JP2644491B2 - Image processing device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus that binarizes image data representing an image by an error diffusion method.

<Prior Art> Conventionally, there is an error diffusion method as a binarization method for reproducing a halftone in a digital printer, a digital facsimile, or the like. This method is a method of calculating the density difference between the image density of the original document and the output image density for each pixel, and dispersing the error resulting from the calculation after giving a specific weight to the peripheral image. This error diffusion method is described in reference R.
W. Floyd and L. Steinberg “An Adaptive Algorithm for
Presented in “Speatial Gray Scale” SID 75 Digest (1976).

Since this method has no periodicity, the moiré phenomenon, which is a problem in the other dithering methods such as the dithering method and the density pattern method, does not occur, but it is unique in the uniform density portion of the image (the portion other than the edge portion). There is a drawback in that a stripe pattern is generated, and in particular, in a highlight portion of an image, dots are dispersed, so that granular noise is conspicuous, which causes deterioration in image quality.

<Objective> The present invention is to eliminate the above-mentioned disadvantages of the prior art, and performs a halftone process on a portion other than an edge portion of an image to form a pseudo halftone dot, and then performs a binarization process by an error diffusion method. The purpose of the present invention is to provide an image processing apparatus capable of suppressing the generation of a unique stripe pattern generated when the binarization processing is performed by the error diffusion method.

<Example> Hereinafter, an example of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing one embodiment of the present invention.
The input sensor unit 11 includes a photoelectric conversion element such as a CCD and a driving device that scans the photoelectric conversion element, and performs scanning of a document. Image data of the document read by the input sensor unit 11 is sent to the A / D converter 12. Here, the data of each pixel is converted into 8-bit digital data and quantized into data having 256 levels of gradation.

Next, in the correction circuit 13, shading correction for correcting unevenness in sensitivity of the CCD sensor and unevenness in illuminance caused by the illumination light source is performed by digital arithmetic processing. Next, the corrected signal 100 is input to the edge detection circuit 14 and the selector 15.

At this time, the data is transferred by block transfer with m × n pixels as one block. In this embodiment, m = 3 and n = 3.
The case will be described.

The edge detection circuit 14 determines whether an edge exists in the block. 3 × 3
Minimum value in a method or block performed by the Laplacian, and calculates the difference between the maximum value and compare the value threshold T ₁ (here, T ₁ = 15) and, edge exists if above T _1, T ₁ If it is smaller, there is a method of determining that the edge does not exist, and the like, but in the present embodiment, the latter will be described. Incidentally, T ₁ is determined experimentally. Here, it is determined whether or not an edge exists in the block in order to avoid a decrease in resolution that occurs when the block in which the edge exists is halftoned. The edge detection circuit 14 determines the presence of the edge, and outputs “1” if the edge exists, and outputs “0” as the signal 200 if the edge does not exist. The signal 200 is input to the selector 15. If the signal 200 is "0", the signal 100 input to the selector 15 is transferred to the preprocessing circuit 16 as a signal 300 for each block. If the signal 200 is “1”, the signal 100 is changed to the signal 40
It is transferred to the memory 17 for each block as 0.

Here, the selector 15 is provided so as to perform the pre-processing only on the portion having no edge determined by the edge detection circuit 14.

The preprocessing circuit 16 performs pseudo halftone dot conversion by taking the total of the block data input as the signal 300 and replacing the total of the density data in the block with the density of one or more pixels in the block. In other words, by forming a pseudo halftone dot in a uniform density portion of an image determined not to be an edge portion, dots can be arranged periodically, and a unique dot that occurs when binarization is performed by an error diffusion method or the like. Stripe pattern can be reduced. Data 500 output from the preprocessing circuit 16
Is input to the memory 17. Data 600 read out from the memory 17 in pixel units enters a binarization circuit 18 where binarization is performed. The result is input to the printer 19 as a signal 700, and is output as an image by the printer 19. FIG. 2 is a block diagram of the pre-processing circuit 16. The signal 300 output from the selector 15 enters the sum calculation circuit 21, where the sum S of the densities of nine data (3 × 3 pixels) in the block is calculated by the following equation.

