JP2003333612A - Hue noise filter, solid-state image pickup device, and photographing device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce a color noise included in a color difference signal of image information. <P>SOLUTION: In a hue noise filter which obtains a color difference signal in a processing object pixel position from color difference signals in individual pixel positions within a prescribed range including the processing object pixel position, change of each color difference signal I within the prescribed range is obtained, and the value of the color difference signal in the processing object pixel position is held in a preset prescribed level and is outputted in the case that the change crosses the prescribed level. Thus the color noise is reduced. Peculiarly, when the prescribed level is '0', color is not given to an achromatic color to make the color noise visually inconspicuous. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a hue noise filter for reducing hue noise which is a cause of color noise in an image, and more particularly to a hue noise filter in which a change in hue signal is reduced to make the color noise visually inconspicuous. The present invention relates to a noise filter and a solid-state imaging device and a photographing device equipped with this hue noise filter.

[0002]

2. Description of the Related Art An image pickup device such as a digital still camera or a digital video camera is equipped with a solid-state image pickup device such as a CCD or a CMOS sensor. This solid-state image sensor,
The subject image is output as a signal for each color of red (R), green (G), and blue (B) on a pixel-by-pixel basis, and the image processing device outputs the RGB signal to a luminance signal (Y) and a color difference signal (Cr, C).
The image is converted into b) and the captured image is reproduced.

At this time, if a large amount of noise is included in the color difference signal, the color noise becomes noticeable, resulting in an unnatural image. Therefore, a low-pass filter has been conventionally used to reduce the color noise.

FIG. 8A is a diagram showing an example of the color difference signal Cr. The horizontal axis shows the positions of the pixels p1, p2, ... And the vertical axis shows the magnitude of the color difference signal Cr. If the magnitude of the color difference signal Cr obtained at each pixel position is as shown in FIG. 8A, the conventional color noise filter (low-pass filter) performs a process of taking an average value of neighboring pixel data. We are trying to reduce color noise.

That is, as the magnitude of the color difference signal Cr of the pixel p2, the average value of the color difference signals a, b and c of the pixels p1, p2 and p3 is obtained to determine the magnitude of the color difference signal of the pixel p2 (FIG. 8 (b)).
Black point B) (B = (a + b + c) / 3), and as the magnitude of the color difference signal Cr of the next pixel p3, the pixels p2, p3, and
The average value of the color difference signals b, c, d of p4 is obtained and set as the magnitude C (= (b + c + d) / 3) of the color difference signal of the pixel p3. The low-pass filter sequentially performs such average value calculation and outputs color difference data B, C, ... Shown in FIG.

[0006]

By applying the above-described low-pass filter processing to the color difference signals, the color difference signals a, b, c, ... Shown in FIG. 8 (a) are replaced by the color difference signals B shown in FIG. 8 (b). , C, D, ... Reduces the color noise by adding the surrounding pixel data. However, the change in the color difference signal is still large, and for example, if there is a large change in the color difference signal as shown in FIG. 8B in an achromatic image (for example, “white”) (the value of the color difference signal is “0”), Has a problem in that color is conspicuous in places that should be "white", and color noise is noticeable.

An object of the present invention is to provide a hue noise filter which reduces color noise and makes the color noise visually inconspicuous, and a solid-state image pickup device and a photographing device equipped with this hue noise filter.

[0008]

A hue noise filter for achieving the above object is a hue noise filter for obtaining a color difference signal at a pixel position to be processed from a color difference signal at each pixel position within a predetermined range including the pixel position to be processed. , Obtaining the change of each color difference signal within the predetermined range,
When the change crosses a predetermined level set in advance, hue noise reduction processing for holding the value of the color difference signal at the pixel position to be processed at the predetermined level is performed.

With this configuration, it is possible to reduce the hue noise from the color difference signal on which the low frequency hue noise is superimposed. Further, since the cycle of the points that cross the predetermined level is shorter than the cycle of the waveform of the color noise, it is possible to detect noise of a lower frequency and reduce low frequency noise.

Preferably, the predetermined level is zero indicating an achromatic color. With this configuration, it is possible to prevent color from adhering to an achromatic image.

A hue noise filter that achieves the above object is a hue noise filter that obtains a color difference signal at a pixel position to be processed from a color difference signal at each pixel position within a predetermined range including the pixel position to be processed, within the predetermined range. The hue noise reduction process is performed by comparing the values of the color difference signals and using the minimum absolute value of the values of the color difference signals as the value of the color difference signal at the pixel position to be processed. With this configuration, the change in the hue signal can be reduced, and the color noise in the image becomes inconspicuous.

A solid-state image pickup device which achieves the above-mentioned object is a solid-state image pickup device which senses a subject image, a conversion means which converts a color signal output from the solid-state image pickup device into a luminance signal and a color difference signal, and the color difference signal. On the other hand, a hue noise filter that performs the above-described hue noise reduction processing is mounted. With this configuration, it is possible to reproduce the captured image using the color difference signal with less hue noise.

