JPH07322060A - Image area separation device, facsimile equipment provided with the device and electronic file device - Google Patents

Abstract

PURPOSE:To clearly read coexisting characters and to display the patterns of a dot and a photograph without generating a moire pattern by judging the respective edges of a character area and a dot area and selecting picture element values which are respectively processed in accordance with judgement results. CONSTITUTION:A line sensor 101 reads an original 103 where the patterns of the character and the dot coexist, and an edge detection means 106 detects the edge part of the character area. Then, the edge of the character area and the edge of the dot area are judged. A selection part 112 selects a character information picture element which is processed in an edge emphasis part 107, and a density correction part 108 based on the judged result and whose edge is emphasized or the pattern information of the dot, in which background density is set to be white by a gamma correction part 109 and a moire suppression edge emphasis part 111 and the generation of the moire pattern is prevented without the over-emphasis of edge density.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a facsimile, an electronic file, etc. having a function of reading an image, and more particularly, when an image containing characters, halftone dots, photographs and the like is read, the edge area of the character portion is The present invention relates to an image area separating device suitable for easily and satisfactorily separating the image area.

[0002]

2. Description of the Related Art As a conventional technique for distinguishing an image read by an image scanner or the like into a character image region and a pattern image region, there is a method described in Japanese Patent Publication No. 5-51225. In this conventional technique, an average density value in a window composed of a pixel to be image-processed and its peripheral pixels is calculated, and if the average density value is within a range in which there is a difference with the density value of the pixel of interest. For example, it is considered as a halftone image.

[0003]

In the above-mentioned prior art, since the image area is discriminated only by the information of the density value and the average density value of the pixel of interest, the image area is represented by the black pixel and the white pixel like the character. There is a problem that a halftone dot photograph area is erroneously determined as a character area.

An object of the present invention is to provide an image area separating apparatus which does not erroneously determine a halftone dot picture area as a character area, a facsimile apparatus and an electronic file apparatus using the same.

[0005]

In the image area separation device for capturing an image in which a character image and a halftone image are mixed, is the pixel belonging to the edge of the character region in the captured image? An edge detecting means for determining whether or not there is a pixel, a first means for outputting a pixel value after performing edge enhancement processing and density correction processing for a pixel determined to be an edge in the captured image, and the captured image On the other hand, the gamma correction processing and the second means for outputting the pixel value after performing the edge enhancement for the continuous edge and the processing for the isolated edge as the smoothing, and the edge detecting means determine the edge of the character area. In this case, the pixel value output from the first means is selected, and when the edge detection means determines that the edge is not an edge, the pixel value output from the second means is selected. And-option unit, the pixel value selected by said selection means by providing the means for error diffusion processing on the binarized output is achieved.

[0006]

When the edge of the character area and the edge of the halftone dot area exist, the edge of the character area is emphasized and the halftone dot area edge is not so emphasized. As a result, the characters are clearly displayed, and the halftone dot area is prevented from being over-emphasized to form a moire pattern.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a configuration diagram of a main part of an image area separation device according to an embodiment of the present invention. The sensor 101 is, for example, a contact line sensor used in a facsimile, and has a configuration in which photoelectric conversion elements for 2048 pixels are arranged side by side, and image information is provided in pixel units for each scanning line on the document 103. Convert to analog voltage. The A / D converter 104 includes the sensor 10
The analog signal output of 1 is converted into a digital signal.

The sensor characteristic correcting means 105 is one of the sensors.
A signal obtained by correcting the characteristic variation of the photoelectric conversion element for each pixel, the temperature variation, and the deterioration characteristic over time, and normalizing the image signal is output. For example, before reading the original, the white reference plate 102 is read by the sensor 101, and the sensor output for one main scanning line is stored in the memory. This signal is S (x). Next, the sensor output for one main scanning line at the time of reading the document 103 is set to P (x). Further, the output of the A / D converter 104 has 8-bit precision, black information is "0", and white information is "25".
If 256 gradation output of 5 "is made, the sensor characteristic correction 105
Output V (x) can be expressed by the following equation 1. x is a variable indicating a pixel position, and the leftmost pixel position of the main scanning 2048 pixels is “0” and the rightmost pixel position is “2047”.

