JPH0851538A - Image processor - Google Patents

Abstract

PURPOSE:To provide an output of an image density signal with fidelity to an original by discriminating a state of a notice picture element of the original in multi-stage between gradation images from a binary image, selecting a density conversion curve depending on the result of discrimination and applying density conversion to each picture element to reduce a sense of incongruity of the density of an outputted image. CONSTITUTION:Notice picture element signals di-1, j-5 from a line memory 2 are converted into a density signal by a density conversion circuit 3 and the resulting signal is outputted. In this case, an edge detection circuit 27 calculates the sharpness of the edge of the density of a center picture element based on picture element signals of a center picture element and adjacent picture elements of an area in the vicinity of the notice picture element and compares the result with a threshold level and three stages of sharpness signals are fed to a characteristic discrimination circuit 28 via delay circuits 8, 9, 26. The circuit 3 selects a conversion curve for a binary image signal and a conversion curve for a multi-value image signal based on the result of discrimination of the circuit 28 to convert the picture signal into a density signal. A sense of incongruity due to a rapid change in the density of an output image through the multi-value processing is reduced and an image with fidelity to an original in which thin lines, characters and gradation images are in existence in mixture is obtained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention reads an image of an original,
Image processing of the read digital image data to perforate the heat-sensitive stencil paper (digital printing machine), electrophotographic technology to form a dry image on the photoconductor and transfer it to paper (digital copying machine), Alternatively, it relates to an apparatus for copying on thermal paper or the like. In particular, it relates to an image processing apparatus that binarizes a read image and outputs it as binary data.

[0002]

2. Description of the Related Art In order to obtain the best result when a document in which a binary image such as a character or a line drawing and a gradation image such as a photograph are mixed is output using the above-mentioned image processing apparatus. It is necessary to binarize the value image area into either the maximum density or the minimum density with a single threshold value, and to perform density conversion that preserves the input and output density in consideration of the characteristics of the input and output devices in the gradation image area. . For that purpose, it is necessary to determine whether each part of the image is a binary image region or a gradation image region and separate the two regions.

Conventionally, in order to discriminate between a binary image area and a gradation image area, an image is divided into blocks of n × n pixels, feature extraction is performed for each block, and the block is extracted using the feature extraction result. (JP-A-3-153167), a method of performing feature extraction using a target pixel and its peripheral pixels, and a determination of each pixel using the feature extraction result (JP-A-1-27573). ) Etc.

[0004]

However, since the former method discriminates each block, the shape of the block is not formed in the erroneously discriminated portion or the boundary portion between the binary image area and the gradation image area. There is a problem that it will appear. When the latter method is used, the influence of misjudgment is less noticeable than that of the former method, but a difference in density occurs between the misjudged portion and the correctly discriminated portion, which causes a feeling of strangeness.

Further, it is difficult to distinguish between a thick line drawing or black solid in the binary image area and a high density portion of a photograph in the gradation image area. If a discrimination parameter is adjusted so that a thick line drawing or solid black can be discriminated as a binary image, a crushed portion will appear in the photographic image. If the discrimination parameter is adjusted so that the high density portion of the photograph can be discriminated as a gradation image, the density of a thick line drawing or black solid becomes light.

In order to solve the above-mentioned problems, the present invention provides a method for a document in which a binary image area and a gradation image area are mixed,
An object of the present invention is to provide an image processing apparatus that outputs an image density signal in which the contrast is strong in the binary image area, the gradation is preserved in the gradation image area, and a sense of discomfort due to a difference in density does not occur.

[0007]

In order to solve the above-mentioned problems, according to the present invention, the original image input by the original reading means is discriminated into a binary image and a gradation image and converted by a corresponding density curve. In an image processing device for outputting an image signal, a means for judging the state of a target pixel of an original from a binary image to a gradation image in multiple stages, a density conversion curve for a binary image, a gradation image for a gradation image according to the judgment. A density conversion curve selecting means for selecting an appropriate density conversion curve from the density conversion curve and one or more density conversion curves interpolating the two curves and converting the density is provided.

There are various methods for identifying the state of the image of interest used in the apparatus of the present invention.
(1) The sharpness of the edge of the target pixel is detected according to the density of the neighboring pixel to determine the state of the target pixel. (2) Whether or not the target pixel is an edge pixel is detected according to the densities of neighboring pixels, and the distance between the target pixel and the edge pixel closest to the target pixel is counted to determine the state of the target pixel. . (3)
Obtained by whether or not the pixel of interest is a thin line, whether or not the pixel of interest is an edge pixel, the degree of sharpness of the edge of the pixel of interest, the distance of the pixel of interest from the edge pixel, and the length of the line segment that sandwiches the pixel of interest. Judgment is made according to the characteristics of each pixel.