An image pickup apparatus that achieves the above-mentioned object is a solid-state image pickup device that senses a subject image, an optical system that forms a subject image on the solid-state image pickup device, and a color signal output from the solid-state image pickup device as a luminance signal. And a conversion unit for converting the color difference signal into a color difference signal, a hue noise filter that performs the hue noise reduction process according to claim 1 on the color difference signal, a color difference signal with reduced hue noise, and luminance. Storage means for storing the subject image reproduced from the signal. With this configuration, it is possible to capture an image with little color noise.

[0014]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of a digital still camera equipped with a hue noise filter according to an embodiment of the present invention. In this example, the hue noise filter of the present embodiment is mounted on the digital still camera, but the hue noise filter of the present embodiment can be mounted on other types of photographing devices such as a digital video camera as well.

Further, the hue noise filter of the present embodiment does not necessarily have to be mounted on the photographing device, but may be mounted on an image processing system constituted by an image processing device, a personal computer or the like. Although the hue noise filter is described as a hardware circuit in the following, it is needless to say that this may be realized by software.

The digital still camera shown in FIG. 1 includes a taking lens 10, a solid-state image pickup device 11 such as a CCD, an aperture 12 provided between the two, an infrared cut filter 13, and an optical low-pass filter 14. Prepare The CPU 15, which controls the entire digital still camera, controls the lens driving unit 18 based on the distance information to the subject obtained by controlling the infrared light emitting unit 16 and the infrared light receiving unit 17, and sets the position of the taking lens 10 to the focus position. Then, the aperture amount of the aperture 12 is controlled via the aperture drive unit 19 so that the exposure amount becomes an appropriate exposure amount.

Further, the CPU 15 has an image pickup device driving section 20.
The solid-state imaging device 11 is driven via the, and the subject image captured through the taking lens 10 is output as a color signal.
Further, a user instruction signal is input to the CPU 15 through the operation unit 21, and the CPU 15 performs various controls according to the instruction. The solid-state imaging device 11 is a CCD having a honeycomb pixel arrangement, a Bayer CCD, or a CMOS sensor.

The electric control system of the digital still camera has an analog signal processing unit 22 connected to the output of the solid-state image pickup device 11 and an R output from the analog signal processing unit 22.
An A / D conversion circuit 23 for converting a GB color signal into a digital signal is provided, and these are controlled by the CPU 15.

Further, the electric control system of this digital still camera has a memory control unit 2 connected to the main memory 24.
5, a digital signal processing unit 26 described later in detail, a compression / expansion processing unit 27 that compresses a captured image into a JPEG image or expands a compressed image, and integrates photometric data to adjust the gain of white balance. External memory control unit 30 to which the integrating unit 28 and the removable recording medium 29 are connected
And a display control section 32 to which a liquid crystal display section 31 mounted on the back surface of the camera or the like is connected.
And the data bus 34 and the CPU 1
It is controlled by the command from 5.

The digital signal processing unit 26, the analog signal processing unit 22, the A / D conversion circuit 23, etc. shown in FIG. 1 can be mounted on a digital still camera as separate circuits, but these are solid-state imaging. It is preferable to manufacture the element 11 on the same semiconductor substrate by using an LSI manufacturing technique to form one solid-state imaging device.

FIG. 2 shows the digital signal processing unit 2 shown in FIG.
It is a detailed block diagram of FIG. This digital signal processing unit 26
Is a digital RG output from the A / D conversion circuit 23.
An offset correction circuit 41 that takes in the B color signal and performs offset processing, a gain correction circuit 42 that performs white balance, and a gamma correction circuit 43 that performs gamma correction on the corrected color signal are provided. When performing linear matrix processing or knee correction on the signal after the offset correction, it is performed between the gain correction circuit 42 and the gamma correction circuit 43.

The digital signal processor 26 further interpolates the RGB color signals after gamma correction to obtain RGB three color signals at each pixel position.
An RGB / YC conversion circuit 45 for obtaining the luminance signal Y and the color difference signals Cr, Cb from the RGB signal, a noise filter 46 for reducing noise from the luminance signal Y and the color difference signals Cr, Cb, and a luminance signal after noise reduction. A contour correction circuit 47 for performing contour correction on Y, and a color difference signal C after noise reduction
A color difference matrix circuit 48 for performing color tone correction by multiplying r and Cb by a color difference matrix.

The RGB interpolation calculation section 44 is not necessary if it is a three-plate type image pickup element, but the solid-state image pickup element 11 used in this embodiment is a single-plate type solid-state image pickup element, and each pixel is connected to R, Since only one color signal of G and B is output, in a pixel that outputs a color that is not output, that is, R, how much the color signals of G and B become at this pixel position is determined by G of surrounding pixels. , B signal by interpolation calculation.

