JPH0779349A - Picture recording method - Google Patents

Abstract

PURPOSE:To prevent deterioration in the reproducibility of thin lines by pseudo multi-value processing by executing the pseudo multi-value processing and edge extract processing independently and synthesizing pseudo multi-value processing data and edge data obtained through the processing as above. CONSTITUTION:A picture input signal 101 read by an image scanner or the like and transmitted is received by a 5X5 filter 102. The 5X5 filter 102 has a predetermined coefficient to eliminate noise in the picture input signal 101 to generate a signal with less comparatively and without deterioration of the reproducibility of thin lines. Moreover, the picture input signal 101 is processed by a 3X3 Laplacian filter 106 in parallel independently of the 5X5 filter 102 and subjected to binarization processing 108 based on an edge data slice level 107 to generate edge data 108. The pseudo multi-value processing data 108 and the edge data 108 are 0Red selectively by an OR circuit 110 to form a binarized picture output 111 and the picture is recorded based on the output.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image recording method suitable for a digital copying machine or the like for forming a digital image data recording image.

[0002]

2. Description of the Related Art In an image recording apparatus such as a digital copying machine, an original image to be recorded / reproduced is pixel-wise scanned by an image scanner or the like and converted into digital image data (binary electric signal). The digital image data is processed by a spatial filter for "noise removal" and "feature extraction". The spatial filter typically takes out 3 × 3 or 5 × 5 pixels in a certain pixel space and multiplies the pixel density of each pixel by a coefficient to obtain a total sum. For example, a certain point (i , f) is the pixel density of f (i, f),

[Equation 1] Coefficient matrix for the 3 × 3 region

[Equation 2] The sum of products is calculated as follows: g (i, j) = af (i-1, j-1) + bf (i, j-1) + cf (i
+ 1, j-1) + lf (i-1, j) + mf (i, j) + nf (i + 1,
j) + xf (i-1, j + 1) + yf (i, j + 1) + zf (i + 1, j +
1) To obtain a new pixel density, which removes unnecessary noise.

Known as a representative of spatial filters,

[Equation 3] The average value filter having the coefficient of (1) removes noise by averaging the pixel densities of the peripheral pixels including the pixel of interest and crushing the projected density values. The central value is set to "2" so that the pixel of interest is weighted to reduce the blur of the image in the averaging.

In addition to the average value filter, a Laplacian filter is known as a filter for extracting a density change of an image.

[Equation 4] An image in which only the portion where the density has changed is extracted can be obtained by passing it through a filter having a coefficient.

On the other hand, in the method of binarizing image data having a certain value width, such as "1" if the slice level is larger than that value and "0" if it is smaller, a photograph This causes a remarkable deterioration in image quality with respect to an image having gradation (halftone) like an image. In order to eliminate this, pseudo multi-value conversion is performed by a binarization method called a dither method. In this method, for example, as shown in FIG. 2, 16 slice levels are prepared in a 4 × 4 pixel space, arranged in a predetermined order, and binarized by the value.

[0006]

The method as described above works properly and effectively in removing noise.
When the target pixel is weighted by using the averaging filter and the averaging process is performed, some blurring of the image cannot be avoided, and the reproducibility of the fine line is particularly unfavorable.

[0007]

SUMMARY OF THE INVENTION The present invention has been made in view of the above points, and provides an image recording method for forming a recorded image based on a processed signal by subjecting image input data to filter processing and pseudo-multivalue processing. Pseudo multi-valued processing and edge extraction processing are independently performed on the image input signal, and the pseudo multi-valued data obtained by the pseudo multi-valued processing and the edge data obtained by the edge extraction processing are combined. It is characterized by forming an image output.

[0008]

With this method, noise is removed by the pseudo multi-value quantization process, and the edge data suppresses the deterioration of the reproducibility of the fine line.

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to the drawings.
FIG. 1 shows an example in which the image recording method according to the present invention is embodied. An image input signal 101 read by an image scanner or the like and sent is input to a 5 × 5 filter 102. The 5 × 5 filter 102 is, for example,

[Equation 5] Has a coefficient of, removes noise of the image input signal 101,
A signal is formed that suppresses the deterioration of the reproducibility of a fine line with relatively no blur. However, if the signal obtained by performing such filter processing is subjected to the pseudo multi-value quantization processing 104 using the dither matrix slice level 103 as a reference by the dither method to form the pseudo multi-value data 105, noise or the like can be removed. , A signal with poor reproducibility of thin lines.

Therefore, the image input signal 101 is processed by the 3 × 3 Laplacian filter 106 in parallel independently of the 5 × 5 filter 102, and then binarized 108 based on the edge data slice level 107. The edge data 109 is formed. The pseudo multi-valued data 105 and the edge data 109 thus formed are connected to the OR circuit 110.
To selectively combine to form a binary image output 110,
Image recording is performed based on this output.

[0011]

As described above, according to the present invention, the pseudo multi-value processing and the edge extraction processing are independently performed, and the pseudo multi-valued data obtained by these processings is combined with the edge data to realize the pseudo multi-valued processing. It is possible to suppress deterioration of reproducibility of thin lines due to processing. Further, since the slice level for the dither matrix can be set independently, it is possible to obtain a coordinated image signal even for a fine line in a whitish halftone original.

[Brief description of drawings]

[Figure 1]

[Fig. 2]

[Explanation of symbols]

 102 5 × 5 filter 104 Pseudo multi-valued processing 105 Pseudo multi-valued data 106 3 × 3 Laplacian filter 108 Binarization processing 109 Edge data 110 Binary image output

Claims (1)

[Claims] 1. An image recording method for forming a recorded image based on a processed signal by subjecting image input data to a filtering process and a pseudo-multi-valued process, wherein:
Pseudo multi-value processing and edge extraction processing are performed independently,
An image recording method, wherein an image output is formed by synthesizing pseudo multivalued data obtained by the pseudo multivalued processing and edge data obtained by the edge extraction processing.