[0009]

According to the present invention, if there is a line segment that scissors the pixel of interest at the rising and falling edges of the density in the main scanning direction or the sub-scanning direction, the length thereof is counted. If the length of the line segment is short, it is highly possible that the line segment is a line forming a character. The present invention also counts the distance between the pixel of interest and the edge pixel closest to the pixel of interest. Here, the shorter the distance from the edge, the more like a character image, and the longer the distance, the more like "a gradation image".

According to the present invention, in addition to a binary image density conversion curve suitable for a binary image and a gradation image density conversion curve suitable for a gradation image, the gradation conversion is suitable for interpolating both curves. It has a number of density conversion curves.

From the above-mentioned several feature quantities estimated for each pixel, when the feature quantity exhibits a strong binary image-likeness, the binary image density conversion curve is selected, and the tone image likeness with a strong feature quantity is determined. In the case of showing, the gradation conversion density conversion curve is selected.
If neither of these is the case, several curves for interpolating the curves for the binary image and the gradation image are interpolated according to the degree of the "binary image-likeness" or "gradation-likeness" of the feature amount. An appropriate one is selected from the density conversion curves of and the density is converted. As a result, the density of the output image due to misjudgment of the region occurs when the feature amount does not show "binary image-likeness" or "gradation image-likeness" or when only the image of the region where the feature amount is narrow is seen. It is possible to reduce the discomfort caused by a sudden change in the.

By selecting the density conversion curve closer to the density conversion for the binary image as the edge is closer to the edge and the density conversion curve closer to the density conversion for the gradation image as the edge is farther from the edge, thick characters and solid portions are selected. Can be made to appear black, and the gradation of the high density portion of the gradation image can be preserved.

[0013]

EXAMPLES Examples of the apparatus of the present invention will be described below. (Embodiment 1) FIG. 1 is a block diagram of an image processing apparatus according to an embodiment of the present invention. The original is irradiated with light and the reflected light from the original is sent as an electric signal (image density signal) by a line sensor such as CCD
Converted to and output. The image density signal d _ij read by the reading means (not shown) is
It is input to the memory 2.

The image density signal output from the reading means is directly input to the line memory 2 or via the MTF correction circuit 1 shown by a dotted line for correcting the image density signal closer to the original image. Input to line memory 2.

In the case of passing through the MTF correction circuit, each coefficient of the matrix is multiplied by the density value of the pixel corresponding to the position of each coefficient by the MTF correction coefficient matrix as shown in FIG. 2 and the respective results are added. And the corrected image signal d ′ _ij is calculated.
Is input to the line memory 2.

FIG. 3 shows the positions on the original of the pixels corresponding to the image density signals stored in the line memory 2. The latest pixel corresponding to the image density signal output from the reading device is represented by p _ij . Here, j is the pixel number on the main scanning, i is the pixel number on the sub-scanning, and p _ij is the i-th pixel on the sub-scanning and the j-th pixel on the main scanning. The density of the pixel p _{i, j} is represented by d _{i, j} . The line memory 2 stores two main scans of a hatched portion and image density signals of two pixels.

Image density signals (d _{i, j} , d _{i, j-1} , d _{i, j-2} , d _{i-1, j} ,) output from the line memory 2
d _{i-1, j-1} , d _{i-1, j-2} , d _{i-2, j} , d _{i-2, j-1} , d
_{i-2, j-2} ) forms a 3 × 3 matrix centered on the pixel p _{i-1, j-1} .

As shown in FIG. 1, the image density signals d _{i-1, j-5} output from the line memory 2 are converted into the density conversion circuit 3
Is input to. Image density signals (d _{i, j} , d _{i, j-1} , d _{i, j-2} , d _{i-1, j} , d) output from the line memory 2
_{i-1, j-1} , d _{i-1, j-2} , d _{i-2, j} , d _{i-2, j-1} ,
d _{i-2, j-2} ) is input to the edge detection circuit 27. The edge detection circuit 27 is a circuit for detecting whether or not the pixel of interest is a density conversion point with a pixel adjacent to the pixel of interest.

FIG. 4 is a diagram showing the edge detection circuit 27. In the arithmetic circuit 13, (a), (b), (c) of FIG.
Convolution calculation is performed by the edge detection coefficient matrix in the horizontal direction, the vertical direction, the diagonal direction to the left, and the diagonal direction to the right shown in (d), and the largest value among the absolute values of the obtained four values is the edge. Output as signals e _{i-1, j-1}
It is input to the comparator 14, the comparator 15, and the comparator 25.