The noise filter 46 is provided with the hue noise filter 46a according to this embodiment. This color noise filter 46a receives the color difference signals Cr and Cb, processes the color difference signals Cr and Cb as described later in detail, and outputs the color difference reduced color difference signals Cr and Cb to the color difference matrix circuit 48 in the subsequent stage. Output. Hereinafter, the color difference signal Cr
Although only the processing for the color difference signal Cb will be described, color noise is also reduced from the color difference signal Cb by performing the same processing on the color difference signal Cb.

FIG. 3 is a detailed block diagram of the hue noise filter 46a. The hue noise filter 46a of the present embodiment
Is a three-stage flip-flop 51 connected in cascade.
(Hereinafter referred to as FF1), 52 (hereinafter referred to as FF2), 53 (hereinafter referred to as FF3), and each FF1,
The output determination circuit 54 reads the values held by FF2 and FF3 and determines the output value. The output determination circuit 54 outputs "0" when all the data of FF1, FF2, FF3 are not "positive" or all "negative",
At other times, the data of FF2 is output.

FIG. 4 shows input data of the same color difference signal Cr as that shown in FIG. The magnitude of the color difference signal of the pixel p1 is data a, the magnitude of the color difference signal of the pixel p2 is data b, ..., The magnitude of the color difference signal of the pixel p11 is data k. FF1, FF2, FF3 shown in FIG.
The data of each pixel p1, p2, ... Is stored and shifted in order.

That is, when the data c, b, and a of the pixels p3, p2, and p1 are stored in the FF1, FF2, and FF3, respectively, and the next data d of the pixel p4 is stored in the FF1, the storage of the FF3 is performed. The data a is discarded, the stored data b of FF2 is shifted to FF3, and the stored data c of FF1 is shifted to FF2.

The hue noise filter 46a having such a configuration
When the color difference signal Cr after the RGB / YC conversion is input, the hue noise filter 46a of the present embodiment determines and outputs the magnitude of the color difference signal in each pixel using the data of three consecutive pixels. When the color difference signal data a, b, ..., K shown in FIG. 4 are sequentially input to the hue noise filter 46a, the data indicated by black dots in FIG. 4 are output as the color difference signal Cr.

That is, for example, the color difference signal of the pixel p2 is determined from the data a, b, and c of the pixels p1, p2, and p3, but in the case of this embodiment, the average value is not output and is output. In the case a <0, b <0, c> 0,
The output determination circuit 54 outputs "0". Similarly, "0" is output as the data of the next pixel p3. Next pixel p
In case of 4, c> 0, d> 0, e> 0, and therefore FF
The stored data of 2, that is, the data d in this case is output.

FIG. 5 is a diagram showing an example of the input signal waveform and the output signal waveform when the magnitude of the color difference signal Cr is determined and output as described above. It is assumed that a color difference signal I (input signal: shown by envelope of digital data) formed of digital data after RGB / YC conversion is input to the hue noise filter 46a.

The hue noise filter 46a has been described as having a predetermined processing range (a range of 3 pixels has been described in FIGS. 3 and 4;
The number of pixels to be processed is arbitrary, and the larger the value, the wider the range of data is referenced, so noise can be reduced.
There is a trade-off relationship that the image resolution decreases. Unless a zero-cross point is included in (), the input color difference signal is output as it is. However, if even one zero-cross point X is included in the processing range, "0" is output as the color difference signal. In FIG. 5, the output signal waveform II is shown by a thick line.

For example, if the spatial waveform of the noise contained in the color difference signal is a sine wave, the zero-crossing cycle is 1/2 of the cycle of this sine wave. That is, the zero-cross points X can be detected at intervals shorter than the cycle of the waveform, and it becomes possible to detect noise of a lower frequency.

For this reason, in the present embodiment, it is possible to reduce noise at a lower frequency as compared with the related art by the presence / absence of the zero cross of the color difference signals, and in particular, color noise in the vicinity of achromatic color is achromatic. You can Further, since low-frequency noise can be detected, color noise can be effectively reduced without widening the processing range.

In the above-described embodiment, the value of the color difference signal is set to zero (that is, achromatic color) when the zero-cross point falls within the processing range, but "zero-cross" is required to reduce the hue noise. Instead, the level of the color difference signal for which noise is desired to be reduced is set as a predetermined level, and it is determined whether the input signal waveform (color difference signal I) crosses this predetermined level within the processing range. The value of may be held at this "predetermined level" and used as the output value.

FIG. 6 is a block diagram of the hue noise filter 46a according to the second embodiment of the present invention. The hue noise filter 46a is different only in that an output determination circuit 55 is provided instead of the output determination circuit 54 shown in FIG. The output determination circuit 55 of the present embodiment includes FF1, FF2, FF
Of the three stored data, the data with the smallest absolute value is output.