Here, D _ij is the density of the (i, j) pixel in the block. FIG. 3 is a block diagram of the summation circuit 21.
The total sum S of the densities is input to the halftoning circuit 22 as a signal 301.

FIG. 4 is a diagram showing density data in a block (3 × 3 pixels). D _ij (i = 1,2,3, j = 1,2,3) is (i,
j) Represents the pixel density.

FIG. 5 is a diagram showing the density after the halftone processing. A _ij
(I = 1, 2, 3, j = 1, 2, 3) represents the density of the (i, j) pixel.

Here, assuming that the maximum density that can be expressed by the printer is D _max , in the halftoning circuit 22, i) when S ≦ D _max A ₂₂ = S other pixel densities = 0 ii) A when D _max <S ≦ 5D _{max 22} = D _max A ₁₂ = A ₂₁ = A ₂₃ = A ₃₂ = (S-D _max ) / 4 Other pixel density = 0 iii) When S <5D _max A ₁₂ = A ₂₁ = A ₂₂ = A ₂₃ = A halftone dot is formed such that A ₃₂ = D _{max and} other pixel densities = (S−5D _max ) / 4. Here, S is the sum of the densities output from the sum calculation circuit 21 in FIG.
_ij (i = 1,2,3, j = 1,2,3) represents the density of the (i, j) pixel in the block after the pseudo halftoning process.

With the above configuration, pseudo halftone dots can be formed. By creating this periodic structure before binarization, a unique pattern generated when binarization is performed by an error diffusion method or the like can be reduced. When the halftoning process is completed, the data is input to the memory 17 as a signal 500.

FIG. 6-1 is a block diagram showing details of the binarization circuit 18. The image data 600 (X
_ij) is the error epsilon _ij stored in the error Batu Hua memory 53, the weighting generator 52 multiplied by a weighting factor alpha _kl, are added in normalized (divided by Σα _kl) the data adder 51.

 This is represented by the following equation.

FIG. 6-2 shows an example of the weighting coefficient. Numeral 57 indicates a pixel position currently being processed, and the closer the pixel position is, the larger the value in the matrix. This is a weight generator
At 52, the data a of the error buffer memory 53 is multiplied by 1/48, b and 3/48, c and 5/48... L and 7/48 in FIG. 6-2, and the sum data is added. This is because weighting is performed using data in the error buffer memory 60 which is sent to the pixel position 62 and is close to the pixel position 66 currently being processed.

The error ε stored in the error buffer memory 53
_ij stores the difference data between the correction data x ′ _ij added by the adder 51 before the data currently being processed and the binarized output data y _ij .

Next, the correction data x ' _ij added by the adder 51 is compared with a threshold value T by a binarization circuit to output data y _ij .
Here, y _ij is binarized data such as y _max or y _max (for example, 255 and 0).

On the other hand, in the arithmetic unit 55, the correction data x ′ _ij and the output data y _ij
Is calculated, and the result is stored in the error buffer memory 53.
At the memory location corresponding to the pixel position 57 currently being processed. Next, the same processing as described above is performed on the image data.
In this case, the error ε _ij of the error buffer memory 60 is shifted by one to the right. By repeating this operation sequentially, the binarization process of the error diffusion method is executed.

In this way, by detecting the edges of the image and converting the uniform density portions other than the edge portions into halftone dots, the dots can be aligned, so that a unique pattern generated during the error diffusion method can be reduced. it can.

