JPH02214258A - Picture binarization control system - Google Patents

Abstract

PURPOSE:To attain binarization while eliminating noise by applying median filter processing, contour emphasis processing and dither processing to a read multi-level picture data. CONSTITUTION:A read only memory 12 and a random access memory 13 are used to allow a processor 11 to realize functions of read of multi-level picture data, median filter processing, contour emphasis and dither processing. That is, the multi-level picture data is read and the pre-processing such as median filter processing and contour emphasis is applied, and in the median filter processing, plural picture elements in a window around a noted picture element are arranged in the order of levels and the center level is matched with the level of the notes picture element. Then since a black isolating point is at a white level when other picture elements in the window are close to a white level, noise is eliminated. Moreover, contour is emphasized and the level is dither-processed with the systematic dither method to obtain a binary picture signal. Thus, the reproduced picture without distorted contour and any noise is obtained.

Description

[Detailed Description of the Invention] [Summary] An image binarization control method that binarizes multivalued image data such as a natural image using a systematic dithering method.
In an image binarization control method in which multivalued image data is binarized by systematic dithering for the purpose of digitization, there is provided a means for reading the multivalued image data, and a multivalued image data read by the means. Means for performing median filter processing on value image data; means for performing edge enhancement processing on multivalued image data filtered by said means; and means for applying dithering processing on multivalued image data whose contours have been enhanced by said means. and is configured to output multivalued image data as binary image data by the means.

[Industrial application field]

The present invention relates to an image binarization control method that bisulfates multivalued image data such as a natural image using a systematic dithering method.

When transmitting multivalued image data such as a natural image using a binary transmission system such as a facsimile, or when outputting it using a binary thermal printer, etc., it is common to convert it into a binary image using systematic dithering. be. When binarized using this systematic dithering method, the outline becomes blurred, so preprocessing such as outline enhancement is performed. Such preprocessing often introduces noise, resulting in degraded reproduced image quality. Therefore,
There is a need for a binarization method that involves less noise.

[Conventional technology]

A conventional example of binarizing multi-value image data such as a natural image is, for example, as shown in FIG. etc., is read into a processor etc., performs dithering processing using the systematic dither method, and outputs the resulting binary image data to a binary thermal printer, etc. be.

It is also known to perform edge enhancement as a preprocessing of dithering processing. For example, as shown in FIG. 6, multi-value image data is read into a processor, etc., the multi-value image data is subjected to contour enhancement processing to 0, the contour-enhanced multi-value image data is dithered to 0, It is a [phase] that outputs binary image data.

In the systematic dithering method, for example, as shown in Bayer's pattern in Figure 7, a darkness value of O to 3 is determined for 2 x 2 pixels, and compared with the level of each pixel, the value of the pixel is lower than the threshold value. If the darkness is large, it is black, and if the darkness is smaller than the darkness value, it is white. In addition, in the case of a pattern for 4 x 4 pixels, it corresponds to a pattern in which a constant is multiplied and added to a 2 x 2 pixel pattern, and in the thick line area, the darkness value of each 2 x 2 pixel is multiplied by 4. This is the threshold value.

Since this type of dark value pattern is used for binary conversion, even pixels of the same gradation will have different binary representations depending on their position.For example, the contours of linear parts such as letters and figures can be changed by dithering processing. The image becomes distorted, making it difficult to discern details on the playback screen, and in many cases making it impossible to read characters.

Therefore, as shown in FIG. 6, contour enhancement processing 0 is added as preprocessing. Various means are known for this edge enhancement, and for example, a means for processing using an edge enhancement operator shown in FIG. 8 is known.

[Problem to be solved by the invention]

As mentioned above, by binarizing multivalued image data using systematic dithering after pre-processing for edge enhancement, it is possible to prevent distortion of linear parts such as characters and figures. Noise such as isolated points is also emphasized, which has the disadvantage of deteriorating the reproduced image quality.

The present invention aims to remove noise and perform binarization.

[Means to solve the problem]

The image disulfuration control method of the present invention removes noise by median filter processing, and will be explained with reference to FIG.

A means 1 such as a processor that reads multivalued image data from a recording device, a television camera, etc.; a means 2 that performs median filter processing on the multivalued image data read by this means 1; The output means 5, such as a binary thermal printer, is equipped with means 3 for performing edge enhancement processing on multivalued image data, and means 4 for performing dithering processing on the multivalued image data whose contours have been enhanced by this means 3.
It outputs value image data.

[Effect]

Multivalued image data is read and preprocessed with median filter processing and edge enhancement.Median filter processing arranges multiple pixels in a window centered on the pixel of interest in order of level, and sets the center level as the pixel of interest. Since a black isolated point becomes white when other pixels within the window are close to white, noise is removed.

Then, after the contour is emphasized, it is dithered using a systematic dithering method to become a binary image signal, resulting in a reproduced image in which the contour does not collapse and noise is removed.

It is also possible to read the multivalued image data by means 1, perform edge enhancement processing by means 3, and then perform median filter processing by means 2.