FIG. 7 is an explanatory diagram showing the relationship between the input signal (same as in FIG. 8A) and output data of the hue noise filter 46a according to the second embodiment. The hue noise filter 46a outputs the data c of the color difference signal of the pixel p3 as the data of the color difference signal of the pixel p2, and outputs the data c having the smallest absolute value among the data a, b, and c of the pixels p1, p2, and p3. Among the data b, c, d of the pixels p2, p3, p4, the data c having the smallest absolute value is output as data. The output data of each pixel is shown by a black dot.

Also in the embodiment of FIG. 7, it is possible to suppress the change of the color difference signal and reduce the color noise as compared with the conventional example in which the average value shown in FIG. 8B is taken. In particular, since the color is made achromatic by holding a value close to “0” from the hue signal in the vicinity of the achromatic color, it is possible to reduce the magnitude of the change in the hue signal to make the color noise visually inconspicuous. it can.

In the above-described embodiment, for convenience of description, the output data is obtained from the color difference signals of three consecutive pixels arranged in a line, but the actual pixels of the solid-state image pickup device are arranged vertically and horizontally. Since the pixels are arranged two-dimensionally, in order to reduce the noise of a certain pixel, the same hue noise reduction process as described above is performed using the signals of the pixels within the two-dimensional predetermined range around the pixel. . For example, 3 around a pixel
Noise reduction is performed using the color difference signals of the pixels of × 3.

[0040]

According to the present invention, it is possible to reduce the hue noise in the color difference signal obtained by converting the color signal obtained from the image pickup device. In particular, since the color noise in the vicinity of the achromatic color is made achromatic, the color noise becomes visually inconspicuous, and an image having an excellent appearance can be obtained.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a digital still camera equipped with a hue noise filter according to an embodiment of the present invention.

FIG. 2 is a detailed configuration diagram of a digital signal processing unit shown in FIG.

FIG. 3 is a configuration diagram of a hue noise filter according to the present embodiment shown in FIG.

FIG. 4 is a diagram for explaining signal processing of the hue noise filter shown in FIG.

5 is a diagram showing an input signal waveform and an output signal waveform of the hue noise filter shown in FIG.

FIG. 6 is a configuration diagram of a hue noise filter according to another embodiment of the present invention.

FIG. 7 is a diagram illustrating signal processing of the hue noise filter shown in FIG.

FIG. 8 is a diagram illustrating an average value process performed by a conventional hue noise filter (low-pass filter).

[Explanation of symbols]

11 Solid-state image sensor 15 CPU 22 Analog signal processor 23 A / D conversion circuit 23 26 Digital signal processor 46 noise filter 46a Hue noise filter 51,52,53 flip-flops 54, 55 output judgment circuit I Input signal waveform (color difference signal Cr) II Output signal waveform (color difference signal Cr)

Continued front page    F-term (reference) 5C055 AA00 BA05 CA03 EA04 FA21                       GA01                 5C065 AA01 AA03 BB22 CC03 CC09                       DD02 GG02 GG11 GG18 GG22                 5C066 AA01 BA20 CA07 EC12 GA02                       GA05 GB03 HA01 KC01 KD02                       KE07 KE19 KM02

Claims (5)

[Claims] 1. A hue noise filter for obtaining a color difference signal at a processing target pixel position from a color difference signal at each pixel position within a predetermined range including the processing target pixel position, and obtaining a change of each color difference signal within the predetermined range, A hue noise filter that performs a hue noise reduction process of holding the value of the color difference signal at the pixel position to be processed at the predetermined level when the change crosses a preset predetermined level. 2. The hue noise filter according to claim 1, wherein the predetermined level is zero indicating an achromatic color. 3. A hue noise filter that obtains a color difference signal at a processing target pixel position from a color difference signal at each pixel position within a predetermined range including the processing target pixel position, and compares the values of the respective color difference signals within the predetermined range. A hue noise filter that performs a hue noise reduction process using the minimum absolute value of the color difference signals as the value of the color difference signal at the processing target pixel position. 4. A solid-state image sensor for exposing a subject image,
A conversion unit that converts a color signal output from the solid-state imaging device into a luminance signal and a color difference signal, and a hue noise that applies the hue noise reduction process according to any one of claims 1 to 3 to the color difference signal. A solid-state imaging device comprising a filter. 5. A solid-state image sensor for exposing a subject image,
An optical system for forming a subject image on the solid-state image sensor, a conversion unit for converting a color signal output from the solid-state image sensor into a luminance signal and a color difference signal, and the color difference signal. 3. A hue noise filter that performs the hue noise reduction process according to any one of 3 above, and a storage unit that stores the subject image reproduced from the color difference signal and the luminance signal with reduced hue noise. Imaging device.