JPH04234266A - Noise reducing processor for moving image - Google Patents

Abstract

PURPOSE:To reduce a random noise and a burst noise by outputting the maximum value out of the minimum values in the combination of picture element values as the value of a picture element targeted to process in a frame in which the picture element targeted to process is included. CONSTITUTION:Assuming the values of the picture elements located at the upper and lower sides for the picture element targeted to process as U, D, the values of the picture elements located at the same position as that of the picture element targeted to process in frames before and behind the frame temporarily in which the picture element targeted to process is included as P and L, and the value of the picture element targeted to process as A in the frame in which the picture element targeted to process is included, a selection circuit 2 selects a prescribed pairs of combination of at least one of U, D, P, and L and A. A generation circuit 4 outputs the combination with a small value out of the combination of two values. An output circuit 6 receives each output value of the circuit 4, and outputs the maximum value as the value of the picture element targeted to process.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving image noise reduction processing apparatus.

0002

2. Description of the Related Art There are the following noises that occur in moving images.

(A) Random Noise This noise is isolated in the time direction and also in the horizontal and vertical directions within the same frame.

(B) Burst noise This noise is isolated in the time direction and vertically within the same frame (that is, from the pixels in the horizontal lines above and below the horizontal line containing the pixel to be processed). However, they are continuous in the horizontal direction of the same frame.

[0005] Conventionally, there are the following methods for reducing such noise in a moving image consisting of a plurality of temporally consecutive frames, each frame consisting of a plurality of pixels.

(1) A method using a temporal filter that performs linear filtering in the time direction. According to this method, the value of the pixel to be processed is averaged with the values of temporally preceding and succeeding pixels, and noise is dispersed.

(2) A method using a smoothing filter that limits the frequency band within one frame, ie, spatially. According to this method, noise having unnecessary frequency components is removed.

(3) A method using a spatial nonlinear filter.

[0009]

[Problem to be solved by the invention] The above-mentioned conventional (1),
Methods (2) and (3) are the same as those described in (A) and (B) above.
) cannot be reduced, and there is a risk that edges of significant images may be degraded. An object of the present invention is to provide a moving image noise reduction processing device that can reduce both random noise and burst noise without deteriorating the edges of a significant image.

0010

[Means for Solving the Problems] A moving picture noise reduction processing device configured as claimed in claim 1 is a moving picture noise reduction processing device consisting of a plurality of temporally consecutive frames, each frame of which is composed of a plurality of pixels. In the frame that includes the pixel to be processed, the values of the pixels located above and below the pixel to be processed are U and D, respectively, and the processing in the frames temporally before and after the frame that includes the pixel to be processed is calculated. The value of the pixel at the same position as the target pixel is P and L,
When the value of a pixel to be processed is A, means for selecting a predetermined set of combinations of A and at least one of U, D, P, and L, and for each of the selected combinations, a minimum value among the combinations. It is characterized by comprising means for generating the maximum value among the minimum values and means for outputting the maximum value among the minimum values as the value of the pixel to be processed.

The moving picture noise reduction processing device according to claim 2 is a moving picture noise reduction processing device consisting of a plurality of temporally consecutive frames, each frame of which is composed of a plurality of pixels, and wherein the processing target pixels are The values of the pixels located above and below the pixel to be processed in the included frame are set to U and D, respectively, and the values of the pixels located above and below the pixel to be processed are set to the same position as the pixel to be processed in the frames temporally before and after the frame in which the pixel to be processed is included. When the values of a certain pixel are P and L, and the value of the pixel to be processed is A, at least one of U, D, P, and L
means for selecting a predetermined set of combinations of A and A; means for generating the maximum value in each selected combination; and means for outputting the minimum value of the maximum values as the value of the pixel to be processed. It is characterized by having the following.

0012

[Operation] In the moving image noise reduction processing device configured as claimed in claim 1, the values U and D of pixels located spatially perpendicular to the pixel to be processed, and the values before and after the pixel to be processed in the temporal direction. A predetermined set of combinations of at least one of the values P and L of the pixel located at , and the value A of the pixel to be processed is selected, and for each of these selected combinations,
The minimum value in the combination is determined, and the maximum value among these determined minimum values is output as the value of the pixel to be processed. Both random noise and burst noise can be reduced.

[0013] In the moving image noise reduction processing device having the structure of claim 2, the values U and D of pixels located spatially perpendicular to the pixel to be processed, as well as the values U and D of pixels located before and after the pixel to be processed in the temporal direction. A predetermined set of combinations of at least one of the values P and L of the pixel located at Since the minimum value among the maximum values obtained is output as the value of the pixel to be processed, it is possible to reduce both random noise and burst noise in the zero direction (that is, in the negative direction) without degrading the edges of the significant image. I can do it.