In the comparator 15, the input edge signal e
_{If i-1, j-1} is equal to or greater than the first edge threshold T2, it is valid (a signal indicating that the pixel is an edge pixel), and if less than T2, it is invalid (a signal that the pixel is not an edge pixel) and the first edge detection signal e1.
_{i-1, j-1} are output. The first edge threshold T2 is set to a value capable of detecting only a sharp edge (having a large density difference) that hardly appears in the gradation image.

The comparator 14 outputs the second edge detection signal e2 _i-1.j-1 as valid if the edge signal e _i-1.j-1 is greater than or equal to the second edge threshold T3 and invalid if less than T3. To be done. The second edge threshold T3 is smaller than the first edge threshold T2, and is set to a value at which a slightly sharp edge (with a slightly large difference in density) can be detected.

In the comparator 25, the input edge signal e
_{If i-1.j-1} is greater than or equal to the third edge threshold T8, the third edge detection signal e3 _i-1.j-1 is output. The third edge threshold T8 is smaller than the second edge threshold T3, and is set to a value at which a dull edge with almost no difference in density can be detected.

As shown in FIG. 1, the edge detection signals are supplied to the dot delay circuits 8, 9 and 26 in order to match the density conversion timing with respect to the image density signal d _i-1.j-5 of the pixel to be processed. Is entered.

The first edge detection signal e1 _i-1.j-5 , the second edge detection signal e2 _i-1.j-5 and the third edge detection signal e3 output from the dot delay circuits 8, ₉ and _26.
i-1.j-5 is input to the feature determination circuit 28. In the characteristic judging circuit 28, the input first edge detection signal e
_{1 i-1, j-5} , the second edge detection signal _{e2 i-1, j-5} , the third edge detection signal _{e3 i-1, j-5} , the density conversion selecting the rule shown in FIG. 6 signal g _{i -1, j-5} is determined and output.

The density conversion selection signal g _i-1.j-5 is 1. Edge pixels mainly exist in the binary image area. The degree of edge sharpness (density difference) is strong (large)
The density conversion curve is selected in consideration of image characteristics such that the higher the degree of sharpness of the edge is, the lower the degree of sharpness of the edge is.

In FIG. 6, ∘ indicates that the binary signal is valid, × indicates that it is invalid, and − indicates that it is ignored. For signals with three or more values, the numbers represent the signal values. Here, the value of the density conversion selection signal g _{i-1, j-5} is the binary image if g _{i-1, j-5} = 0, and the gradation if the value is g _{i-1, j-5} = 3. Image g _{i-1, j-5} = 1 means a binary image, and g _{i-1, j-5} = 2 means a gradation image.

As shown in FIG. 7, the density conversion circuit 3 binarizes a density conversion curve A for binary image (shown by a solid line) for binarizing the input density to either maximum or minimum, and Density conversion curve B for gradation image that stores gradation in output density
(Indicated by a one-dot chain line), a density conversion curve C (indicated by a dotted line) from a binary image density conversion curve that interpolates the above two curves, and a density from a gradation image density conversion curve It has a total of four density conversion curves of two density conversion curves consisting of the conversion curve D (shown by a two-dot chain line).

If g _{i-1, j-5} = 0 according to the density conversion selection signals g _{i-1, j-5} input from the feature judging circuit 28, the binary image density conversion curves A, g _{If i-1, j-5} = 3, gradation density conversion curve B, and if g _{i-1, j-5} = 1, density conversion curve C interpolating two curves, g _{i-1, j-5} = 2, the density conversion curve D that interpolates two curves is selected as a density conversion curve, and the image density signal d
The density conversion of _{i-1, j-5} is performed, and the converted density signal cd _{i-1, j-5} is output.

In the density conversion circuit 3, each of the four density conversion curves has four data conversion tables corresponding to each density conversion curve, and the density conversion selection signal g output from the feature judgment circuit 28. and outputs the converted density signal cd _i-1.j-5 performs density conversion of the image density signal d _i-1.j-5 with reference to the selected data conversion table in accordance with the _i-1.j-5 .

The converted density signals cd _{i-1, j-5} converted by the density conversion circuit 3 are input to the binarization circuit 4. The binarization circuit 4 binarizes the input converted density signal cd _{i-1, j-5} by an error diffusion method and outputs it as a binary signal.

In this embodiment, the edge detection coefficient matrix is 3 × 3. Matrix size is n × m
(N> 0, m> 0: n and m are integers), the coefficients are changed, or the number of matrices is increased,
The edge detection accuracy of the edge detection circuit can be improved.

