JPH04122931A - Image processing device - Google Patents

Abstract

PURPOSE:To change a threshold level and to remove the noises of a color signal by independently setting lookup tables for each of respective colors. CONSTITUTION:A circuit 18 for generating the average value of nearby picture elements computes a color signal SYY and determines an average value UY when the color signal SY (yellow) is inputted. The conversion values Q(UY) corresponding to the threshold levels of the respective colors are outputted by the loopup tables with the above-mentioned value as an address. On the other hand, the output signal of an adder circuit 26, i.e. the output signal FY of a noise removing circuit 12, is determined by an arithmetic circuit 24, a multiplier circuit 22, and the circuit 26. FY is expressed by FY=SY+Q (UY)X(UY-SY) where the conversion value Q(UY) denotes 0 to 1 values. The input signal SY is outputted as it is as the output of the circuit 12 when this value is 0. The average value UY is outputted when the above-mentioned value is 1. The above-mentioned processing is executed for each of the respective colors.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image processing device that can suitably remove noise from color images.

[Prior Art] In the fields of printing and plate making, when a continuous tone image is converted into an electrical signal using a solid-state image sensor such as a CCD and then reproduced on a film, etc., the influence of the graininess of the solid-state image sensor or film etc. Noise appears on the reproduced image. For example, in the technology disclosed in Japanese Patent Publication No. 63-36189 to reduce this noise, density information corresponding to each pixel is once stored in a storage device, and then a plurality of pixels stored in the storage device are Using the density information of , find a density distribution in which those pixels are arranged into density classes, define the density of the pixels located at both ends of the density distribution as the white level and black level, and calculate the density distribution from the white level position in the direction of the black level position. A predetermined ratio of pixels to the total number of pixels is counted, and the level of the pixel closest to the black level among the counted pixels is set as the threshold hold level.

Furthermore, the technology disclosed in Japanese Patent Application Laid-Open No. 63-156475 is based on the fact that human vision has a characteristic that is close to logarithmic with respect to brightness (that is, a characteristic that is sensitive to changes in brightness in dark places). Therefore, if the same amount of noise is mixed into the highlights and shadows of an image, there will be much more noise in the shadows than in the highlights, which will hinder accurate reproduction of tones. In addition, human vision can only have a low spatial resolution in a portion of the shadow, so even if the resolution is slightly lowered in a portion of the shadow, the quality of the image obtained does not appear to the human eye to have deteriorated. Focusing on the fact that it is invisible,
By creating and outputting an unsharp signal only in the shadow part of the image, for example by averaging adjacent pixels, it is possible to eliminate noise for the human eye without reducing resolution in one reading. The aim is to reduce the Furthermore, regarding noise removal in color image processing, the magnitude relationship between the density of a center pixel and the density average value of its neighboring pixels is determined by referring to a lookup table (LUT) and the density of the center pixel is output or the density average of neighboring pixels is output. Techniques for determining whether to output a value are known.

[Problem to be solved by the invention] However, in the case of color signals, the graininess of the film, the CCD
Since the magnitude of noise varies depending on the color due to the photoelectric conversion characteristics and the like, there is a problem that good results cannot be obtained by removing noise using a black and white signal.

The present invention has been made in order to solve the above problems, and provides color signals (Y (yellow), M (magenta)).
, C (cyan) or R (red), G (green), B (blue))
An object of the present invention is to provide an image processing device that can perform optimal noise removal by independently installing lookup tables with different conversion threshold levels for each conversion.

[Means for Solving the Problems] In order to solve the above-mentioned problems, the present invention includes a neighboring pixel average value generation circuit that generates an average value of image signals in pixels in the vicinity of a pixel of interest, and an image signal. A look-up table that is set for each color and holds coefficients corresponding to the average value; and calculation means for obtaining a desired image signal from the image signal from the pixel of interest, the average value, and the coefficient. It is characterized by

[Function] Since the magnitude of noise differs depending on the color, the image and image processing device according to the present invention change the threshold level for each color using a lookup table.
Removes noise from color images.

[Embodiments] Next, preferred embodiments of the image processing apparatus according to the present invention will be described in detail with reference to the accompanying drawings.

In FIG. 1, reference numeral 10 indicates a noise removal circuit incorporated in the image processing apparatus according to this embodiment.

The noise removal circuit 10 receives, as input signals, color signals Sy, Sm, and Sc from a photoelectric conversion element (not shown) using a CCD, for example. Here, S
Y, S, , and sc are color signals of yellow, magenta, and cyan, respectively, and each signal is input to a noise removal circuit 12, 14, and 16 for each color, respectively. F, ,F,,F
o is the output signal of each noise removal circuit 12, 14, 16, and indicates the output after noise removal.

FIG. 2 is a diagram showing the configuration of the noise removal circuit 12 related to the color signal S shown in FIG. 1. Note that the configurations of the other noise removal circuits 14 and 16 are the same as that of the noise removal circuit 12, and a description thereof will be omitted.

The color signal S, is converted into an average value Uy by a neighboring pixel average value generation circuit 18. The neighboring pixel average value generation circuit 18 is, for example, a circuit that adds human input signals of neighboring pixels of a pixel related to the color signal S and calculates an average.

A look-up table (LOT) 20 stores the above average value U.
, is used as an address to change the threshold level during image processing.

The subtraction circuit 24 receives the average value U and the color signal S.

Find (Uy　Sy) from .

The multiplication circuit 22 calculates the product of the converted value Q (UY ), which is the output of the look-up table 20, and (U, -3 units).

Addition circuit 26 adds the outputs of multiplication circuit 22 and subtraction circuit 24.

Next, the operation of the embodiment of the present invention configured as above will be explained.

When the color signal S is input, the neighboring pixel average value generating circuit 18 calculates the color signal SY to obtain the average value Uy. Next, using this average value U as an address, the lookup table 20 outputs a converted value Q (U, ) according to the threshold level of each color.

On the other hand, the average value is determined by the subtraction circuit 24 (U.

−8,) is required.

The obtained (UY SY) is converted by the multiplier circuit 22 to the converted value Q (u
, ) to find the product.

This product, i.e., Q (UY ) X (UYSy
) is further added to the color signal S by the adder circuit 26, and the output signal F of the adder circuit 26 is obtained.

t is found.

Fy = Sv + Q (TJy) × (I
jy SY ) Here, the converted value Q (U, ) takes a value from 0 to 1, and when this value is O, the output of the noise removal circuit is the input signal as it is, and when it is 1, it is the average value UY is output. The above processing is performed for each color.

In this way, the noise removal circuit 12 uses the lookup table 20 for each color to generate the color signal Sy,
When the noise of Sv and Sc is removed, the noise of the image is suitably removed and the roughness is reduced.

[Effects of the Invention] The image processing device according to the present invention focuses on the fact that the magnitude of noise differs depending on the color, and by setting a lookup table independently for each color, the threshold level is changed to improve the color signal. Remove noise. As a result, a good image without any roughness can be obtained.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a noise removal circuit in an image processing apparatus according to the present invention, and FIG. 2 is a block diagram of each color of the noise removal circuit. 10... Noise removal circuit 18... Neighboring pixel average value generation circuit 20... Lookup table 22... Multiplication circuit 24... Subtraction circuit 26... Addition circuit

Claims (1)

[Claims] (1) A neighboring pixel average value generation circuit that generates the average value of image signals in pixels in the vicinity of the pixel of interest, and a lookup table that is set for each color that makes up the image signal and holds coefficients corresponding to the average value. An image processing device comprising: arithmetic means for obtaining a desired image signal from the image signal from the pixel of interest, the average value, and the coefficient.