[0014]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows the structure of a first embodiment of a moving image noise reduction processing apparatus according to the present invention. This embodiment is an example in which the present invention is applied to a moving image in which one frame is composed of a plurality of lines (for example, 525) and one line includes a plurality of effective pixels (for example, 720). This embodiment includes a pixel combination selection circuit 2, a minimum value set generation circuit 4, and a maximum value output circuit 6.
Consists of. The pixel combination selection circuit 2 sets the values of pixels located above and below the pixel in the frame including the pixel to be processed as U and D, respectively, in the frames temporally before and after the frame including the pixel to be processed. When the values of pixels located at the same position as the pixel to be processed are P and L, and the value of the pixel to be processed is A, a combination of A and at least one of U, D, P, and L is set as a predetermined set (in this example, So, (
A, U) (A, D), (A, P) and (A, L) 4
combination).

The pixel combination selection circuit 2 includes one line delay selection circuits (LD) 21, 23, 24 and 26 and one frame delay circuits (FD) 22 and 25. 1
The line delay circuit 21 delays the value of the digitized input pixel by one line so that the pixel to be processed is processed after the frame including the pixel to be processed (in this example, one frame later). The value L of the pixel at the same position as is output. The one-frame delay circuit 22 outputs the value D of a pixel located below (in this example, one line below) the pixel to be processed in a frame that includes this pixel. The 1-line delay circuit 23 receives the output of the 1-frame delay circuit 22 and outputs the value A of the pixel to be processed. The 1-line delay circuit 24 receives the output of the 1-line delay circuit 23 and calculates the value U of the pixel located above (in this example, 1 line above) the pixel in the frame that includes the pixel to be processed.
Output. The one-frame delay circuit 25 delays the output of the one-frame delay circuit 22 by one frame. The 1-line delay circuit 26 receives the output of the 1-frame delay circuit 25 and is located at the same position as the processing target pixel temporally previous (in this example, one frame before) with respect to the frame containing the processing target pixel. Outputs the pixel value P.

FIG. 2 shows the spatial and temporal positional relationships of A, U, D, P and L output by the pixel combination selection circuit 2.

The minimum value set generating circuit 4 includes four minimum value output circuits (MIN) 42, 44, 46 and 48. The minimum value output circuit 42 outputs a combination of two values (A, L
), the smaller value min(A
,L). The minimum value output circuit 44 receives the combination of two values (A, D) and outputs the smaller value min(A, D) of these values. Minimum value output circuit 4
6 receives a combination of two values (A, U) and outputs the smaller value min(A, U). The minimum value output circuit 48 receives the combination of two values (A, P) and outputs the smaller value min(A, P) of these values.
P) is output. The maximum value output circuit (MAX) 6 receives the output values of the four minimum value output circuits 42, 44, 46 and 48, and calculates the maximum value of these values A1=max{min(
A, U), min (A, D), min (A, L), mi
n(A,P)} is output as the value of the pixel to be processed.

In the moving image noise reduction processing apparatus shown in FIG. Combinations (A, U), (A
, D), (A, L), and (A, P) are selected, and the minimum value is found for each of these combinations,
Since the maximum value among the minimum values obtained is output as the pixel to be processed, both random noise and burst noise in the peak direction (i.e., positive direction) as shown in Fig. 3 can be suppressed without deteriorating the edges of the significant image. can be reduced. To explain the principle of noise removal in more detail with reference to FIG. 4, the original pixel signal sequence contained no noise, but due to burst noise being added for some reason, the signal level value changed from 10 to 30. Even if the noise level increases, the noise level 30 can be removed and the original signal level 10 can be reproduced by taking the minimum value of the adjacent pixels and then taking the larger of the adjacent minimum values.

FIG. 5 shows the configuration of a second embodiment of the moving image noise reduction processing apparatus of the present invention. In this embodiment, the output values L, D, A,
It includes a maximum value set generation circuit 7 that receives U and P, and a minimum value output circuit 8. Maximum value set generation circuit 7 has four maximum value output circuits 72, 74, 76 and 78. The maximum value output circuit 72 receives the combination of two values (A, L) and outputs the larger value max (A, L) of these values. The maximum value output circuit 74 outputs a combination of two values (
A, D), the larger of these values m
Output ax(A,D). The maximum value output circuit 76 has 2
It receives a combination of two values (A, U) and outputs the larger value max (A, U) of these values. The maximum value output circuit 78 receives the combination of two values (A, P) and outputs the larger value max (A, P) of these values. The minimum value output circuit 8 includes four maximum value output circuits 72
, 74, 76, and 78, the minimum value of these values A2=min{max(A,U), max(
A, D), max (A, L), max (A, P)} as the value of the pixel to be processed.

In the moving image noise reduction processing device shown in FIG. A combination (A, U) of the values P and L of the pixels located before and after each and the value A of the pixel to be processed,
(A, D), (A, L) and (A, P) are selected,
The maximum value is determined for each of these combinations, and the minimum value among the determined maximum values is output as the pixel to be processed. Therefore, the zero direction as shown in Figure 6 is obtained without deteriorating the edges of the significant image. (ie, in the negative direction) and burst noise can be reduced.