[0009]

## EQU1 ## V (x) = 255 / S (x) .P (x) This function is generally well known as shading correction.

The edge detecting means 106 detects the edge portion of the character area from the image information, as will be described later in detail.

The edge emphasizing means 107 converts the image information input pixel by pixel into black information if it is close to black and white information if it is close to white, thereby converting the edge information. It has the function of correcting blur in the area and adding sharpness to the binary output. For example, as shown in FIG. 4, an image area composed of 3 pixels in the main scanning and 3 pixels in the sub-scanning centered on the target pixel for edge enhancement indicated by a square in a thick frame, and a coefficient 107 corresponding to each pixel position. Edge enhancement is realized by performing a convolution operation with a digital filter such as. The density value at pixel position x is V
(X), where the edge emphasis output is E (x), the edge emphasis output E (P12) at the pixel position P12 is the convolution operation of the following equation 2.

[0012]

[Equation 2] E (P12) = 5 · V (P12) -V (P02) -V (P11)
It is represented by -V (P13) -V (P22).

As shown in the graph of FIG. 4, the characteristics of this edge enhancement filter are such that the gain becomes higher in the main scanning direction, the sub-scanning direction and the oblique 45 degree direction as the spatial frequency (the number of changes in density per 1 mm) increases. Therefore, in both the character region and the halftone dot region, the edge where the density change is sharp is emphasized.

The density correction means 108 adjusts the density by adding or subtracting a certain fixed value to the edge-enhanced output. For example, when the density of the edge emphasis output is further increased, it is realized by adding a preset constant value.

The gamma correction means 109 performs density correction on the normalized image signal according to the input pixel density, and adjusts the contrast of gradation expression and the brightness of pseudo halftone output. . This function can be realized, for example, by inputting the pixel density to an address of the random access memory and assigning the gamma correction result to the data output of that address.

The document background density detecting means 110 reads several main scanning lines at the leading end of the document 103 by the sensor 101 at the start of reading the document and detects the minimum density value thereof.
Then, the correction characteristic of the gamma correction 109 is corrected so that the detected minimum density value is output as white when binarized. As a result, even if the background of the original is not white as in a newspaper, for example, the binarized output can make the background white, which is effective in shortening the transmission time in the facsimile.

The moiré suppressing edge emphasizing means 111, when the image information input on a pixel-by-pixel basis is a pixel in the edge portion of a character area or a picture area, is closer to black, and is closer to white, and is closer to white. The blur of the edge area is corrected by emphasizing the information of (1) to add sharpness to the binary output. However, if it is a pixel at the edge of the halftone dot photographic area, the edge emphasis is reduced to prevent moire from occurring in the halftone dot area during binary output. For example, as shown in FIG. 5, an image region 301 composed of 5 pixels in the main scanning and 3 pixels in the sub-scanning centering on the target pixel for edge enhancement indicated by a square in a thick frame, and a coefficient corresponding to each pixel position. It is realized by performing a convolution operation with a digital filter such as 111. Assuming that the density value of the pixel position x is V (x) and the moire suppression edge emphasis output is M (x), the moire suppression edge emphasis output M (P12) of the pixel position P12 is the convolution calculation of the following formula 3.

[0018]

[Equation 3] M (P12) = 5/2 ・ V (P12) +3/8 (V (P02)
+ V (P11) + V (P13) + V (P22))-5/8 ・ (V (P01)
+ V (P03) + V (P21) + V (P23))-1/8 ・ (V (P00)
It is represented by + V (P04) + V (P20) + V (P24)).

The moire suppressing edge enhancement filter 111
As shown in the graph in Fig. 5, the characteristics of the spatial frequency (1 mm
The gain becomes higher in the main scanning direction and the sub-scanning direction as the number of gradation changes per hit increases, but in the oblique 45 ° direction, the gain decreases as the spatial frequency increases. This is to increase the edge emphasis degree when high-density pixels are continuous in the vertical and horizontal directions, such as the outline of a character or pattern, and to check high-density and low-density pixels in a checkered area such as a halftone dot area. This is to reduce the edge enhancement degree when the patterns are alternately present. With this function, it is possible to express the contours of characters and patterns with sharpness in the binarized output, and it is possible to prevent the occurrence of moire due to the edge emphasis of the halftone dot area.

