JP2002084422A - Image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processor that enhances the reproducibility of an edge part in the case of applying binary processing to a multi-value image signal. SOLUTION: A dither matrix circuit 1 adopts a dither matrix method to apply binary processing to a received multi-value image signal. Furthermore, an edge extract circuit 2 extracts the edge part from the received multi-value image signal and provides an output of a binary image signal consisting only of the edge parts to an arithmetic circuit 3. The arithmetic circuit 3 synthesizes the image signal binarized by the dither matrix circuit 1 with the binary image signal whose edge parts are extracted by the edge extract circuit 2 through a logical arithmetic operation. Thus, the image of the edge parts extracted by the edge extract circuit 2 corrects the image quality of the edge parts deteriorated by the dither matrix method so as to enhance the image quality of characters or the like.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for binarizing a multilevel image signal.

[0002]

2. Description of the Related Art There are two types of binarization methods for pseudo-representing a halftone image as a binary image: a dither matrix method and an error diffusion method. Error diffusion method is 2
The values are set to, for example, the minimum density and the maximum density, and the multi-valued image signal is compared with a predetermined threshold and replaced with one of two values. At this time, a density error occurring between the replaced minimum density or maximum density and the pixel density is obtained, added to the density of surrounding pixels, and compared with a threshold value. In this manner, binarization is performed by transmitting the density error to the surrounding pixels. For example, when binarizing a multi-valued image in a density range of 0 (black) to 10 (white) with a threshold value 5, an image having a uniform density 6
First, since the density is 5 or more, the value after binarization is set to a value 0 (white) corresponding to the density 10. At this time, density 10 and density 6
Is propagated to surrounding pixels. For example, if 100% is propagated to the next pixel, the density of the next pixel becomes 6
-4 = 2, and the value after binarization is the value 1 corresponding to the density 0
(Black). Since the error at this time is 2, 2 is added to the density 6 of the next pixel, and the result is compared with the threshold to obtain 2.
The value after the value conversion is 0 (white). In this way, by sequentially transmitting the density error to the surrounding pixels, the multi-valued image can be pseudo-expressed by binary dots of 1 or 0.

In such an error diffusion method, even when an error is propagated in a high density portion and a low density portion, the probability of becoming "0" or "1" by binarization is high. Therefore, in a high density part or a low density part, the dots of “0” or “1” tend to be binarized so as to concentrate. Due to such a tendency, even when, for example, a black character is included in an image on a white background, the image is relatively well reproduced. However, there is a problem in that the circuit configuration is complicated due to error propagation and the cost is high.

On the other hand, the dither matrix method is a method of sequentially comparing thresholds arranged in a matrix and multi-valued image signals. FIG. 8 is an explanatory diagram of an example of the dither matrix. For example, the density range is set to 0 to 10, and 1 to 9 are arranged as threshold values in a 3 × 3 matrix as shown in FIG. Then, 3 × of the input image signal
The three pixels are compared with the threshold at the corresponding position in the dither matrix. When the pixel whose density value of the image signal is equal to or smaller than the threshold value is “1”, the uniform image having the density 5 is 3 × 3 pixels in FIG.
5 is "1", and the other pixels are "0". By binarizing the image signal in this way, a halftone can be reproduced in a pseudo manner.

In such a dither matrix system, only the image signal is compared with the threshold value of the dither matrix to obtain a 2
Since the value can be converted, the circuit configuration is simple, and the cost can be reduced. However, since the dots are arranged according to the shape in which the thresholds are arranged (for example, a Bayer type or a spiral type), there is a problem that the dots are dispersed even at an edge portion of, for example, a character and cannot be reproduced well.

FIG. 9 is a block diagram showing an example of a conventional image processing apparatus. In the figure, 11 is an edge enhancement circuit, 12
Is a dither matrix circuit. The dither matrix circuit 12 is a circuit that performs a binarization process according to the above-described dither matrix method. As described above, since the dots are dispersed in the edge portion of a character or the like, the edge portion is blurred. Therefore, conventionally, an edge enhancement circuit 11 is provided before the dither matrix circuit 12 to enhance the edge of the multi-valued image signal,
Has been configured to perform binarization by using.

However, even if the image signal has been subjected to the edge enhancement processing by the edge enhancement circuit 11, deterioration of the edge portion is inevitable due to the characteristics of the dither matrix system, and the image quality is deteriorated. there were.

[0008]

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and provides an image processing apparatus which has improved reproducibility of an edge portion when binarizing a multi-valued image signal. It is intended to provide.