In this case, noise components such as isolated points increased by the contour enhancement process can be removed by the median filter process.

〔Example〕

Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 2 is a block diagram of an embodiment of the present invention, in which 11 is a processor (CPU), 12 is a read-only memory (
ROM), 13 is random access memory (RAM)
), 14 is a hard disk device (HD), 15 is a television camera (TV), 16 is an AD converter (A/D), 1
7 is a binary thermal printer.

Read-only memory 12 and random access memory 13
The processor 11 realizes the functions of the multivalued image data reading means 1, the median filter processing means 2, the edge enhancement means 3, and the dithering means 4 in FIG. .

Therefore, the multivalued image signal obtained by the television camera 15 is converted into a digital signal by the AD conversion section 16,
The information is recorded on the hard disk device 14 under the control of the processor 11.

The processor 11 reads multivalued image data from the hard disk device 14 based on the program stored in the read-only memory 12 and stores it in the random access memory 13.
Median filter processing is performed using, etc. This median filter has, for example, a window configuration shown in FIG.
It is a cross-shaped window centered on the pixel of interest A. For example, if the level of the pixel of interest A is 1 and the level of the other pixels a w d is O, if you arrange them in order of level, it will be "10000" and the window will be in the center. Since the pixel level of the third bit located is 0, the target pixel A is set to O. Therefore, isolated points can be removed.

The multivalued image data subjected to the median filtering process is subjected to contour enhancement processing, and then dithered using a systematic dithering method to form binary image data, which is output to the binary thermal printer 17.

FIG. 4 is a flowchart of an embodiment of the present invention, in which multivalued image data is written from an image digitizer (an AD converter that converts an analog image signal into a digital signal) to the hard disk 14, and the multivalued image data is transferred from the hard disk 14. The processor 11 reads (■) and performs median filter processing on the read multivalued image data (■). That is, isolated points are removed using a median filter having a window configuration as shown in FIG.

The multivalued image data subjected to the median filter processing is subjected to contour enhancement processing (1), and the multivalued image data subjected to the contour enhancement processing is subjected to a systematic dithering method to be binarized (2). Then, the binarized image data is transferred to a binary thermal printer 17.
, and print out the gradation display image.

It is also possible to reverse the order of steps ① and ② in the above embodiment, and the same effect can be obtained.

Further, after step (2) in the above embodiment, as shown by the dotted line path, the level of the pixel of interest within the window of the median filter is determined (2), and the level of the pixel of interest within the window of the median filter is determined as shown by the dotted line path. It compares the level of the pixel with the pixel level and determines whether it is the highest or lowest. If it is the highest or lowest, proceed to step ■ and perform median filter processing. If it is not the highest or lowest, the pixel of interest is subjected to median filtering. The image is output as is to the edge enhancement processing means without any processing.

That is, when the pixel of interest is an isolated point, for example, the pixel of interest A within the window of the median filter shown in FIG.
Since this is the highest or lowest level compared to the level ~d, this isolated point can be removed by median filter processing. Furthermore, unnecessary median filter processing can be applied to a screen with few noise components to prevent the outline from becoming blurred.

The window configuration of the median filter in the above embodiment is as follows:
Various shapes other than a cross and configurations with a number of pixels other than the 5-pixel configuration can be used. It is also possible to transmit binary image data to the other party's terminal device via a transmission path.

〔Effect of the invention〕

As explained above, in the second image binarization control method for converting multivalued image data into binary image data using a systematic dither method, the present invention pre-processes multivalued image data by the median filter processing means 2. At the same time, the contour enhancement processing means 3 performs contour enhancement processing and binarizes it by a systematic dithering method, and the contour enhancement process prevents the collapse of the contour due to the binarization by the systematic dithering method. However, since the increase in noise caused by this edge enhancement processing can be prevented by the median filter processing, there is an advantage that deterioration in the image quality of the binarized image can be prevented. Also, when edge enhancement processing is performed after median filter processing is performed first, edge enhancement is performed after noise components are removed by median filter processing, making it possible to prevent image quality deterioration of the binarized image. .

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram of the principle of the present invention, FIG. 2 is a block diagram of an embodiment of the present invention, FIG. 3 is an explanatory diagram of a median filter window, FIG. 4 is a flowchart of an embodiment of the present invention, and FIG. 6 and 6 are flowcharts of conventional binarization, FIG. 7 is an explanatory diagram of Bayer's pattern, and FIG. 8 is an explanatory diagram of an edge enhancement operator. 2 is a means for reading, 2 is a means for median filter processing, 3 is a means for edge enhancement processing, and 4 is a means for dithering processing.

Claims (1)

[Scope of Claims] An image binarization control method for binarizing multivalued image data by a systematic dither method, comprising means (1) for reading the multivalued image data;
); means (2) for performing median filter processing on the multivalued image data read by the means (2); means (3) for performing edge enhancement processing on the multivalued image data filtered by the means (2); An image 2 characterized in that it comprises means (4) for performing dithering processing on the multivalued image data whose contours have been enhanced by step 3), and outputs the multivalued image data as binary image data by the means (4). Value control method.