JP2003009165A - Image signal processing method and imaging apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image signal processing method that allows an output image to be seen easier than a conventional black/white wrist watch type imaging device, and can reduce power consumption. SOLUTION: LPF processing is made only horizontally in an LPF processing section 7A after interpolation processing, an error diffusion processing section 10A is used for gradation conversion processing, and error diffusion is made only horizontally. Based on the configuration, nearly the same picture quality as the case where processing is made both horizontally and vertically can be achieved with a small-scale circuit by reducing processing in a vertical direction in LPF processing and error diffusion processing.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image signal processing method for digital signal processing and outputting a video signal obtained by imaging an object with an image sensor.

[0002]

2. Description of the Related Art In recent years, a small image pickup device (wrist camera) that a user can carry around by wearing it on a wrist like a wristwatch has been developed.

The image signal processing method of the wristwatch type image pickup device is processed as shown in FIG. The captured image is converted by the black and white image sensor 1 into a still image of about 8 bits, and the output image of a still image of about 8 bits and a moving image after the signal processing by the signal processing unit 2 is a monochrome liquid crystal of about 8 bits. It is displayed on panel 3.

In this block diagram, the output of the monochrome image sensor 3 has already been digitally converted by an A / D converter (not shown). As shown in FIG. 10B, the signal processing unit 2 emphasizes the OB clamp / AGC unit 2a that fixes the black level and controls the gain, the gamma processing unit 2b that corrects the characteristics of the image data, and the contour portion. It is composed of an aperture processing section 2c.

[0005]

Such a wrist watch type image pickup device is a wrist watch type in which a battery of a relatively large size cannot be used. Therefore, in order to reduce power consumption, lower power consumption than color display is realized. The current situation is that it can be displayed in black and white.

However, since the output image is black and white, there is a problem that the output image is difficult to see depending on the subject. Therefore, an object of the present invention is to provide an image signal processing method in which an output image is easier to see than a conventional black-and-white wristwatch-type imaging device and power consumption can be reduced as much as possible.

[0007]

According to a first aspect of the present invention, there is provided an image signal processing method, wherein a video signal read from a color image sensor is converted into a digital signal, which is interpolated to obtain a color component for each pixel. When interpolating into output information,
It is characterized in that horizontal low-pass processing and vertical low-pass processing are performed on the interpolated data to reduce noise generated during the interpolation from the output information.

According to a second aspect of the present invention, there is provided an image signal processing method in which a video signal read from a color image sensor is converted into a digital signal, which is interpolated to obtain a color component for each pixel and output information. In this case, the horizontal low-pass processing and the vertical low-pass processing are performed on the interpolated data to reduce the noise generated during the interpolation, and the horizontal low-pass processing and the vertical low-pass processing are performed. It is characterized in that the processed signal is subjected to horizontal error diffusion processing and vertical error diffusion processing, and the gradation of the output information is converted into a gradation lower than that of the input information and outputted.

According to a third aspect of the present invention, there is provided an image signal processing method, wherein a video signal read from a color image sensor is converted into a digital signal, interpolation processing is performed to interpolate a color component for each pixel, and output information is output. In this case, the noise after the interpolation is reduced by performing the horizontal low-pass processing on the post-interpolation data without performing the vertical low-pass processing, and by performing the horizontal low-pass processing on the signal. The horizontal error diffusion process is performed without performing the vertical error diffusion process, and the grayscale of the output information is converted into a lower grayscale than the input information and is output.

According to a fourth aspect of the present invention, there is provided an image pickup apparatus in which an interpolation process is performed to interpolate a color image sensor and a digital signal of a video signal read from the color image sensor to interpolate a color component for each pixel. Section, an LPF processing section that applies an LPF to the interpolated data, and a gradation conversion processing section that performs gradation conversion processing of the interpolated data into data having a smaller number of bits. The unit is configured to perform the LPF processing only in the horizontal direction, and the gradation conversion processing unit is configured to perform the error diffusion processing only in the horizontal direction.

[0011]

BEST MODE FOR CARRYING OUT THE INVENTION The image signal processing method of the present invention will be described below based on each specific embodiment.

(Embodiment 1) FIGS. 1 and 2 show a wristwatch type image pickup apparatus which realizes an image signal processing method of the present invention. In the image pickup apparatus employing the color image sensor 4 and the color liquid crystal panel 5 (Embodiment 1), the color image sensor 4 outputs about 8 bits of each color of RGB, and the input of the color liquid crystal panel 5 is about 8 bits. It is three times 24 bits.