FIG. 7 shows the configuration of a third embodiment of the moving image noise reduction processing apparatus of the present invention. This embodiment includes the minimum value set generation circuit 4 and maximum value output circuit 6 shown in FIG. 1, the maximum value generation circuit 7 and minimum value output circuit 8 shown in FIG. 5, and the same pixel as in FIG. The output values L, D, of the combination selection circuit 2
It includes an averaging circuit 9 that receives A, U, and P and outputs an average value AVE of these values, subtracters 10 and 11, a comparator 12, and a selector 13. The subtractor 10 is
Output A1 of minimum value output circuit 6 and output AV of average circuit 9
Outputs the difference from E. The subtracter 11 outputs the difference between the output A2 of the minimum output circuit 8 and the output AVE of the average circuit 9. The comparator 12 compares the outputs of the subtractors 10 and 11, and outputs "1" when the former is larger, and outputs "0" when the latter is larger. The selector 13 is the comparator 1
When receiving “1” from 2, the output A of the minimum value output circuit 8
2 is output as the value A3 of the pixel to be processed. Further, when the selector 13 receives "0" from the comparator 12, it outputs the output A1 of the maximum value output circuit 6 as the value A3 of the pixel to be processed.

In short, the embodiment shown in FIG. 7 selects the output value A1 of the maximum value output circuit 6 and the output value A2 of the minimum value output circuit 8, whichever is closer to the average value of L, D, A, U, and P. Since the value of is outputted as the value A3 of the pixel to be processed, both the peak direction noise shown in FIG. 3 and the zero direction noise shown in FIG. 6 can be removed.

Note that the embodiments shown in FIGS. 1, 5, and 7 described above have a combination of one of U, D, P, and L with A (for example, (A, U), (A, D )) is selected, but
For example, a combination of two of U, D, P, and L and A (for example, (A, U, D), etc.) may be selected. However, if a combination of three or more values is used, the image will look flat, so a combination of two values including A is preferable.

Furthermore, although the above-described embodiment uses pixels above and below one line, pixels above and below a plurality of lines, such as two or three lines, may be used.

Further, in the above-described embodiment, frames before and after one frame are used, but frames before and after a plurality of frames, such as two or three frames, may be used.

[0026]

As described above, according to the moving image noise reduction processing device according to claim 1, it is possible to perform nonlinear filtering on a moving image data string spatially in the vertical direction and in the temporal direction. Both random noise and burst noise in the peak direction (ie, in the positive direction) can be reduced without deteriorating the edges of the image.

According to the moving image noise reduction processing apparatus according to the second aspect of the present invention, both random noise in the zero direction (negative direction) and burst noise can be reduced without deteriorating the edges of a significant image.

[0028] Furthermore, both the apparatuses according to claims 1 and 2 can obtain the same noise reduction effect even when the noise increases, and can perform noise reduction processing independently of other image processing. can.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a first embodiment of a moving image noise reduction processing device of the present invention.

[Fig. 2] An explanatory diagram showing the spatial and temporal positional relationship of pixels used in the apparatus of Fig. 1.

[Figure 3] Waveform diagram showing an example of noise that can be reduced by the embodiment in Figure 1

FIG. 4 is an explanatory diagram of the noise reduction operation of the embodiment in FIG. 1.

[Figure 5]
A block diagram showing a second embodiment of the moving image noise reduction processing device of the present invention

FIG. 6 is a waveform diagram showing an example of noise reduced by the embodiment of FIG. 5;

FIG. 7 is a block diagram showing a third embodiment of the moving image noise reduction processing device of the present invention.

[Explanation of symbols]

2 Pixel combination selection circuit 4 Minimum value set generation circuit 6 Maximum value output circuit 7 Maximum value set generation circuit 8 Minimum value output circuit 10 Subtractor 11 Subtractor 12 Comparator

Claims (2)

[Claims] 1. A noise reduction processing device for a moving image consisting of a plurality of temporally consecutive frames, each frame of which is composed of a plurality of pixels, wherein in a frame including a pixel to be processed, a and the values of the pixels located below are U and D, respectively, and the values of pixels located at the same position as the pixel to be processed in frames temporally before and after the frame including the pixel to be processed are P and L, respectively. When the value of the pixel to be processed is A, means for selecting a predetermined set of combinations of at least one of the U, D, P, and L and the A, and for each of the selected combinations, 1. A moving image noise reduction processing apparatus, comprising: means for generating a minimum value; and means for outputting a maximum value of the minimum values as a value of the pixel to be processed. 2. A noise reduction processing device for a moving image consisting of a plurality of temporally consecutive frames, each frame of which is composed of a plurality of pixels, wherein in a frame including a pixel to be processed, a and the values of the pixels located below are U and D, respectively, and the values of pixels located at the same position as the pixel to be processed in frames temporally before and after the frame including the pixel to be processed are P and L, respectively. When the value of the pixel to be processed is A, means for selecting a predetermined set of combinations of at least one of the U, D, P, and L and the A, and for each of the selected combinations, 1. A moving image noise reduction processing apparatus, comprising: means for generating a maximum value; and means for outputting a minimum value of the maximum values as a value of the pixel to be processed.