The selection means 112 is an edge detection means 106.
One of the two inputs is selected and output according to the output result of. If it is determined to be an edge, the edge enhancing means 107
The pixel having the higher density is selected from the pixels processed by the density correction unit 108 and the pixels processed by the gamma correction unit 110 and the moire suppression edge enhancement unit 111. If it is determined to be a non-edge, the pixel processed by the gamma correction unit 110 and the moire suppression edge enhancement unit 111 is selected.

The error diffusion means 113 converts the average gradation of a plurality of pixel areas into an output ratio of the number of black pixels and the number of white pixels, so that a pseudo halftone is expressed.

Next, the operation of the image area separation device shown in FIG. 1 will be described. First, the white reference plate 102 is read by the sensor 101, and the characteristics of the sensor 101 for one main scanning line are stored in the memory in the sensor characteristic correction unit 105. Manuscript 10
When the input of No. 3 is started, the document background density detecting unit 110 detects the minimum density value from the pixel density data of the document front end that is normalized and output using the sensor characteristics stored in the sensor characteristic correcting unit 105.

Then, the correction characteristic of the gamma correction means 109 is changed so that the minimum density value is converted into white. After that, the edge detection is performed pixel by pixel until the reading of the original is completed.
The higher of the pixel densities edge-enhanced by 7 and further shifted to a higher density by the density correction means 108 and the pixel density gamma-corrected by the gamma correction means 109 and processed by the moire suppression edge emphasis means 111 is selected. When it is determined that the image is not an edge, the gamma correction unit 109 corrects the background density of the document to white, the contrast and brightness of the pattern area are corrected, and the moire suppression edge enhancement unit 111 corrects the pixels of the contour portion in the pattern. Is edge-enhanced and the pixel density in which the edge enhancement in the halftone dot is prevented is selected.

Finally, the pixels output from the selection means 112 are uniformly subjected to the pseudo halftone binarization processing by the error diffusion means 113, so that the character area is dark and sharp, and the halftone area and the halftone dot photographic area are It is possible to perform high-gradation expression, and the background of the document can be binary output with white.

FIG. 2 is a diagram showing a reference area composed of 5 pixels in the main scanning direction and 3 pixels in the sub-scanning direction for determining whether or not the pixel of interest is an edge by the edge detecting means 106. is there. If the position of the pixel of interest is P12, P
The edge value of the pixel of interest P12 is detected with reference to the density values of the pixels located from 00 to P24.

FIG. 3 is a detailed functional block diagram of the edge detecting means 106. The background gray value detecting means 201 detects the average minimum value of two pixels (the level closest to white) from the pixels around the pixel to be subjected to edge determination. The purpose is
This is because the background density of the document is highly likely to exist around the edge in the character area, and therefore the background density of the document is selected as the minimum average value of two pixels. Assuming that the minimum average value of the two surrounding pixels is MNA,

[0027]

[Formula 4] MNA = MIN (Th1, Th2, Th3, T
h4, Th5, Th6) where Th1 = 1/2 (P00 + P01) Th2 = 1/2 (P03 + P04) Th3 = 1/2 (P10 + P11) Th4 = 1/2 (P13 + P14) Th5 = 1 / 2 (P20 + P21) Th6 = 1/2 (P23 + P24)

Specifically, as indicated by reference numeral 201 in FIG. 6 which is a detailed circuit diagram of the edge detecting means, the density values of two adjacent pixels are input to the averaging circuit 407 and the minimum value thereof is selected as the minimum value selecting circuit. It is sufficient to output from 408. The averaging circuit 407 can be realized by an adder, and the minimum value selection circuit 408 can be easily realized by a combination of comparators.

