JPH0211064A - Picture signal processor - Google Patents

Abstract

PURPOSE:To remove noise without the reduction of a contrast by using a media window matrix method as a noise removing means in a picture signal. CONSTITUTION:Light from an original picture, after it is converted to an electric signal with a photo-electric converter 10, is converted to a digital signal corresponding to the picture element concentration with an A/D converter 11. The digital signal is accumulated in a frame memory 12 as picture data, an mXn picture element is extracted out of the picture signals accumulated at the frame memory 12 by an arithmetic processor 13, one out of these respective ones is successively selected, and further, after the picture element matrix of an (i) row and a (j) column is set, a media window (MD) operation is executed, the obtained value is replaced to the original picture element value and the new picture element arrangement is executed. Thus, without the reduction of a contrast, the high frequency noise can be removed from the picture signal.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an image signal processing device, and more particularly to an image signal processing device that can remove high frequency noise in a signal without reducing image contrast. It is related to.

(Prior art) In digital image signal processing devices, such as electrophotographic devices, copying machines, and facsimile devices, when converting an original image into a digital signal and performing various signal processing, dust on the contact glass, unevenness in the background of the document, etc. Two types of filtering techniques are used to remove the noise generated and mixed in and reproduce a clear image.

Generally, in facsimile machines and other electrophotographic devices, an original image is divided into fine pixels in a matrix, and the signals are converted into digital electrical signals corresponding to density or color and transmitted or input into a memory device. In this case, the averaging calculation matrix method has traditionally been common as a noise removal method performed at various stages. This method locally averages when there are small fluctuations in the image signal. It eliminates unnecessary noise and functions as a kind of spatial low-pass filter.

For example, for each pixel arranged in a matrix and digitized, the average value of several nearby pixel values including surrounding pixels is calculated, and the original value is replaced with this average value.

To briefly explain this, let us consider a case in which averaging filtering is performed on one of the pixels l arranged as shown in FIG. 4, for example.

Now, for each pixel of interest 2, we consider a 3×3 matrix 3 with this as the center, the row of the pixel array is 1, the column is j, the pixel of interest is indicated as di, j, and this part is extracted. The diagram is shown in FIG. 5, and the averaging is performed using the following equation for this matrix value.

However, D i,j indicates the average value for the pixel di,j.

However, although this method acts as a kind of low-pass filter and is effective in removing high-frequency noise, the average value is affected by the magnitude of the surrounding pixel values and is not effective at edges where pixel values change greatly. There was a drawback of reduced contrast.

That is, in the relationship between the density and position of the original image shown by the solid line in FIG. 6, there is a problem in that due to averaging, steep changes in pixel density as shown by the dotted line become blurred, resulting in a decrease in contrast.

(Objective) The present invention was made in response to the above circumstances, and an object of the present invention is to provide an image signal processing device equipped with a method for removing high frequency noise from an image signal without reducing contrast. .

(Structure of the Invention) In order to achieve the above object, the present invention provides a memory device that converts the pixels of an image into digital signals, arranges the digital signals in a matrix, and sequentially stores the digital signals, and a memory device that sequentially stores the digital signals. An arithmetic unit that extracts a predetermined number of pixel matrices including a pixel and some of its surroundings and calculates a median value from these, and replaces the median value with the value of the pixel of interest to create a new pixel matrix. It is characterized by having the means and.

The present invention can be widely applied to devices that handle image signals such as electrophotographic devices using digital signals, and specific examples thereof include facsimile machines and copying machines.

For example, in a copying machine, in addition to the above-mentioned means, a photoelectric converter that reads the original image, an A/D converter that converts the read signal into a digital signal, and a median value created by the above method are used as a new pixel signal. It is characterized by comprising a photoreceptor that forms a latent image based on the output driver signal.

DESCRIPTION OF THE PREFERRED EMBODIMENTS The image signal processing device of the present invention will be specifically described below based on an embodiment shown in the accompanying drawings.

FIG. 1 shows an embodiment of the present invention, in which the present invention is applied to a latent image forming device of an electrophotographic apparatus such as a copying machine.

In this figure, reference numeral 10 is a photoelectric converter, and the light from the original image is converted into an electrical signal by the photoelectric converter 10, and then the A/D converter 11 (analog-to-digital converter)
is converted into a digital signal corresponding to the pixel density.

This digital signal is sent to the frame memory 1 as image data.
At the same time, the signals are stored in the frame memory 12 and subjected to noise removal calculation processing, which will be described in detail below, by the arithmetic processor 43, and then stored in the frame memory 12 again.