In this embodiment, there are two density conversion curves for interpolating the density conversion curve for binary images and the density conversion curve for gradation images.
However, the number of edges may be one, and the number of comparators having different thresholds of the edge detection circuit 27 is increased to increase the classification of the degree of sharpness of the edge, thereby interpolating both curves. It is also possible to extend the concentration conversion curve to three or more.

(Embodiment 2) Another embodiment of the apparatus of the present invention will be described below. FIG. 8 is a block diagram of the image processing apparatus according to the embodiment of the present invention.

Image density signals (d _{i, j} , d _{i, j-1} , d _{i, j-2} , d _{i-1, j)} output from the line memory 2
d _{i-1, j-1} , d _{i-1, j-2} , d _{i-2, j} , d _{i-2, j-1} , d
_{i-2, j-2} ) constitutes a 3 × 3 matrix centered on the pixel p _{i-1, j-1} as in the first embodiment.

As shown in FIG. 8, the image density signals d _{i-1, j-5} output from the line memory 2 are converted into the density conversion circuit 3
Is input to. Image density signals (d _{i, j} , d _{i, j-1} , d _{i, j-2} , d _{i-1, j} , d) output from the line memory 2
_{i-1, j-1} , d _{i-1, j-2} , d _{i-2, j} , d _{i-2, j-1} ,
d _{i-2, j-2} ) is input to the edge detection circuit 29.

FIG. 9 is a diagram showing an edge detection circuit. In the arithmetic circuit 13, as in the case of the first embodiment, as shown in FIG.
Convolution calculation is performed by the edge detection coefficient matrix in the horizontal direction, the vertical direction, the left diagonal direction, and the right diagonal direction shown in (b), (c), and (d), and the absolute value of the obtained four values is calculated. The largest value among the values is the edge signal e _{i-1, j-1}
And is input to the comparator 15.

In the comparator 15, the input edge signal e
_A first edge detection signal e1 that is valid (a signal indicating that the pixel is an edge pixel) and is invalid (a signal indicating that the pixel is not an edge pixel) if _{i-1 and j-1} are equal to or greater than the first edge threshold value T2 and is less than T2.
_{i-1, j-1} are output. The first edge threshold T2 is set to a value capable of detecting only a sharp edge (having a large density difference) that hardly appears in the gradation image.

As shown in FIG. 8, the first edge detection signal e1 _i-1.j-5 output from the dot delay circuit 9 is input to the feature determination circuit 30. The distance classification circuit 11 counts and classifies the distance from the pixel of interest to the edge pixel. FIG. 10 shows the details of the distance classification circuit 11. The line memory 23 stores the distance data from the edge pixel of each pixel for one line.

FIG. 16 shows the positions of the pixels on the original corresponding to the distance data to the edge pixels of each pixel stored in the line memory 23. First edge detection signal el _{i-1, j-5}
Is input to a distance counting circuit 22, the pixel p _i-1, the distance de _i-1 from the edge of the _{j-5, j-5} is left pixel _{p i-1, j-6} , the upper pixel p _{i- 2, j-5} , the distance from the edge of the upper right pixel p _{i-2, j-4} , de _{i-1, j-6} , de _{i-2, j-5} , de, respectively
_{i-2, j-4} are used to count as follows.

If e1 _{i-1, j-5} = 1 then de _{i-1, j-5} = 0 If e1 _{i-1, j-5} = 0 then de _{i-1, j-5} = min ( de _{i-1, j-6} , de _{i-2, j-5} , d
e _{i-2, j-4} ) +1

The above de _{i-1, j-6} , de _{i-2, j-5} , de
_{i-2 and j-4} are fetched from the line memory 23 and de
_{i-1 and j-5} are newly stored in the line memory 23. The distance de _{i-1, j-5} from the edge is input to the comparator 24. If de _{i-1, j-5} <T6, then f _{i-1, j-5} = 0 T6 ≤ de _{i-1 , j-5} <T7, then f _{i-1, j-5} = 1 T7 ≦ de _{i-1, j-5} <T8, then f _{i-1, j-5} = 2 T8 ≦ de _{i-1 , j-5} , the classification is made according to the rule that f _{i-1, j-5} = 3. Distance classification signal f
_{i-1, j-5} are input to the feature determination circuit 30.

In the feature judging circuit 30, the density conversion selection is made according to the rule as shown in FIG. 11 according to the input first edge detection signals e1 _{i-1, j-5} and the distance classification signals f _{i-1, j-5} . Signal g
_{i-1, j-5} are determined. This density conversion selection signal g _i-1.j-5
Is 1. Edge pixels mainly exist in the binary image area. The shorter the distance between the pixel of interest and the edge pixel closest to the pixel of interest, the more likely it is to be a character image, and the longer the distance, the more likely it is to be a gradation image. The density conversion curve is selected in consideration of image characteristics such as.

