduction processing device.

  Here, the computer-readable recording medium means a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, a semiconductor memory, or the like. Alternatively, the computer program may be distributed to the computer via a communication line, and the computer that has received the distribution may execute the program.

Hereinafter, the processing procedure of the noise reduction processing method realized by the CPU executing the noise reduction processing program will be described with reference to FIGS.
First, in step SA1 in FIG. 18, camera shake correction is performed on the video signal input from the input image portion, and in the subsequent step SA2, a motion vector is calculated from two consecutive frames. Subsequently, in step SA3, a screen position indicating what position on the screen the region where the motion vector has been calculated is detected, and a luminance value and a frequency component are detected.

  Subsequently, in step SA4, it is determined whether or not the motion vector is equal to or greater than a preset threshold value. As a result, when the motion vector is greater than or equal to a preset threshold value, it is recognized as a moving image region, and when it is less than the threshold value, it is recognized as a still image region. Subsequently, if it is a moving image area, the process proceeds to step SA5 to determine whether the area is a screen attention area. As a result, if it is the screen attention area, the process proceeds to step SA6, and it is determined whether or not the brightness value of the area is a brightness where noise is not noticeable.

  Subsequently, when the luminance is conspicuous, the process proceeds to step SA7, where it is determined whether or not the frequency component in the region is high. As a result, when the frequency component is high, the process proceeds to step SA8, and the in-plane NR process using the NR coefficient a1 is executed. When the frequency component is not high, the process proceeds to step SA9 and the NR coefficient b1 is used. In-plane NR processing is executed. On the other hand, if the luminance is not inconspicuous in step SA6, the process proceeds to step SA10 to determine whether the frequency component in the region is high. As a result, when the frequency component is high, the process proceeds to step SA11 and the in-plane NR process using the NR coefficient c1 is executed. On the other hand, when the frequency component is not high, the process proceeds to step SA12 and the NR coefficient d1 is set. The in-plane NR process used is executed.

  In step SA5 described above, if the area is not a screen attention area, the process proceeds to step SA13 (see FIG. 19), and it is determined whether or not the brightness is inconspicuous. Subsequently, when the luminance is conspicuous, the process proceeds to step SA14 to determine whether or not the frequency component in the region is high. As a result, if the frequency component is high, the process proceeds to step SA15 and the in-plane NR process using the NR factor e1 is executed. On the other hand, if the frequency component is not high, the process proceeds to step SA16 and the NR coefficient f1 is set. The in-plane NR process used is executed.

On the other hand, if the luminance is not inconspicuous in step SA13, the process proceeds to step SA17, where it is determined whether the frequency component in the region is high. As a result, when the frequency component is high, the process proceeds to step SA18 and the in-plane NR process using the NR coefficient g1 is executed. On the other hand, when the frequency component is not high, the process proceeds to step SA19 and the NR coefficient h1 is set. The used in-plane NR process is executed.
Here, the following relationship is established among the NR coefficients a1, b1, c1, d1, e1, f1, g1, and h1.
h1>g1>f1>e1>d1>c1>b1> a1

  On the other hand, if the motion vector is less than the threshold value in step SA4 in FIG. 18, the region is recognized as a still image region, and the process proceeds to step SA20 (see FIG. 20). In step SA20, it is determined whether the area is a screen attention area. As a result, if it is a screen attention area, the process proceeds to step SA21, and it is determined whether or not the brightness value of the area is a brightness where noise is not noticeable. Subsequently, when the luminance is a conspicuous luminance, the process proceeds to step SA22, and it is determined whether or not the frequency component in the region is high. As a result, if the frequency component is high, the process proceeds to step SA23, where the inter-plane NR process and the in-plane NR process using the NR coefficient a2 are performed. On the other hand, if the frequency component is not high, the process proceeds to step SA24. Then, the in-plane NR process and the in-plane NR process using the NR coefficient b2 are executed.

  On the other hand, if the brightness is not inconspicuous in step SA21, the process proceeds to step SA25 to determine whether or not the frequency component in the region is high. As a result, if the frequency component is high, the process proceeds to step SA26, where the inter-plane NR process and the in-plane NR process using the NR coefficient c2 are performed. On the other hand, if the frequency component is not high, the process proceeds to step SA27. Then, the in-plane NR process and the in-plane NR process using the NR coefficient d2 are executed.