[0009]

According to the present invention, there is provided an image processing apparatus for binarizing a multi-valued image signal, comprising: a dither matrix means for binarizing the image signal by a dither matrix method; An edge extracting means for extracting, and an arithmetic means for performing a logical operation on an output signal of the dither matrix means and an output signal of the edge extracting means.

[0010] In the binarization processing by the dither matrix means, the reproducibility at the edge portion is poor as described above. However, in the present invention, an image of only the edge portion is generated by the edge extracting means, and the image of the edge portion and the image after binarization by the dither matrix method are logically operated by the arithmetic means and combined. As a result, the reproducibility of the edge portion is improved, and a binarized image signal with good image quality can be obtained.

As the edge extracting means, a Laplacian filter or a local average filter can be used.

[0012]

FIG. 1 is a block diagram showing an embodiment of an image processing apparatus according to the present invention. In the figure, 1 is a dither matrix circuit, 2 is an edge extraction circuit, and 3 is an arithmetic circuit. The dither matrix circuit 1 has a threshold matrix of a predetermined size, and binarizes the input multi-level image signal by comparing the input multi-level image signal with a threshold at a corresponding position in the threshold matrix. The binarized image signal is passed to the arithmetic circuit 3.

The edge extraction circuit 2 extracts an edge portion from the input multi-valued image signal, and outputs a binary image signal consisting of only the edge portion to the arithmetic circuit 3. As the edge extraction circuit 2, a Laplacian filter, a local average filter, or the like can be used as described later. In particular, by extracting as an edge only a portion having a large tone difference, it is possible to extract only an edge of a portion where a contour of a character or the like is desired to be clearly reproduced, and to reduce an influence on a portion having a light gradation change.

The arithmetic circuit 3 performs a logical operation on the image signal binarized by the dither matrix circuit 1 and the binary image signal from which the edge is extracted by the edge extracting circuit 2. For example, when the edge portion is extracted as “1” by the edge extraction circuit 2, a logical sum operation may be performed on the image signal binarized by the dither matrix circuit 1. In addition, for example, when the edge portion is extracted as “0” by the edge extraction circuit 2, the AND operation may be performed on the image signal binarized by the dither matrix circuit 1.

In this way, the arithmetic circuit 3 combines the image signal binarized by the dither matrix circuit 1 and the binary image signal whose edges have been extracted by the edge extraction circuit 2. By this synthesizing process, an edge portion that has deteriorated when binarized by the dither matrix circuit 1 can be corrected with a good edge image extracted by the edge extraction circuit 2, and a good edge image can be obtained. it can. Of course, the halftone portion is binarized by the dither matrix circuit 1, so that the halftone can be reproduced in a pseudo manner.

The logical operation performed by the arithmetic circuit 3 is based on the value of each pixel when binarized by the dither matrix circuit 1 (for example, whether black is 1 and white is 0, or whether white is 1 and black is 0). ), And the logic of the edge portion extracted by the edge extraction circuit 2 (whether the extracted edge is the black side or the white side, and whether the edge portion is 0 or 1), and the like.
In any case, the edge portion of the image binarized by the dither matrix circuit 1 may be combined with the edge image extracted by the edge extraction circuit 2.

The above-described edge extraction circuit 2 will be further described. As the edge extraction circuit 2, for example, a Laplacian filter can be used. FIG. 2 is an explanatory diagram for a Laplacian filter. Each rectangle in the figure indicates a pixel. The Laplacian filter is a filter for obtaining a change rate (derivative) of an image signal. Now, pixel p
Referring to pixels a to d (density Da to Dd) adjacent to the upper, lower, left and right of (density Dp), the emphasis coefficient in the horizontal (main scanning) direction is x, and the emphasis coefficient in the vertical (sub scanning) direction is y. In this case, the processing result Fp of the Laplacian filter is given by: Fp = x · Da + y · Db + x · Dc + y · Dd− (2
x + 2y) · Dp. Simply, x = y = 1 and Fp = Da + Db + Dc + Dd-4Dp.

FIG. 3 is an explanatory diagram of an example of processing an image signal by a Laplacian filter. Here, it is assumed that the input image signal is an image signal read by a scanner or the like, and a larger value is white and a smaller value is black. Further, as the image signal after binarization, white is output as “0” and black is output as “1”.

Now, consider the case where the input image signal changes from white to black as shown in FIG. When there is such a density change, the Laplacian filter outputs a signal corresponding to the change rate of the input image signal as shown in FIG. 3B. By binarizing the output of such a Laplacian filter with a predetermined threshold value S, an image signal of an edge portion is obtained as shown in FIG.