In FIG. 11, ∘ indicates that the binary signal is valid, × indicates that it is invalid, and-indicates that it is ignored. For signals with three or more values, the numbers represent the signal values. Here, the value of the density conversion selection signal g _{i-1, j-5} is the binary image if g _{i-1, j-5} = 0, and the gradation if the value is g _{i-1, j-5} = 3. Image g _{i-1, j-5} = 1 means a binary image, and g _{i-1, j-5} = 2 means a gradation image.

The density conversion circuit 3 has four density conversion curves as in the case of the first embodiment.
Depending on the density conversion selection signal g _{i-1, j-5} input from 0

If g _{i-1, j-5} = 0, the density conversion curve for binary image A g _{i-1, j-5} = 3, the gradation density conversion curve B g _{i-1, j- If 5} = 1, a density conversion curve C that interpolates two curves (from the density conversion curve for binary image) If gi-1, j-5 = 2, a density conversion curve D that interpolates two curves (gradation image) The density conversion curve is selected according to the following rule.
The density conversion of _{i-1, j-5} is performed and the converted density signal cd _{i-1, j-5} is output.

In the density conversion circuit 3, the four density conversion curves have four data conversion tables corresponding to the respective density conversion curves, and the density conversion selection signal g output from the feature determination circuit 30. and outputs the converted density signal cd _i-1.j-5 performs density conversion of the image density signal d _i-1.j-5 with reference to the selected data conversion table in accordance with the _i-1.j-5 .

The converted density signals cd _{i-1, j-5} converted and output by the density conversion circuit 3 are input to the binarization circuit 4.
In the binarization circuit 4, the input converted density signal cd _{i-1, j-5}
Is binarized by the error diffusion method and output as a binary signal.
In this embodiment, similarly to the first embodiment, the accuracy of edge detection can be improved. Further, as in the first embodiment, the density conversion curve for interpolating the density conversion curve for the binary image and the density conversion curve for the gradation image may be one, and both curves are increased by increasing the number of classifications of the distance classification circuit 11. The density curve to be interpolated can be extended to three or more.

(Embodiment 3) Another embodiment of the apparatus of the present invention will be described below. FIG. 12 is a block diagram of the image processing apparatus according to the embodiment of the present invention.

The image density signals (d _{i, j} , d _{i, j-1} , d _{i, j-2} , d _{i-1, j)} output from the line memory 2
d _{i-1, j-1} , d _{i-1, j-2} , d _{i-2, j} , d _{i-2, j-1} , d
_{i-2, j-2} ) constitutes a 3 × 3 matrix centered on the pixel p _{i-1, j-1} as in the first embodiment.

As shown in FIG. 12, the image density signals d _{i-1, j-5} output from the line memory 2 are input to the high density line detection circuit 10 and the density conversion circuit 3. Image density signals (d _{i, j} , d _{i, j-1} , d) output from the line memory 2
_{i, j-2} , d _{i-1, j} , d _{i-1, j-1} , d _{i-1, j-2} , d _{i-2, j} , d
_{i-2, j-1} , d _{i-2, j-2} ) are input to the fine line detection circuit 5 and the edge detection circuit 7. The thin line detection circuit 5 is a circuit that detects whether or not the pixel of interest is a thin line. In the thin line detection circuit 5, the horizontal direction (main scanning direction) shown in (a) and (b) of FIG.
And the convolution calculation is performed by the fine line detection coefficient matrix in the vertical direction (sub-scanning direction), and the larger absolute value of the two obtained values is compared with the threshold value T1. Signal indicating that), T1
If it is less than this, the image density signal d of the pixel to be processed is regarded as an invalid (line thin signal) detection signal h _{i-1, j-1.}
It is input to the dot delay circuit 6 in order to match the timing of density conversion for _{i-1 and j-5} . The dot delay circuit 6 outputs thin line detection signals h _{i-1, j-5} ,
It is input to the characteristic determination circuit 12.

FIG. 13 is a diagram showing an edge detection circuit in the device of this embodiment. Similar to the first embodiment, the arithmetic circuit 13 uses the edge detection coefficient matrix in the horizontal direction, the vertical direction, the left diagonal direction, and the right diagonal direction shown in FIGS. The calculation of the volume is performed, and the largest of the absolute values of the four obtained values is output as the edge signal e _{i-1, j-1.}
5 is input.