The spatial frequency detecting means 202 of FIG. 3 detects the spatial frequency characteristic of the peripheral pixel area centering on the pixel for edge detection. Specifically, a value obtained by subtracting the density value of the pixel of interest P12 from the output of the moire suppression edge enhancement filter 111 for the pixel of interest P12 of FIG. 2 is output. When the density value of the target pixel P12 is V (P12), the moire suppression edge enhancement filter output value is M (P12), and the output of the spatial frequency detection 202 is F (P12), the following expression 5 is given. In addition, M (P12) is the above formula 3

[0030]

[Equation 5] F (P12) = M (P12) -V (P12)

As a result, F (P12) outputs a large value in the character area, a small value in the halftone dot area, and almost "0" in an area where the gradation changes gently such as a photograph area.
Is output.

The judgment threshold value calculation means 203 determines a judgment threshold value for judging whether or not the pixel of interest P12 is an edge of a character area. Since the output of the spatial frequency detection 202 is F (P12), assuming that the output of the background density value detecting means 201 is K (P12) and the output of the constant setting register 404 is a, the determination threshold TH is Formula 6

[0033]

[Equation 6] TH = a−F (P12) + K (P12)

As a result, in the character area, F (P12) has a large value and K (P12) has a small value, so TH has a small value. In the halftone dot area and photo area, F
Since (P12) has a small value and K (P12) has a large value, TH has a large value. Specifically, as indicated by reference numeral 203 in FIG. 6, the output of the constant setting register 404 and the output of the spatial frequency detection 202 are input to the subtractor 405, and then the output and the output of the background density value detection 201 are added. Bowl 40
The result of addition in 6 is used as the judgment threshold value. Also, the constant a
Depending on the characteristics of the reading system such as the sensor 101.
It can be set freely from 14.

The judgment target value calculating means 204 determines the pixel P of interest.
12 is replaced with the maximum value MXV of the average value of two pixels adjacent to each other. The judgment target value calculation output PB is expressed by the following formula 7

[0036]

## EQU7 ## MXV = 1/2 (P12 + PP) where PP = MAX (P01, P02, P03, P11, P13, P21,
P22, P23).

The purpose of this is that, in a region such as a character region where high-density pixels are continuous in the vertical, horizontal, and diagonal directions, even if the maximum value of the two-pixel average value is replaced, the same gray value as the target pixel is obtained. On the other hand, in a checkerboard-like area where high-density pixels and low-density pixels alternate such as a halftone dot area, even if the maximum value of the two-pixel average value is replaced, the density value becomes lower than the pixel of interest. This is because the edge determining means 205 described below acts to make the value smaller than the determination threshold value and contributes to the prevention of erroneous edge detection in the halftone dot area.

The edge determining means 205 compares the determination target value with the determination threshold value and determines whether the target value P12 is the edge of the character area. When the judgment target value is MXV and the judgment threshold value is TH, the judgment is made by the following formula 8

[0039]

If MXV> TH, P12 = edge If MXV ≦ TH, P12 = non-edge

In the case of the edge of the character area, the judgment target value MX
Since V has a large value and the determination threshold TH has a small value, the pixel of interest P12 is determined as an edge. On the other hand, in the case of the halftone dot area and the halftone area, the determination target value MXV
Becomes a small value and the determination threshold value TH becomes a large value, so that the pixel of interest P12 is determined as a non-edge. Specifically, as indicated by reference numeral 206 in FIG. 6, the comparator 409
Can be configured with.

[0041]

According to the present invention, when an image in which characters and patterns such as halftone dots and photographs are mixed is binarized, the character region has high density and a clear outline, and halftone dots are used. In a picture area such as a photograph, moire does not occur, high gradation expression is performed, and high image quality binary output in which the background density of a document is white can be obtained with a simple circuit configuration.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an image area separation device according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram of edge detection.

FIG. 3 is a functional configuration diagram of an edge detection unit shown in FIG.

FIG. 4 is an explanatory diagram of edge enhancement processing.

FIG. 5 is an explanatory diagram of moire suppression edge enhancement processing.

FIG. 6 is a specific circuit configuration diagram of an edge detection unit.