FIG. 3 shows a case where the density change is steep. When the density change is gradual, the rate of change obtained by the Laplacian filter becomes small. Therefore, when binarized by the threshold value S, it cannot be obtained as an edge. Thus, by appropriately setting the threshold,
For example, as in the case where a character is drawn on a white background, it is possible to extract only an edge portion where a sharp change in density exists.

In this example, the output of the Laplacian filter is binarized using a positive threshold value, and the edge on the black side is extracted. However, the present invention is not limited to this, and binarization can be performed using a negative threshold to extract a white-side edge. In this case, the edge portion is set to "0", and the other portions are set to "1".
The arithmetic circuit 3 may calculate a logical product.

Further, in FIG. 3, the larger the value of the input image signal is white, the smaller the value is black. However, the present invention is not limited to this, and the larger the value of the input image signal is black, The smaller one may be white. In this case, by inverting the sign of the output of the Laplacian filter, edges can be extracted in the same manner as described above. Of course, the value after binarization may be black at “0” and white at “1”.

The edge extraction circuit 2 can extract an edge by applying a local average filter. FIG. 4 is an explanatory diagram for a local average filter. Each rectangle in the figure indicates a pixel. The local average filter is for calculating the average of the pixel values in a region of a predetermined size, for example, 3 × 3 in FIG. In the example shown in FIG.
Da, Db, Dc, Dd, De, D
When f, Dg, Dh, and Di, the local average value Save is: Save = (Da + Db + Dc + Dd + De + Df + Dg
+ Dh + Di) / 9. A value obtained by subtracting a predetermined value α from the local average value Save, that is, S = Save−α is used as the threshold value S, and the input image signal is
Value.

The threshold value S is a value that changes as the pixel value in the area changes. However, since the average of the pixel values in the region is calculated, if there is a sudden change in the pixel value, the threshold value S shows a behavior in which the change is averaged. It will be reversed at the edge. By this operation, an edge part can be extracted.

FIG. 5 is an explanatory diagram of an example of processing an image signal when a local average filter is used. Here, as in FIG. 3, it is assumed that the input image signal is an image signal read by a scanner or the like, and a larger value is white and a smaller value is black. Further, as the image signal after binarization, white is output as “0” and black is output as “1”.

Now, consider the case where the input image signal changes from white to black as shown by the solid line in FIG. When there is such a density change, the threshold value S obtained by subtracting the predetermined value α from the local average Save obtained by the local average filter changes as shown by a broken line in FIG. At this time, if a portion where the input image signal falls below the threshold value S is “1”, an edge image in which only the black edge portion is “1” as shown in FIG. 5B is obtained. Can be.

FIG. 5 shows a case where the density change is steep. When the density change is gradual, the change in the threshold value S follows the change in the density of the input image, so that the input image signal does not fall below the threshold value S. How much density change is extracted as an edge can be adjusted by the value of the predetermined value α subtracted from the local average Save.

Further, in this example, a black side edge is extracted by binarizing an input image signal using a threshold value S obtained by subtracting a predetermined value α from the local average Save. In this case, the arithmetic circuit 3 at the subsequent stage performs a logical sum operation. However, the present invention is not limited to this, and a white-side edge can be extracted by adding the value of the predetermined value α to the local average Save to obtain the threshold value S and binarizing the input image signal. In this case, it is preferable to extract the edge portion as “0” and calculate the logical product in the arithmetic circuit 3.

In FIG. 5, the input image signal having a larger value is white and the smaller value is black. However, the present invention is not limited to this. The input image signal having a larger value is black. The smaller one may be white. In this case, the addition and subtraction of the predetermined value α from the local average Save may be reversed. Of course, the value after binarization can also be black at "0" and white at "1".

FIG. 6 is an explanatory diagram of a specific example of the processing steps in an embodiment of the image processing apparatus of the present invention. Now
Consider a case where a vertical edge exists as shown in FIG. By reading the image shown in FIG. 6A with a scanner or the like, an image signal having a small value in a solid portion and a large value in a white portion is obtained. Such an image signal is input to the dither matrix circuit 1 and the edge extraction circuit 2.

The dither matrix circuit 1 binarizes the input image signal according to a predetermined threshold matrix.
Thereby, for example, an image as shown in FIG. 6B is obtained. Here, it is assumed that the pixel painted black is “1” and the image extracted white is “0”. As can be seen from FIG. 6B, white pixels and black pixels are mixed in the edge portion. When the resolution is high, it is difficult to visually check that such white pixels and black pixels coexist, but the edge is visually recognized as blurred.