In the comparator 15, the input edge signal e
_A first edge detection signal e1 that is valid (a signal indicating that the pixel is an edge pixel) and is invalid (a signal indicating that the pixel is not an edge pixel) if _{i-1, j-1} is greater than or equal to the first edge threshold T2, and is less than T2.
_{i-1, j-1} are output. The first edge threshold T2 is set to a value capable of detecting only a sharp edge (having a large density difference) that hardly appears in the gradation image.

In the comparator 14, the second edge detection signal e2 _i-1.j-1 which is valid when the edge signal e _i-1.j-1 is equal to or more than the second edge threshold T3 and is invalid when the edge signal is less than T3. Is output. The second edge threshold value T3 is smaller than the first edge threshold value T2, and is set to a value at which a slightly sharp edge (with a slightly large difference in density) can be detected.

FIG. 14 shows a high density line detection circuit. The high density line detection circuit is a circuit used for recognizing lines of preset thickness and density, and whether the target pixel is sandwiched between a rising edge pixel and a falling edge pixel, and whether the density is high. This is a circuit for detecting whether or not.

The first edge detection signal e1 _i-1.j-1 is inputted to the dot delay circuit 16 and the logical sum circuit 17 which are composed of four pieces. At the same time, the outputs of the dot delay circuit 16 e1 _i-1.j-2 , e1 _i-1.j-3 , e1 _i-1.j-4 , e1 _i-1.j-5
Is input to the logical sum circuit 17, and the leading pixel edge signal m
It is output as _i-1.j-5 .

The comparator 19 outputs a high density detection signal c4 _{i-1, j-5 which} is valid if the input image density signal d _{i-1, j-5} is greater than or equal to the threshold value T4, and invalid if less than T4. To be done. The output e1 of the dot delay circuit 16 is supplied to the OR circuit 21.
_{i-1, j-5} and the high density line detection signal _{m'i-1, j-5} are input.
Edge or high density line signal bm from the OR circuit 21
_{i-1, j-5} are output and input to the dot delay circuit 20 to detect the edge of the previous pixel or the high density line detection signal bm _i-1.j-6
Is output.

The preceding pixel edge signal m _{i-1, j-5} , the high density detection signal c4 _{i-1, j-5} , and the edge or high density line detection signal bm _{i-1, j-6} of the previous pixel are the logical product. It is input to the circuit 18. The high density line detection signals m ′ _{i-1, j-5} are output from the logical product circuit 18 and input to the logical sum circuit 21 and the feature determination circuit 12.

FIG. 15 shows the details of the distance classification circuit 11 which counts and classifies the distance from the pixel of interest to the edge pixel.
The line memory 23 stores the distance data from the edge pixel of each pixel for one line.

FIG. 16 shows the position of the pixel on the original corresponding to the distance data to the edge pixel of each pixel stored in the line memory 23. First edge detection signal el _{i-1, j-5}
Is input to a distance counting circuit 22, the pixel p _i-1, the distance de _i-1 from the edge of the _{j-5, j-5} is left pixel _{p i-1, j-6} , the upper pixel p _{i- 2, j-5} , the distance from the edge of the upper right pixel p _{i-2, j-4} , de _{i-1, j-6} , de _{i-2, j-5} , de, respectively
_{i-2, j-4} are used to count as follows.

If e1 _{i-1, j-5} = 1 then de _{i-1, j-5} = 0 If e1 _{i-1, j-5} = 0 then de _{i-1, j-5} = min ( de _{i-1, j-6} , de _{i-2, j-5} , d
e _{i-2, j-4} ) +1

The above de _{i-1, j-6} , de _{i-2, j-5} , de
_{i-2 and j-4} are fetched from the line memory 23 and de
_{i-1 and j-5} are newly stored in the line memory 23. The distance de _{i-1, j-5} from the edge is input to the comparator 24. If de _{i-1, j-5} <T6, then f _{i-1, j-5} = 0 T6 ≤ de _{i-1 , j-5} <T7, then f _{i-1, j-5} = 1 T7 ≦ de _{i-1, j-5} <T8, then f _{i-1, j-5} = 2 T8 ≦ de _{i-1 , j-5} , the classification is made according to the rule that f _{i-1, j-5} = 3. Distance classification signal f
_{i-1, j-5} are input to the feature determination circuit 12.

In the characteristic judging circuit 12, the fine line detection signals hi-1, j-5, the first edge detection signals e1 _{i-1, j-5} and the second fine line detection signals inputted are inputted.
Edge detection signal e2 _{i-1, j-5} , high density line detection signal m '
_The density conversion selection signals g _{i-1, j-5} are determined and output from _{i-1, j-5} and the distance classification signal f _{i-1, j-5 according} to the rule shown in FIG.