It should be noted that if the edge detection circuit 14 in FIG. 1 is changed to the highlight part detection circuit 20 in FIG. 7, only the highlight part can be converted into a halftone dot. The highlight part detection circuit 20 determines that the maximum value of the density of nine pixels (3 × 3 pixels) in the block is equal to the threshold value T ₃ (for example, T ₃ =
20) In the following cases, it is configured to be a highlight part. Then, by highlighting the highlighted portion as described above and adding regularity, dispersed dots can be aligned in a short period, so that regularity without noise can be created and highlighting can be performed. It is possible to prevent the occurrence of granular noise felt in the section.

In the above-described embodiment, the pseudo halftoning process is performed based on whether the edge portion or the highlight portion exists. In the embodiment described below, smooth image processing is performed by switching the amount of halftoning at this time stepwise.

FIG. 8 is a diagram showing an embodiment in which the amount of halftoning in the pseudo halftoning process is switched stepwise.

1 and 7 denote the same processing unit or signal.

The signal 100 corrected by the correction circuit 13 is input to the edge detection circuit 14, the preprocessing circuit 16, and the mixer 86. At this time, the data is subjected to block transfer with m × n pixels as one block. Here, it is assumed that m = n = 3. However, it is needless to say that m = n = 5, m = 3, m = 5, and the like are not limited to m = n = 3.

The edge detection circuit 14 determines whether an edge exists in the block. Then, the presence of an edge is determined by the edge detection circuit 84. If an edge exists, “1” is output.
If not, “0” is output as signal 200.

FIG. 9 is a block diagram of the preprocessing circuit 16. The corrected signal 100 is input to the sum calculation circuit 90, where the sum S of the densities in the block is calculated. Is calculated. The calculated sum S is output as a signal 802 and input to the mixer 86 and the halftone circuit 91. In the halftone conversion circuit 91, pseudo halftoning is performed according to the sum S. The summation circuit 90 and the halftone circuit 91 can be realized by the same configuration as the summation circuit 21 and the halftone circuit 22 in FIG. The data on which the halftone processing has been performed is input to the mixer 86 as a signal 803.

FIG. 10 is a diagram for explaining the processing in the mixer 86. The signal 802 output from the preprocessing circuit 16 and the signal 200 output from the edge detection circuit 14 are
And outputs weight signals 811, 812 according to the signals 200, 802. At this time, the weight signal 811 is expressed as 1-α, and the weight signal 812 is expressed as α (0 ≦ α ≦ 1). These signals are input to weighting circuits 93 and 94, respectively, and the weighted data is output as signals 813 and 814. The adder 95 adds the signal 813 and the signal 814 and outputs the result as a signal 804.

Next, the operation will be described. A _ij (i, j = 1,2,3) represents the halftoned data 803, and D _ij (i, j =
1, 2, 3) represent the original data 100. The signals 200 and 802 are input to the weight determination circuit 92, and the signal 200 is "1".
, That is, when there is an edge in the block, α
= 1. At this time, the signal 811 is "0" and the signal 812 is "1". That is, in the edge portion, as in the above-described embodiment, the halftoning process is not performed because the resolution is not deteriorated. Then the signal
When 200 is "0", that is, when there is no edge in the block, the value of α is determined according to the value of the signal 802 as shown in FIG. For example, when the value of the signal 802 is 1275, α
The value of (signal 812) is 0.56, and signal 811 is 0.44 (1-α).
And FIG. 11 is an example showing the relationship between (signal 802) and α, and is not limited to this. In addition, by changing the relationship between the signal 802 and α in FIG. 11, it is also possible to make a halftone dot at an arbitrary density (signal 802).
This part can be realized by the LUT. Then, the halftone-processed data 803 (A _ij ) is multiplied by 1-α, and the original data 100 (D _ij ) is multiplied by α. And this weighted data 813
And 814 are added by the adder 95. This can be written as:

804 (addition result) = (1−α) A _ij + α · D _ij (i, j = 1,2,3) The addition result is output as a signal 804.