As shown in FIG. 2A, the color image sensor 4 has a C-MOS photoelectric conversion unit 4a having 176 pixels in the horizontal direction and 144 pixels in the vertical direction, and each pixel of the photoelectric conversion unit 4a. There is a one-to-one correspondence between the incident light P and the photoelectric conversion unit 4a.
The color filter 4b is arranged on the side of h and has a specific repeating pattern of filter elements of RGB colors arranged in a grid pattern. The output of the color image sensor 4 has already been digitally converted by an A / D converter (not shown).

The interpolation processing unit 6 interpolates the 8-bit output of each pixel of the photoelectric conversion unit 4a to convert the output of the color image sensor 4 into a total of 24 bits of 8 bits for each color of RGB. Specifically, the case of a target pixel that detects light that has passed through the R color filter element will be described as an example. With respect to the 8-bit output of the target pixel, the 8-bit output of the target pixel is Two outputs of G color 8 bits and B color 8 bits are interpolated and produced from the outputs of the pixels around this pixel of interest, and converted into a total of 24 bits of 8 bits for each color of RGB.

In the horizontal / vertical LPF processing section 7, based on the 1H delayed interpolation signal D1 which can be read from the 1HRAM 8 which repeatedly reads the interpolated signal and the latest 1H signal D2 after the interpolating, etc. Performs horizontal low-pass processing and vertical low-pass processing of signals.

Here, "1H" is the scanning time in the horizontal direction,
This is the time required to sequentially scan and read 176 pixels in a horizontal row of the color image sensor 4. The horizontal low-pass processing will be specifically described.

[0017] outputs a moving average of the output values of the pixels adjacent in the horizontal direction as a result, T ₁₁ the output value of the latest target pixel shown in FIG. 2 (b), before one horizontally adjacent When the output value of the pixel is T ₁₀ , (T ₁₀ + T ₁₁ ) / 2 is output, and when the output value of the latest pixel at the next timing is T ₁₂ , (T ₁₁ + T ₁₂ ) / 2 is output. The averaged value is output to.

The vertical low-pass processing will be described in detail. A moving average of the output values of vertically adjacent pixels is used as the output. The latest output value of the target pixel is T ₁₁ , and the vertically adjacent pixels are vertically adjacent. If the output value of the pixel one hour before is T ₀₁ ,
When (T ₀₁ + T ₁₁ ) / 2 is output and the output value of the latest pixel at the next timing is T ₁₂ , and the output value of the pixel 1H before is T ₀₂ , (T ₀₂ + T ₁₂ ) / 2 is output. The averaged value is output.

The 24-bit signal output from the LPF processor 7 is displayed on the color liquid crystal panel 5 having a gradation of 24 bits via the signal processor 9. The signal processing unit 9 is configured to perform color signal processing. Specifically, as shown in FIG. 3, the white balance processing unit 9a, the gamma processing unit 9b, and the RGB components are converted into YUV components. YU
It is composed of a V conversion unit 9c, an aperture processing unit 9d that performs edge enhancement correction on the Y component, a matrix processing unit 9e that adjusts the hue with the UV component as an input, and an RGB conversion unit 9f that converts the YUV component into the RGB component. Has been done.

Because of this configuration, the interpolation processing unit 6 is provided for color display. However, the LPF processing unit 7 performs the horizontal low-pass processing and the vertical low-pass processing to obtain the interpolation processing unit. Noise due to interpolation by 6 Noise can be removed. A color image that is easy to see can be provided.

(Embodiment 2) FIGS. 4 to 6 show a wrist watch type image pickup apparatus (embodiment 2) which realizes the image signal processing method of the present invention.

In the image pickup apparatus of the first embodiment, the 24-bit output signal of the signal processing unit 9 is displayed on the 24-bit color liquid crystal panel 5 as it is. The difference from the image pickup apparatus of the second embodiment is that a horizontal / vertical error diffusion processing unit 10 is interposed between the output of the processing unit 9 and the color liquid crystal panel 5A.

In the first embodiment, color display makes the display easier to see than the conventional example of monochrome display. However, the amount of information is 8 bits for each color component RGB, that is, 24 bits, which is three times the number of gradations. It is necessary to use a color liquid crystal panel that consumes a lot of power, that is, consumes a large amount of power, and it is desirable to reduce power consumption.

On the other hand, in this (Embodiment 2),
A horizontal / vertical error diffusion processing unit 10 is provided to perform gradation conversion to a small number of bits and display the low gradation color liquid crystal panel.

Here, the 24-bit output image of the signal processing unit 9 is converted into 12-bit gradation, and a 12-bit color liquid crystal panel 5A having a gradation number smaller than 24 bits is used as the color liquid crystal panel. There is.