  In step SA20, if the area is not a screen attention area, the process proceeds to step SA28, where it is determined whether or not the brightness is inconspicuous. Subsequently, if the luminance is a conspicuous luminance, the process proceeds to step SA29 to determine whether or not the frequency component in the region is high. As a result, when the frequency component is high, the process proceeds to step SA30, where the in-plane NR process and the in-plane NR process using the NR coefficient e2 are performed. On the other hand, if the frequency component is not high, the process proceeds to step SA31. Then, the in-plane NR process and the in-plane NR process using the NR coefficient f2 are executed.

On the other hand, if the luminance is not inconspicuous in step SA28, the process proceeds to step SA32 to determine whether or not the frequency component in the region is high. As a result, when the frequency component is high, the process proceeds to step SA33, where the in-plane NR process and the in-plane NR process using the NR coefficient g2 are performed. On the other hand, if the frequency component is not high, the process proceeds to step SA34. Then, the in-plane NR process and the in-plane NR process using the NR coefficient h2 are executed.
Here, the following relationship is established among the above-described NR coefficients a2, b2, c2, d2, e2, f2, g2, and h2.
h2>g2>f2>e2>d2>c2>b2> a2

  When NR processing using appropriate NR coefficient processing is executed in this way, the image signal after NR processing is synthesized in step SA35 to create a composite image signal for one frame, and in subsequent step SA36. Then, boundary processing is performed on the composite image signal, and the processed composite image signal is output.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes design changes and the like within a scope not departing from the gist of the present invention.

It is the block diagram which showed schematic structure of the noise reduction processing apparatus which concerns on the 1st Embodiment of this invention. It is the figure which showed the internal function of the noise reduction process part shown in FIG. It is the block diagram which showed schematic structure of the noise reduction processing apparatus which concerns on the 2nd Embodiment of this invention. It is the block diagram which showed schematic structure of the noise reduction processing apparatus which concerns on the 3rd Embodiment of this invention. It is the figure which showed the internal function of the noise reduction process part shown in FIG. It is the block diagram which showed schematic structure of the noise reduction processing apparatus which concerns on the 4th Embodiment of this invention. It is the figure which showed the internal function of the boundary area | region process part of FIG. It is a figure for demonstrating the boundary area process in the boundary area process part of FIG. It is a figure for demonstrating the process of the in-plane NR process part in the boundary area | region process part of FIG. It is the block diagram which showed schematic structure of the noise reduction processing apparatus which concerns on the 5th Embodiment of this invention. It is the figure which showed the relationship between the NR coefficient determined by the moving image area NR coefficient determination part of FIG. 10, and a position. It is the block diagram which showed schematic structure of the noise reduction processing apparatus which concerns on the 6th Embodiment of this invention. It is the block diagram which showed schematic structure of the noise reduction processing apparatus which concerns on the 7th Embodiment of this invention. It is the block diagram which showed schematic structure of the noise reduction processing apparatus which concerns on the 8th Embodiment of this invention. It is the block diagram which showed schematic structure of the noise reduction processing apparatus which concerns on the 9th Embodiment of this invention. It is the block diagram which showed schematic structure of the noise reduction processing apparatus which concerns on the 10th Embodiment of this invention. It is the figure which showed the internal function of the noise reduction process part shown in FIG. It is the flowchart which showed an example of the process sequence implement | achieved by the noise reduction processing apparatus which concerns on the 10th Embodiment of this invention. It is the flowchart which showed an example of the process sequence implement | achieved by the noise reduction processing apparatus which concerns on the 10th Embodiment of this invention. It is the flowchart which showed an example of the process sequence implement | achieved by the noise reduction processing apparatus which concerns on the 10th Embodiment of this invention. It is a figure explaining the block division | segmentation of the determination method 1 implemented in the motion detection part which concerns on the 1st Embodiment of this invention. It is a figure explaining the block division | segmentation of the determination method 2 implemented in the motion detection part which concerns on the 1st Embodiment of this invention. It is the flowchart which showed an example of the process sequence of the determination method 3 implemented in the motion detection part which concerns on the 1st Embodiment of this invention. It is a figure explaining the block division | segmentation of the determination method 3 implemented in the motion detection part which concerns on the 1st Embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 Image input part 12 Area separation part 13 Noise reduction process part 14 Motion detection part 15 Movie area NR coefficient determination part 16 Image output part 17 Still image area NR process part 18 Movie area NR process part 19 Image composition part 20 Pan / tilt area Detection unit 21 Still image region NR coefficient determination unit 22 Still image region NR processing unit 25 Boundary region processing unit 26 Boundary region detection unit 27 Discontinuity detection unit 28 In-plane NR processing unit 30 Screen position detection unit 31 Luminance detection unit 32 Threshold value Adjustment unit 33 Frequency component detection unit 35 Camera shake correction unit 50 Area division unit