In this way, noise is removed and the frame memory 1
The image signal MW 2 is written as a latent image on the photosensitive drum 16 via the output driver 14 and the write driver 15.

Next, a noise removal device performed in the arithmetic processor 13 will be explained.

The signals stored in the frame memory I2 are considered in the same manner as shown in FIGS. 4 and 5, and a 3×3 matrix is extracted.

In the present invention, instead of averaging these pixel values, the median value of these pixel values is found, and this value is substituted for the pixel di, j and output.

That is, for the pixel of interest di, j, the following calculation is performed for a total of nine pixels, including this and eight surrounding pixels.

Di, j = Med [di-1, j-1, di-1,
j, di-1゜j+l, di, j-l, d
i, j, di, j+I, dDl, j-1, di
+1. j, di÷1. j+1 ] ・ ・
・ ・ ・ (2) However, Med here is an operation to find the median value of the elements in [].

In this way, the median value Di,j =Med[ ] is obtained for all pixels di,j, and this value dij is newly stored in the frame memory 12 as image data of rows i and j.

According to this method, high frequency noise can be removed without deteriorating the contrast compared to the conventional averaging method.

Here, since the signal processing for determining the median value of a predetermined number of pixel matrices is conventionally known as a median window operation matrix (MD matrix), detailed explanation will be omitted.

Incidentally, although the above explanation shows the case of a 3×3 matrix, the present invention does not need to be limited to this, and a 5×5 matrix or more may be used.

Expressing this using a general formula, D is a region (window) indicating the vicinity of each pixel of interest, and med indicates the median value of the function f (k, λ).

At this time, the neighborhood value may be constant, but may be changed depending on the thinness of the image to be preserved or the type of image.

FIGS. 2(a), 2(b), and 2(c) are characteristic diagrams showing experimental results for detailed explanation of the present invention, in which (a)
N x original image signals containing high frequency noise, (b
) shows an image signal obtained as a result of noise removal processing using the conventional average value method, and (c) shows an image signal obtained as a result of noise removal processing using the MD matrix processing of the present invention. .

As is clear from this figure, the conventional averaging method not only cannot completely remove noise but also blurs sharp changes in pixel density, whereas the present invention completely eliminates noise. High frequency noise can be removed and there is no decrease in contrast.

FIG. 3 is a diagram showing an example of a flowchart of the above-mentioned MD matrix calculation.

To simply illustrate this method, 'Frame memory 1 in Figure 1
Extract mXn pixels from the image signals accumulated in 2,
After selecting one for each of these sequentially and setting the pixel matrix of i rows and 5 columns described above for each of them, MD calculation is performed and the obtained value is replaced with the original pixel value to create a new pixel matrix. Get the pixel array.

As mentioned above, the present invention is not limited to what is shown in this flowchart.

Furthermore, the present invention is applicable not only to copying machines but also to a wide range of image processing apparatuses including facsimile machines and electrophotographic apparatuses such as optical disk apparatuses.

Furthermore, although the above calculations may be implemented using an electronic circuit, it is generally easier to perform program processing using a CPU.

(Effects of the Invention) As explained above, the present invention uses the media window matrix method as a means for removing noise in image signals, so it can be used without reducing contrast compared to conventional averaging methods. Noise can be completely removed and the effect is remarkable in obtaining clearer electrophotographic images.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of a copying apparatus to which the present invention is applied, and FIG. 2 is a characteristic diagram showing the relationship between image density and position of an image signal for explaining the present invention in detail. ,
(a) is the original image signal, (b) is the result of noise removal using the conventional averaging method, (C) is the image signal obtained as a result of noise removal according to the present invention, and Figure 3 is the original image signal. A diagram showing an example of a flowchart for carrying out the invention, and FIGS. 4 and 5 are diagrams showing matrix formation of image signals. FIG. 6 is a characteristic diagram showing the effect of noise removal by the conventional averaging method.

Claims (2)

[Claims] (1) In an image processing device that converts the pixels of an image into digital signals and then arranges them in a matrix and performs the required processing, there is provided a memory device that stores the digital image signal matrix, and a memory device that stores the digital image signal matrix, and a memory device that is to be processed from the memory device. an arithmetic unit that extracts a predetermined number of pixel matrices including a pixel and surrounding pixels, and calculates the median value of the pixel matrix; An image signal processing device comprising: means for creating a matrix. (2) The image processing device includes a photoelectric converter that reads an image from a document, and an A/C converter that converts the read signal into a digital signal.
Claim 1 characterized in that it is equipped with a D converter, an output driver that amplifies a signal outputting the median value as a new pixel, and a photoreceptor that forms a latent image using the output driver signal. ).