As shown in FIG. 5, the horizontal / vertical error diffusion processing unit 10 is delayed by the 1H delay signal D11 which can be read from the 1HRAM 11 which is repeatedly reading the signal processed by the signal processing unit 9 and the 1HRAM 11. 2H that can be read from 1HRAM 12 that is repeatedly reading signals
The delayed signal D12 and the latest 1H processed by the signal processing unit 9
The minute signal D10 is converted into 12 bits by the quantization processing unit 10a, and further the quantization error diffusion processing is performed by the quantization error diffusion processing unit 10b in the horizontal and vertical directions to perform gradation conversion.

The horizontal quantization error diffusion process and the vertical quantization error diffusion process will be specifically described with reference to FIG. In FIG. 6, the horizontal scanning direction is the x-axis direction, the vertical scanning direction is the y-axis direction, and the image data of the target pixel before the quantization error diffusion processing is A (th, tv), the target pixel after the quantization error diffusion processing. Output image data of: Ao (th, tv).

The relation between them can be expressed by a general formula as follows. Let Q (A (x, y)) be the function that quantizes the pixel value A (x, y) and QE (A (x, y)) be the function that finds the quantization error. , tv) error diffusion output data Ao (th, tv)
Is represented by the following formula.

Ao (th, tv) = Q [A (th, tv) + 1/4 × QE (A (th-2,
tv)) ＋ 3/4 × QE (A (th-1, tv)) + 1/4 × QE (A (th, tv-2)) + 3
/ 4 × QE (A (th, tv-1))} In this error diffusion output equation, the pixel of interest is A (th, t
At the time of moving to v), the coefficient of neighboring pixels x quantization error has already been added, and the added value is quantized, and that value is used as the output value Ao (th, tv). There is.

This completes the diffusion process for the pixel of interest. By performing this process for all pixel values while shifting the pixel of interest, the gradation conversion process is performed.
As described above, in the conventional error diffusion method, the quantization error is diffused to a large number of horizontal and vertical pixels in order to make the quantization error as inconspicuous as possible and to eliminate the problem due to the diffusion pattern.

(Embodiment 3) FIG. 7 shows a wrist watch type image pickup apparatus which realizes an image signal processing method of the present invention (Embodiment 3).

In the image pickup apparatus according to the second embodiment, the horizontal / vertical LPF processing unit 7 performs the horizontal LPF processing and the vertical LPF processing before the signal processing unit 9, and the signal processing unit 9
The horizontal / vertical error diffusion processing unit 10 executes the horizontal error diffusion processing and the vertical error diffusion processing in the latter stage, but in this (Embodiment 3), the horizontal LP is arranged in the preceding stage of the signal processing unit 9.
The F processing unit 7A is provided to execute only the horizontal LPF processing without executing the vertical LPF processing here, and after the signal processing unit 9, the data after the interpolation processing is gradation-converted into data having a smaller number of bits. The difference is that a horizontal error diffusion processing unit 10A as a gradation conversion processing unit for processing is provided and only the horizontal error diffusion processing is executed here without executing the vertical error diffusion processing.

In consideration of the error diffusion process for diffusing only in the horizontal direction, if the original signal is the same signal in the vertical direction, the diffusion pattern does not change in the vertical direction and becomes vertical streak noise, resulting in image quality. However, if the LPF process in the vertical direction is not performed, noise remains in the vertical direction of the original signal (FIG. 9), and the noise causes a diffusion pattern. It will always change in the vertical direction.

As a result, the state of FIG. 8 does not occur and the image quality is not impaired. That is,
The horizontal streak noise that occurs when the LPF process in the vertical direction is not performed and the vertical streak noise that occurs when the error diffusion process in the vertical direction is not performed are eliminated, thereby eliminating the problems caused by the respective processes. It will be.

Therefore, in the wristwatch type image pickup device of this (Embodiment 3), which is configured so that neither the LPF processing nor the processing in the vertical direction in the error diffusion processing is executed, the horizontal / vertical processing (Embodiment 2) is performed. It is possible to obtain almost the same image quality as the image of the wristwatch type image pickup device.

Since neither the LPF processing nor the error diffusion processing in the vertical direction is executed in this way,
1 HRAM 8, 1 which was necessary in the second embodiment
1 and 12 are not required, the configuration and processing time can be shortened, and power consumption can be reduced. In particular, in a wristwatch-type image pickup device that requires moving image processing, it is possible to obtain a low gradation image having substantially the same image quality as an image that has also been processed in the vertical direction, and the circuit scale can be greatly reduced. As a result, a good low gradation image can be displayed on a color liquid crystal panel having a low display gradation.