The density conversion selection signal g _i-1.j-5 is 1. Fine line pixels and edge pixels mainly exist in the binary image area. The degree of edge sharpness (density difference) is strong (large)
The higher the degree of binary image, the weaker (smaller) the edge sharpness is. 3. If there is a line segment that sandwiches the pixel of interest between the rising edge pixel and the falling edge pixel of the density in the main scanning direction or the sub-scanning direction, if the density of the line is high and the line is thin, the line forming the character (one of the binary image Part) is likely. 4. The shorter the distance between the pixel of interest and the edge pixel closest to the pixel of interest, the more likely the image is a binary image, and the longer the distance, the more likely the image is a gradation image. The density conversion curve is selected in consideration of image characteristics such as.

In FIG. 17, ∘ indicates that the binary signal is valid, × indicates that it is invalid, and − indicates that it is ignored. For signals with three or more values, the numbers represent the signal values. Here, the value of the density conversion selection signal g _{i-1, j-5} is the binary image if g _{i-1, j-5} = 0, and the gradation if the value is g _{i-1, j-5} = 3. Image g _{i-1, j-5} = 1 means a binary image, and g _{i-1, j-5} = 2 means a gradation image.

As shown in FIG. 7, the density conversion circuit 3 binarizes the input density to a maximum or minimum density conversion curve A for binary image (shown by a solid line) and the input density Density conversion curve B for gradation image that stores gradation in output density
(Indicated by a one-dot chain line), a density conversion curve C (indicated by a dotted line) from a binary image density conversion curve that interpolates the above two curves, and a density from a gradation image density conversion curve It has a total of four density conversion curves of two density conversion curves consisting of the conversion curve D (shown by a two-dot chain line).

If g _{i-1, j-5} = 0 according to the density conversion selection signals g _{i-1, j-5} input from the feature determination circuit 12, the density conversion curve for binary image A g _{i- If 1, j-5} = 3, gradation density conversion curve B g _{i-1, j-5} = 1, density conversion curve C interpolating two curves C (from binary image density conversion curve) g _{i -1, j-5} = 2, the density conversion curve is selected according to the rule of the density conversion curve D (from the gradation image density conversion curve) that interpolates the two curves, and the image density signal d
The density conversion of _{i-1, j-5} is performed and the converted density signal cd _{i-1, j-5} is output.

The density conversion circuit 3 is similar to that of the first embodiment. In this embodiment, the fine line detection coefficient matrix and the edge detection coefficient matrix are set to 3 × 3. The size of the matrix is changed to n × m (n> 0, m> 0: n, m is an integer), or the coefficient is changed. Can be changed or the number of matrices can be increased to improve the accuracy of edge detection by the edge detection circuit 7. Further, the thickness of the line to be detected can be increased by increasing the number of delay stages of the dot delay circuit of the high density line detection circuit.

Further, although there are two density conversion curves for interpolating the density conversion curve for binary image and the density conversion curve for gradation image, there may be one density conversion curve, and each edge detection circuit 7 may have one density conversion curve. By increasing the number of comparators to which different thresholds are set, the degree of sharpness of the edge is increased, and by increasing the number of classifications of the distance classification circuit 11, the density conversion curve that interpolates both curves is expanded to three or more. You can also do it.

[0069]

According to the apparatus of the present invention, a sense of discomfort caused by a sudden change in the density of an output image due to erroneous discrimination of areas can be reduced. Even in a document in which character images such as characters and thin lines and gradation images such as photographs are mixed, an image density signal faithful to the document can be output.

Further, by having density conversion information corresponding to a plurality of different density conversion curves for interpolating both the density conversion curve for binary image and the density conversion curve for gradation image, thick characters and solid parts can be obtained. Even if there is, the edge part can be made denser,
It can be black in appearance. Further, as the distance from the edge portion is increased, the density can be converted to have a gradation, so that there is an effect that it is possible to preserve the gradation of a high density portion of the gradation image.

[Brief description of drawings]

FIG. 1 is a block diagram of an image processing apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram showing a coefficient matrix of an MTF correction circuit used in an embodiment of the present invention.

FIG. 3 is a diagram showing positions of pixels on a document in a line memory which stores an image density signal according to an embodiment of the present invention.

FIG. 4 is a diagram showing an edge detection circuit according to an embodiment of the present invention.

FIG. 5 is a diagram showing a coefficient matrix of the edge detection circuit in the embodiment of the present invention.

FIG. 6 is a chart showing a feature determination rule in the embodiment of the present invention.

FIG. 7 is a diagram showing an example of a density conversion curve of a density conversion circuit according to the embodiment of the present invention.

FIG. 8 is a block diagram of an image processing apparatus according to an embodiment of the present invention.