By converting the non-edge portion into a halftone dot with the above configuration, it is possible to avoid a decrease in resolution. Further, by performing the weighting stepwise, the image can be smoothly halftoned in accordance with the density. Therefore, generation of pseudo contours and noise can be suppressed.

The signal 804 output from the mixer 8b is transferred to the memory 17. The signal 805 read from the memory 17 is a binarization circuit 1
Step 8 is entered where the binarization process is performed. This binarization circuit 18 can be realized with the same configuration as that of the binarization circuit 18 shown in FIGS. The signal 806 output from the binarization circuit 18 is output as a binary image in the printer 89.

The present invention is not limited to the above-described embodiment, but can be applied to a color image. In this case, the circuit shown in FIGS. 1, 7, and 8 can be constituted by having the three colors R, G, and B. If the circuit of FIG. 8 has three colors,
It is also possible to change the relationship between (signal 802) and α shown in FIG.

As described above, according to the above-described embodiment, the edge detection is performed, and only the portion where no edge is present or only the highlight portion is detected by detecting the highlight portion, thereby performing the error diffusion method or the like. It is possible to suppress a unique fringe pattern generated at the time of binarization. In addition, by weighting and mixing the original data and the data that has been subjected to halftoning in accordance with the image density, it is possible to suppress noise that occurs at the time of halftoning, and further eliminate pseudo contours. or,
Since the halftoning process can be performed while storing the original image data, a reproduced image faithful to the original can be obtained.

(Effect)

As described above, according to the present invention, halftone dots are formed on non-edge portions of an image to form pseudo halftone dots, and then binarization is performed by an error diffusion method, thereby providing error diffusion. It is possible to suppress the generation of a unique fringe pattern generated when the binarization processing is performed by the method.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of the present invention, FIG. 2 is a block diagram showing details of a preprocessing circuit 16 in FIG. 1, and FIG. 3 is a detail of a summation circuit 21 in FIG. FIG. 4 is a diagram showing data of an original image (3 × 3 pixels), FIG. 5 is a diagram showing data after halftone processing, and FIG. 6-1 is a binarization circuit of FIG. 18 is a block diagram showing details of FIG. 18, FIG. 6-2 is a diagram showing an example of weighting factors, FIG. 7 is a diagram obtained by partially changing FIG. 1, and FIG. 8 is a block diagram showing a second embodiment. 9, FIG. 9 is a block diagram showing details of the pre-processing circuit 85 of FIG. 8, FIG. 10 is a block diagram showing details of the mixer 86 of FIG. 8, and FIG. is a diagram showing the relationship of α, in the figure, 11 is an input sensor, 12 is an A / D converter, 13 is a correction circuit, 14 is an edge detection circuit, 15 is a selector, 16 is a preprocessing circuit, 17 is a memory, and 18 is a memory. Binarization circuit , 19 is a printer, 20 is a highlight detection circuit, 21 and 90 are summation operation circuits, 22 and 91 are halftone circuits, 86 is a mixer, 92 is a weight determination circuit, 93 and 94 are weighting sections, and 95 is a weighting section. It is an adder.

Claims (1)

(57) [Claims] An image is divided into blocks each consisting of image data of a plurality of pixels, and a discriminating means for discriminating whether or not an edge exists in the block; Processing means for performing halftone processing; and for a block having no edge, image data subjected to halftone processing by the processing means is subjected to error diffusion.
Binarizing means for performing binarization processing and performing binarization processing on image data by an error diffusion method without performing halftoning processing by the processing means for a block having an edge; Means for calculating the sum of the densities of the image data of a plurality of pixels in the block; and allocating means for assigning the sum to one or more pixels in the block according to the value of the sum. The allocating means, when allocating the sum to a plurality of pixels, allocates the data such that the sum of the data to be allocated to the plurality of pixels is equal to the value of the sum calculated by the calculating means, and allocates the sum calculated by the calculating means. A central pixel, a pixel adjacent to the central pixel, and other pixels are assigned larger data in order according to the size of the pixel. Place.