[0037]

As described above, according to the image signal processing method of the present invention, the output image is easier to see than the conventional black and white wristwatch type image pickup device, and the power consumption can be reduced as much as possible.

In particular, when the processing in the vertical direction of the LPF and the error diffusion is reduced, it is possible to obtain a low gradation image having substantially the same image quality as the image processed in the vertical direction. The scale can be greatly reduced, and a good low gradation image can be displayed even on a color liquid crystal panel having a low display gradation.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a color wrist camera (embodiment 1) for realizing an image signal processing method of the present invention.

FIG. 2 is a color image sensor and LPF of the same embodiment.
Illustration of processing

FIG. 3 is a configuration diagram of a signal processing unit according to the same embodiment.

FIG. 4 is a configuration diagram of a color wrist camera (embodiment 2) for realizing the image signal processing method of the present invention.

FIG. 5 is a configuration diagram of a horizontal / vertical error diffusion processing unit according to the same embodiment.

FIG. 6 is an explanatory diagram of a horizontal / vertical error diffusion processing unit according to the same embodiment.

FIG. 7 is a configuration diagram of a color wrist camera which realizes an image signal processing method of the present invention (Embodiment 8).

FIG. 8 is an explanatory diagram of horizontal error diffusion processing according to the same embodiment.

FIG. 9 is an explanatory diagram of horizontal LPF processing according to the same embodiment.

FIG. 10 is a block diagram of a conventional black and white wrist camera.

[Explanation of symbols]

4 color image sensor 5 color LCD panel 4a C-MOS photoelectric conversion unit Ph incident light 4b color filter 6 Interpolation processing unit 7 Horizontal / vertical LPF processing unit 8, 11, 12 1 HRAM D1 1H delay interpolation signal The latest 1H signal after D2 interpolation 9 Signal processing unit 9a White balance processing unit 9b Gamma processing unit 9c YUV converter 9d aperture processing section 9e Matrix processing unit 9f RGB converter 10 Horizontal / vertical error diffusion processing unit 5A 12-bit color LCD panel D11 1H delay signal D12 2H delay signal D10 Latest signal for 1H 10a Quantization processing unit 10b Quantization error diffusion processing unit 7A Horizontal LPF processing unit 10A Horizontal error diffusion processing unit (gradation conversion processing unit)

   ─────────────────────────────────────────────────── ─── Continued front page    F-term (reference) 5C021 PA34 PA58 PA78 PA82 PA83                       RA11 RB05 SA21 XA34 XB07                       XB17 YA01 YC09                 5C065 AA01 BB12 BB22 CC02 CC03                       CC09 DD15 GG06 GG07 GG13                       GG18 GG29 GG30                 5C066 AA01 AA03 AA18 CA07 DD07                       EA04 EA11 EC01 EC12 EE01                       GA05 HA04 KA12 KE19 KF05                       KM02

Claims (4)

[Claims] 1. When a video signal read from a color image sensor is digitally converted, and interpolation processing is performed to interpolate a color component for each pixel to obtain output information, the horizontal direction with respect to the data after interpolation is used. An image signal processing method for performing low-pass processing and vertical low-pass processing to reduce noise generated during interpolation from output information. 2. A video signal read from a color image sensor is digitally converted, and interpolation processing is performed to interpolate a color component for each pixel to obtain output information. Low-pass processing and vertical low-pass processing are performed to reduce noise generated during the interpolation, and horizontal error diffusion processing and vertical direction of the horizontal low-pass processing and vertical low-pass processed signals are performed. An image signal processing method in which error diffusion processing is performed and the gradation of output information is converted into a gradation lower than that of input information and then output. 3. A video signal read from a color image sensor is digitally converted, and interpolation processing is performed on the video signal to interpolate a color component for each pixel to obtain output information. Horizontal low-pass processing is performed without performing low-pass processing to reduce noise generated during the interpolation, and horizontal horizontal low-pass processing is performed on the signal after horizontal low-pass processing without performing vertical error diffusion processing. An image signal processing method for performing directional error diffusion processing to convert the gradation of output information into a gradation lower than that of input information and output the converted gradation. 4. A color image sensor, an interpolation processing unit that interpolates a digital signal of a video signal read from the color image sensor to interpolate a color component for each pixel, and to the data after the interpolation. An LPF processing unit that applies an LPF and a gradation conversion processing unit that performs gradation conversion processing on the data after the interpolation processing into data having a smaller number of bits are provided, and the LPF processing unit performs the LPF processing only in the horizontal direction. The image pickup device, wherein the gradation conversion processing unit is configured to perform error diffusion processing only in the horizontal direction.