On the other hand, in the edge extraction circuit 2, for example, the output of the Laplacian filter is
The input image signal is binarized using a value obtained by adding or subtracting (in this case, subtracting) a predetermined value from the output of the local average filter as a threshold value. As a result, FIG.
As shown in (C), an edge image having an edge portion of “1” is obtained.

The arithmetic circuit 3 calculates and outputs the logical sum of the image output from the dither matrix circuit 1 as shown in FIG. 6B and the image output from the edge extraction circuit 2. As a result, as shown in FIG. 6D, the edge part whose image quality has deteriorated due to the binarization processing by the dither matrix circuit 1 is restored by the edge image extracted by the edge extraction circuit 2, and the image whose edge part is improved is obtained. Can be obtained.

FIG. 7 is an explanatory diagram of another specific example of the process in the embodiment of the image processing apparatus of the present invention. As in the example shown in FIG.
Consider an image having a vertical edge as shown in FIG. It is also assumed that an image as shown in FIG. 7B is obtained by binarization by the dither matrix circuit 1.

In the example shown in FIG. 7, the edge extraction circuit 2 extracts a white side edge. In each of the above-described methods, the threshold value is set and the binary logic to be acquired is set. As shown in C), an edge image in which the extracted edge portion is “0” and the other portions are “1” is obtained. This edge image is input to the arithmetic circuit 3 together with the image output from the dither matrix circuit 1 shown in FIG. The arithmetic circuit 3 performs an AND operation of the two to obtain an image as shown in FIG. In this image, the black pixels scattered on the white background when binarized by the dither matrix circuit 1 become white, and the image quality of the edge portion can be improved as in the example shown in FIG.

In the example shown in FIG. 7, the white image portion is set to "0" and the other portions are set to "1" as the edge image. However, the present invention is not limited to this. For example, the white edge portion is set to "1". The same result can be obtained by calculating exclusive OR with the arithmetic circuit 3 by setting the portion to "0".

Of course, both the OR operation using the black side edge image as illustrated in FIG. 6 and the AND operation using the white side edge image as illustrated in FIG. 7 are performed. May be configured. Thereby, the image quality of the edge portion can be further improved.

[0038]

As is apparent from the above description, according to the present invention, the output signal of the dither matrix means and the output signal of the edge extraction means are synthesized by a logical operation.
Outlines of characters and the like are reproduced well. In addition, there is an effect that a high-quality binary image in which characters and the like are well reproduced can be obtained by a low-cost image processing device which can be configured with a simple circuit.

[Brief description of the drawings]

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

FIG. 2 is an explanatory diagram for a Laplacian filter.

FIG. 3 is an explanatory diagram of an example of processing an image signal by a Laplacian filter.

FIG. 4 is an explanatory diagram for a local average filter.

FIG. 5 is an explanatory diagram of a processing example of an image signal when a local average filter is used.

FIG. 6 is an explanatory diagram of a specific example of a processing process in an embodiment of the image processing apparatus of the present invention.

FIG. 7 is an explanatory diagram of another specific example of a process in the embodiment of the image processing apparatus of the present invention.

FIG. 8 is an explanatory diagram of an example of a dither matrix.

FIG. 9 is a block diagram illustrating an example of a conventional image processing apparatus.

[Explanation of symbols]

1. Dither matrix circuit, 2. Edge extraction circuit, 3.
An arithmetic circuit, 11 is an edge enhancement circuit, and 12 is a dither matrix circuit.

 ──────────────────────────────────────────────────続 き Continuing on the front page F term (reference) 2C262 AA24 AA26 AA27 AB13 BB01 BB06 DA03 GA23 5B057 CA08 CA12 CA16 CB07 CB12 CB16 CC01 CE06 CE08 CE13 CH09 CH20 DB02 DB09 DC16 5C077 LL17 LL17 LL19 NN08 PP01 PP47 PQ12 PQ18

Claims (3)

[Claims] 1. An image processing apparatus for binarizing a multi-valued image signal, comprising: dither matrix means for binarizing the image signal by a dither matrix method; edge extracting means for extracting an edge from the image signal; An image processing apparatus comprising: an operation unit that performs a logical operation on an output signal of a dither matrix unit and an output signal of the edge extraction unit. 2. The image processing apparatus according to claim 1, wherein said edge extracting means is a Laplacian filter. 3. An image processing apparatus according to claim 1, wherein said edge extracting means is a local average filter.