FIG. 9 is a diagram showing an edge detection circuit according to an embodiment of the present invention.

FIG. 10 is a diagram showing a distance classification circuit in the embodiment of the present invention.

FIG. 11 is a table showing characteristic determination rules according to the embodiment of the present invention.

FIG. 12 is a block diagram of an image processing apparatus according to an embodiment of the present invention.

FIG. 13 is a diagram showing an edge detection circuit according to an embodiment of the present invention.

FIG. 14 is a diagram showing a high density line detection circuit according to an embodiment of the present invention.

FIG. 15 is a diagram showing a distance classification circuit in the embodiment of the present invention.

FIG. 16 is a line memory 23 according to the embodiment of the present invention.
The position of the pixel on the original corresponding to the distance data to the edge pixel of each pixel stored in FIG.

FIG. 17 is a table showing characteristic determination rules according to the embodiment of the present invention.

FIG. 18 is a diagram showing a coefficient matrix of the thin line detection circuit used in the embodiment of the present invention.

[Explanation of symbols]

 1 MTF correction circuit 2 Line memory 3 Density conversion circuit 4 Binarization circuit 5 Fine line detection circuit 6 Dot delay circuit 7 Edge detection circuit 8 Dot delay circuit 9 Dot delay circuit 10 High density line detection circuit 11 Distance classification circuit 12 feature determination circuit 13 arithmetic circuit 14 comparator 15 comparator 16 dot delay circuit 17 logical sum circuit 18 logical product circuit 19 comparator 20 dot delay circuit 21 logical sum circuit 22 distance counting circuit 23 line memory 24 comparator 25 Comparator 26 Dot delay circuit 27 Edge detection circuit 28 Feature determination circuit 29 Edge detection circuit 30 Feature determination circuit

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical indication H04N 1/40 103 Z

Claims (6)

[Claims] 1. An image processing apparatus for discriminating a document image input by a document reading unit into a binary image and a gradation image and outputting an image signal converted by a corresponding density conversion curve, from the binary image. Means for judging the state of the original pixel of interest between gradation images in multiple stages, and a density conversion curve for a binary image, a density conversion curve for a gradation image, and one or more for interpolating the two curves according to the judgment. An image processing apparatus comprising a density conversion curve selecting means for selecting an appropriate density conversion curve from the density conversion curves and converting the density. 2. A means for judging the state of a target pixel of a document in multiple stages between the binary image and the gradation image is based on information corresponding to the sharpness of the edge of the target pixel according to the densities of neighboring pixels. The image processing apparatus according to claim 1, wherein the image processing apparatus is a determination unit based on the degree of binary image likelihood obtained. 3. The means for judging the state of the original pixel of interest in multiple stages between the binary image and the gradation image detects whether or not the target pixel is an edge pixel according to the density of neighboring pixels. The edge pixel detecting means is a judging means based on the degree of binary image likeness obtained based on information according to the distance between the target pixel and the edge pixel detected by the edge pixel detecting means closest to the target pixel. The image processing apparatus according to claim 1, wherein 4. The means for judging the state of the original pixel of interest in multiple stages between the binary image and the gradation image detects whether or not the target pixel is a thin line pixel according to the densities of neighboring pixels. The thin line pixel detection means is provided, and the edge pixel detection means for detecting whether or not the target pixel is an edge pixel according to the density of the neighboring pixel is included. It has a means for calculating the degree of sharpness of the edge, and if there is a line sandwiching the pixel of interest between the rising edge and the falling edge according to the detection result detected by the edge pixel detecting means, determine the thickness of the line. The thin line pixel has a line thickness counting unit for counting, and a distance counting unit for counting a distance between the target pixel and an edge pixel closest to the target pixel detected by the edge pixel detecting unit. Whether or not there is 2. The image processing apparatus according to claim 1, wherein the image processing apparatus is a determination unit that is obtained based on information according to the sharpness of the edge, the line thickness, and the distance as to whether the pixel is a dipixel. 5. An image processing apparatus for discriminating an original image input by an original reading unit into a binary image and a gradation image, and outputting an image signal converted by a corresponding density conversion curve, from the binary image. Means for judging the state of the original pixel of interest between gradation images in multiple stages, and a density conversion curve for a binary image, a density conversion curve for a gradation image, and one or more for interpolating the two curves according to the judgment. A density conversion curve selecting unit for selecting an appropriate density conversion curve from the density conversion curves and converting the density, and a binarization unit for binarizing the image density signal converted by the density conversion unit are provided. Image processing device. 6. The image processing apparatus according to claim 5, wherein the binarization unit is an error diffusion binarization processing unit.