Claims (18)

An input means for inputting information;
Area dividing means for dividing one frame of the video signal input from the input means into a moving image area and a still image area;
A noise reduction processing device comprising: noise reduction processing means for executing noise reduction processing of a different technique on the moving image region and the still image region.   The noise reduction processing apparatus according to claim 1, wherein the noise reduction processing unit performs inter-plane processing for the still image region and performs in-plane processing for the moving image region.   The noise reduction processing apparatus according to claim 1, wherein the noise reduction processing unit performs inter-plane processing and in-plane processing on the still image area.   The noise reduction processing device according to claim 1, wherein the noise reduction processing unit varies a noise reduction level according to a magnitude of a motion vector for the moving image area. The area segmentation means classifies an area having motion vectors in the same direction and the same size as the pan / tilt area in the one frame,
5. The noise reduction processing device according to claim 1, wherein the noise reduction processing unit performs the same noise reduction processing as that of the still image area on the pan / tilt area.   The noise reduction processing according to claim 1, further comprising boundary region processing means for detecting a boundary region between the moving image region and the still image region and reducing a difference in noise reduction level in the boundary region. apparatus. The boundary region processing means includes
Discontinuity detecting means for detecting a difference in noise reduction level in the boundary region;
Level difference reducing means for executing noise reduction processing according to the difference of the noise reduction level detected by the discontinuity detecting means for a high noise area that is a low noise reduction level in the boundary area; The noise reduction processing apparatus according to claim 6.   The noise reduction processing device according to claim 1, wherein the noise reduction processing unit varies the noise reduction level according to a position in the one frame.   In the one frame, the noise reduction processing means makes the noise reduction level the lowest for the central portion and the still image region in the vicinity of the central portion, and the moving portion in the central portion and in the vicinity of the central portion, the outer peripheral portion and the rest in the vicinity of the outer peripheral portion The noise reduction processing apparatus according to claim 8, wherein the noise reduction level is gradually increased in the order of the image area, the outer periphery, and the moving image area in the vicinity of the outer periphery.   The noise reduction processing device according to any one of claims 1 to 9, wherein the noise reduction processing unit varies the noise reduction level according to luminance in the one frame.   The noise reduction processing device according to claim 10, wherein the noise reduction processing unit performs a noise reduction process having a higher noise reduction level than the other areas on an intermediate luminance area whose luminance is at an intermediate level. Having a frequency detection means for detecting a frequency component of the moving image area information;
The noise reduction processing device according to any one of claims 1 to 11, wherein the noise reduction processing means lowers a noise reduction level in the high frequency component region than a noise reduction level in the low frequency component region.   The noise according to any one of claims 1 to 12, further comprising a threshold adjustment unit that adjusts a threshold used when the region dividing unit divides the frame into the moving image region and the still image region. Reduction processing device. With mode setting means that can set the video mode that emphasizes video and the still image mode that emphasizes still images,
The noise reduction processing apparatus according to claim 13, wherein the threshold adjustment unit adjusts the threshold according to a mode set by the mode setting unit. An input process in which information is input;
An area separation process for dividing one frame of the input video signal into a moving image area and a still image area,
A noise reduction processing method comprising: a noise reduction processing step of performing noise reduction processing of a different method between the moving image region and the still image region. An input step in which information is entered;
A region separation step of dividing one frame of the input video signal into a moving image region and a still image region;
A noise reduction program for causing a computer to execute a noise reduction processing step of executing noise reduction processing of a different technique between the moving image area and the still image area. Optical means for imaging a subject;
Imaging means for photoelectrically converting the subject image formed by the optical means;
Input means for inputting a video signal photoelectrically converted by the imaging means;
Area dividing means for dividing one frame of the video signal output from the input means into a moving image area and a still image area;
A camera apparatus comprising: noise reduction processing means for performing noise reduction processing of a different technique between the moving image area and the still image area. A camera shake correction unit that corrects blurring of an image that occurs during an exposure period of the imaging unit;
The camera device according to claim 17, wherein the region sorting unit divides one frame of the video signal corrected by the camera shake correction unit into the moving image region and the still image region.

Images