[Explanation of symbols]

Reference numeral 101 ... Sensor, 102 ... Blocking white reference plate, 103 ... Original, 104 ... A / D converter, 105 ... Sensor characteristic correction section, 106 ... Edge detection section, 107 ... Edge enhancement section, 1
08 ... Density correction unit, 109 ... Gamma correction unit, 110 ... Original background density detection unit, 111 ... Moire suppression edge enhancement unit,
112 ... Selection unit, 113 ... Error diffusion unit, 114 ... MP
U, 201 ... Background density detection unit, 202 ... Spatial frequency detection unit, 203 ... Judgment threshold value calculation unit, 204 ... Judgment target value calculation unit, 205 ... Edge judgment unit, 301 ... Reference pixel area, 302 ... Reference pixel Area, 401 ... Maximum value selection circuit,
402 ... Averaging circuit, 403 ... Subtractor, 404 ... Constant setting register, 405 ... Subtractor, 406 ... Adder, 407
... averaging circuit, 408 ... minimum value selecting circuit, 409 ... comparator.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical display location G06K 9/20 340 L 7459-5L G06F 15/70 330 Z (72) Inventor Tomoo Kobori Yokohama, Kanagawa Prefecture 292 Yoshida-cho, Totsuka-ku, Yokohama-shi, Ltd.

Claims (11)

[Claims] 1. An image area separation apparatus for capturing an image in which a character image and a halftone image are mixed, and edge detection means for determining whether or not the captured image is a pixel belonging to an edge of a character region. A first means for outputting a pixel value after performing edge enhancement processing and density correction processing for pixels determined to be edges in the captured image; and gamma correction processing and continuous edge processing for the captured image. A second means for outputting a pixel value after performing a process for acting as an edge enhancement on an isolated edge and a process for acting on an isolated edge as a smoothing; and an output from the first means when the edge detecting means determines that it is an edge of a character area. Selecting a pixel value to be selected and selecting the pixel value output from the second means when the edge detecting means determines that the edge is not an edge, and the selecting means. An image area separation device comprising: means for performing error diffusion processing on the processed pixel values to generate a binary output. 2. An image area separating apparatus for capturing an image in which a character image and a halftone image are mixed, and an edge detecting unit for determining a pixel belonging to an edge of a character area, and a pixel determined to be an edge of the character area. For, the means for selecting the higher pixel density value before and after the gamma correction, and the pixel density after the gamma correction for the pixels determined not to be the edge of the character area An image area separation device comprising means for selecting a value. 3. The edge detection means according to claim 1 or 2, wherein the edge detection means sets a threshold value for determining an image area from an image in which a character image and a halftone image are mixed, to a density of peripheral pixels of a target pixel of the determination. An image area separation device comprising means for determining from information and spatial frequency information. 4. The edge detecting means according to claim 3,
An image area separating device comprising means for judging an image area of a target pixel by comparing the calculation result of a pixel which is an object of image area judgment and a peripheral pixel thereof with a threshold value for image area judgment. . 5. The image information according to any one of claims 1 to 4, wherein the edge determination and / or the binarization of the pixel determined to be the edge are density-corrected so that the white reference plane becomes a white peak. An image area separating device, characterized in that it is provided with means for performing the binarization processing of the other portion by using the image information whose density is corrected so that the background density of the document is output in white. 6. The image area separation device according to claim 5, further comprising means for detecting the minimum density at the leading edge of the document and changing the gamma correction characteristic so that the minimum density is corrected to white. 7. The detection of spatial frequency information according to claim 3, wherein a high value is output when the spatial frequency in the main scanning direction and the spatial frequency in the sub scanning direction are significantly different, and the values in both directions are the same value. The image area separation device is characterized by using the output of a filter that outputs a low value. 8. The image area separation device according to claim 3, wherein the grayscale information of the peripheral pixels adopts a value closest to white among the average values of two or more pixels located around the determination target pixel. . 9. An image area separation device for capturing and processing an image in which a character image and a halftone image are mixed is provided with means for performing different image processing for pixels belonging to the edge of the character area and other pixels. Image area separation device. 10. A facsimile apparatus comprising the image area separation device according to claim 1. Description: 11. An electronic file device comprising the image area separation device according to claim 1. Description: