JP2801195B2 - 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 such as a digital copier and a digital facsimile for digitally processing an image, and more particularly, to a method of converting input image data into three or more levels by an error diffusion method. The present invention relates to an image processing apparatus that performs multi-value processing on multi-value data. [Prior art] In general, digital copying machines that sample images using a CCD sensor or the like, output digitized data from a digital printer such as a laser beam printer, and reproduce the images, are conventional analog copying machines due to the development of digital devices. It is becoming widely used instead. This digital copying apparatus generally adopts a method of reproducing gradation by a dither method or a density pattern method for reproducing a halftone image. However, such a method has the following problems. (1) When the original is a halftone image such as a print, a periodic manuscript not included in the original is generated in the copied image. (2) If the original contains line drawings, characters, etc., the dithering process causes the edges to be cut off, resulting in lower image quality. The development in which the manuscript pattern is generated is called moiré development, and the cause thereof is A) a beat between the halftone dot document and the input sampling B) a beat between the halftone dot document and the dither threshold matrix. In particular, the phenomenon (B) is that, when the threshold value of dither is generally arranged in a dot-concentrated manner, the output image also has a pseudo halftone dot structure, which generates a beat between the input image and the moire image. Cause it to occur. On the other hand, there is an error diffusion method as another binarization method. This method is a method of calculating a density difference between an image density of a document and an output image density for each pixel, and distributing an error as a result of the calculation after assigning specific weights to peripheral pixels. For this, see RW Floyd and L. Stei
nberg, SID.17, PP75-77 (1976). This method can preserve the density of the original and has no periodicity, so that moire does not occur in the halftone dot image as compared with dither processing. Another feature is that the image resolution is superior to that of dither. [Problems to be solved by the invention] However, in the case of the error diffusion method in the above conventional example,
An image having a small change in shading, such as a photograph, has drawbacks such as the generation of a unique draft pattern in the output image, and the appearance of grainy noise in highlights and shadows of the image. As a countermeasure, a method of using a multi-value printer to make noise and draft patterns inconspicuous due to intermediate density has been considered. In this case, a method is conceivable in which the output level of error diffusion is set to a multi-valued level of three or more values. The present invention eliminates the above-described disadvantages of the prior art, adds a dither signal having a size corresponding to the edge amount of an image to input image data, and converts the image data to which the digit signal has been added to three or more levels by an error diffusion method. Multi-value processing is performed on multi-valued data. High-quality images can be obtained by multi-value processing in the error diffusion method, and the occurrence of false contours, which was a problem in multi-value processing, is prevented. It is an object of the present invention to provide an image processing apparatus capable of performing the above. Means for Solving the Problems In order to achieve the above object, an image processing apparatus of the present invention performs multilevel processing of input image data into multilevel data of three or more levels by an error diffusion method. Wherein input means for inputting image data, dither signal generating means for generating a dither signal, edge amount detecting means for detecting an edge amount from the image data input by the input means, and the edge amount detecting means Control means for controlling the magnitude of the dither signal generated by the dither signal generation means such that the smaller the detected edge amount is, the larger the dither signal is; and the dither signal whose magnitude is controlled by the control means is input to the input means. Means for adding to the image data input by the means, and an image data to which the dither signal whose size is controlled in accordance with the edge amount is added by the adding means. Processing means for performing multi-level processing on data at three or more levels, and error data addition for adding error data generated at the time of multi-level processing in the processing means to newly input image data Means. Embodiment An embodiment of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram showing one embodiment of the present invention. Image data read by an input sensor unit 10 including a photoelectric conversion element such as a CCD and a drive system for scanning the same is sent to an A / D converter 11. The A / D converter 11 converts the data of each pixel into 8-bit digital data. As a result, the data can be quantized into data having 256 levels of gradation. Next, the digital data quantized to 8 bits is supplied to the correction circuit 12.
Here, shading correction for correcting unevenness in sensitivity of the sensor and uneven illuminance caused by the illumination light source is performed by digital arithmetic processing. Next, the correction signal 100 is supplied to the edge detection circuit 13, the adder 1
Sent to 5. The edge detection circuit 13 detects an edge component of the image and outputs a signal 102 corresponding to the presence or absence of an edge to the additional data setting unit 14. The additional data setting unit 14 supplies additional data 108 suitable for the characteristics of the image to the adder 15. The adder 15 adds the correction signal 100 and the additional data 108, and outputs the added signal 103 to the error diffusion circuit 16. The error diffusion circuit 16 re-quantizes the signal 103 with a quantization number according to the output level of the multi-level printer 16, and outputs the re-quantized signal 101 to the multi-level printer 17. Multi-level printer 17 is a binary printer with ON / OFF dots
While only two states (black and white) can be represented, gray (gray), which is an intermediate level, is represented by controlling the amount of ink, for example. One-level gray printer is a ternary printer, two-level gray printer is 4
It is called a value printer. The multi-value printer 17 controls the amount of ink for each pixel and outputs the image on a predetermined sheet to form an image. FIG. 2 is a block diagram showing details of the edge detection circuit 13 of FIG. Image data 10 corrected by the correction circuit 12
0 is selected by the selector 20 and the line buffer memory 2
Sent to 1a-21d. Selector 20 uses this line buffer 2
One of 1a to 21d is selected and data is written.
The selector 22 selects one of the remaining three line buffers and reads the data. When the writing of the image data to the first line buffer 21a is completed, the next data is transferred to the second line buffer 21a.
Is written to the line buffer 21b. This is sequentially called the third and fourth
The data is written to the line buffers 21c and 21d. When the writing to the fourth line buffer 21d is completed, the process returns to the first line buffer 21a to write the data. As a result, three consecutive line data before the currently written image data line are recorded in the line buffer, and the line data is selected and read out by the selector 22. Next, the line data is sent to a max maximum value (hereinafter referred to as max) detection circuit 23 and a min minimum value (hereinafter referred to as min) detection circuit 24. Next, the values of max205 and min206 are calculated by the subtractor 25 as max205-min206. This value is compared with a specific threshold value T1 by the comparator 26. If the value is larger than T1, it is judged as an edge portion, and if it is smaller, it is judged as a non-edge portion, and an output signal 102 of 0 is obtained. This signal 102 is input to a selector 62 shown in FIG. FIG. 3 is a block diagram showing details of the max detector 23 and the min detector 24 in FIG. The image data 202, 203 and 204 of the line selected by the selector 22 are latches 30a to 30c and 31a to 3
1c, 32a to 32c are delayed by one pixel. The comparison selector 33a compares the outputs of the latches 31a and 32a, which compares the data of a certain pixel with that of the next pixel.
Similarly, the comparator 34a compares the output result of the pixel 33a with the data of the next pixel. Therefore the output of 34a is
It becomes max or min of three consecutive pixels in one line. Next, the comparison selector 35 sets the max of the first line and the second line, or
The min / min is detected, and the comparison / selection unit 36 detects the result and max / min of the third line. As a result, the output of the comparison selector 36 becomes max or min in the 3 × 3 pixel block. Here, max and min are obtained by the configuration of the comparison selector described below. FIG. 4 shows the configuration of the comparison selector when detecting max. The inputs A and B are output to a comparator 40 and latches 41 and 42, respectively. In the comparator 40, when A> B, the comparator
The output of 40 is 1 and this signal is passed through inverter 43
It enters the enable terminal of the latch 41. Assuming that the latches 41 and 42 have negative logic, the output 45 has the value of A. Conversely, when A ≦ B, the output 45 takes the value of B. This
The larger value of A and B is output to 45. On the other hand, the min detector can be easily realized by putting it in the inverter 43 latch 42 side. FIG. 5A shows the edge detection output 102 of FIGS.
FIG. 10 is a view showing another embodiment for obtaining the above. In the data of three lines from the selector 22, the central pixel is 50c and the peripheral pixels 50a, 50b, 50d, 50e. The center pixel is a multiplier 51
, And the result is input to the subtractor 53. On the other hand, the sum of the peripheral pixels is calculated by the adder 52 and then the subtractor 53
to go into. As a result, the output of the subtractor 53 is: output 53 = constant × (50c) − {(50a) + (50b) + (50d) + (50e)}. This corresponds to the Laplacian operation shown in FIG. 5-2 (a). (Constant = 4) Next, the output of the subtractor 53 is compared with a threshold value T4 by a comparator 54. Here, when the output value is larger than the threshold value T4, the edge portion is set to 1; The output 102 is obtained. Similarly, the same result can be obtained by using the Laplacian coefficient of the coefficient as shown in FIGS. 5-2 (b) and (c). FIG. 6 is a block diagram showing the details of the additional data setting unit 14 of FIG. Output from dither matrix circuit 60
106 is multiplied by K with a constant being a multiplier 61 to obtain a signal 107.
The signal 105 is a fixed value, for example, 0. And fixed value 1
05 and the signal 107 are switched in the selector 62 by the edge detection signal 102, and the signal 108 is obtained. That is, the selector 62 selects the signal 105 (= 0) at the time of an edge, and selects the signal 107 at the time of a non-edge. Thereby, the data value 108 added to the image data 100 is determined. Dither matrix 60
In FIG. 8, for example, coefficient values having values from -8 to +8 as shown in FIG. 8 (a) are sequentially and repeatedly output in synchronization with the pixel block. In the figure, a 4 × 4 matrix is shown, but -8, 4, −1,
In the order of −5, in the second line, the coefficient value is 1 in the order of −4, 8, 7, 3.
Outputs 06. The same effect can be obtained by changing the signal value only in the one-dimensional direction as shown in FIG. FIG. 7 is a block diagram showing details of the error diffusion circuit 16 of FIG. The image data 100 is added to correction data 114 described later by the adder 70 to form a correction signal 109. This signal is quantized to a multilevel level by the requantizer 72. Here, three values (0,
128,255) Here, T ₁ and T ₂ are set in the requantizer 72 by the setting signal 112 based on the setting of the threshold value setter 71. The output signal 110 is adjusted by the output buffer 73 with the timing of the multi-level printer 17, and is converted into a signal 101 sent to the multi-level printer 17. On the other hand, the signal 110 and the signal 109 are subtracted by a subtractor 74, and this is used as error data 111 as m in the error buffer memory 75.
Is stored in the location corresponding to. The error buffer memory stores the error data generated by the requantization in the current line and in the 2nd line.
A total of three lines are stored before the line. Assuming that the position of the pixel of interest for which the current processing is to be performed is the position of m in the error buffer memory 75, the correction data 114 is obtained by subjecting the data 113 of the 12 pixels of the error data a to l of the peripheral pixels to a predetermined weighting operation and the sum thereof. It is a signal that has been subjected to arithmetic processing by the weighting arithmetic circuit 26. When this is expressed by an equation, image data x _ij , correction data x ′
_ij , requantized data y _ij , weight coefficient α _kl , and error ε _ij It is. FIG. 9 shows an example of the weighting coefficient. FIG. 10 is a view showing another embodiment of the present invention. The output 100 from the correction circuit 12 is mixed with a value obtained by multiplying the output 108 from the additional data setting unit 14 by α (0 ≦ α ≦ 1) by the mixer 90. This α is the output 120 from the edge detector (FIG. 2),
Alternatively, the output 121 (FIG. 5-1) is input to the mixing ratio determination circuit 91, and the output is output 123 by a function corresponding to FIG.
Is output to the mixer 90. The curve shown in FIG. 11 is experimentally determined. When the outputs 120 and 121 are small, the edge portion is determined to be a non-edge portion, α = 1, and all the additional signals are added to the original signal. Conversely, when the outputs 120 and 121 are large, that is, when the edge amount is large, α = 0 and there is no additional signal, and only the original signal is output from the mixer 90. In this way, by changing the ratio of the signal added to the original signal according to the magnitude of the edge component, the dot component has a large edge component in the halftone dot image, so that the ratio of the additional signal is small.
Since the edge component is small in an image in which pixels are continuous, the ratio of the additional signal increases. This makes it possible to prevent the occurrence of moire in a halftone dot image by the error diffusion method, and to prevent the occurrence of a false contour, which is a drawback of a multi-value printer, in an image having continuous pixels. As described above, according to the above-described embodiment, the edge of the image is detected by the edge detecting means, and the requantized output by the error diffusion method is selected in the edge portion, while the dither signal is added to the original data in the non-edge portion. By providing switching means so as to select an output obtained by requantizing the data by the error diffusion method, it is possible to enhance the edge and maintain a high resolution by the error diffusion method, and there is a problem of outputting the data by a multi-value printer. It is possible to prevent the occurrence of a false contour of an image and obtain a high-quality image. [Effect] As described above, according to the present invention, a dither signal having a larger value is added to input image data as the edge amount of the image is smaller, and the image data to which the dither signal has been added is subjected to an error diffusion method to three or more levels. Multi-value processing is performed on multi-valued data. High-quality images can be obtained by multi-value processing in the error diffusion method, and the occurrence of false contours, which was a problem in multi-value processing, is prevented. can do.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a diagram showing an edge detecting circuit using maximum and minimum values, and FIG. FIG. 4 shows a value detection circuit, FIG. 4 shows a comparison / selection circuit, FIG. 5-1 shows an edge detection circuit using a Laplacian method, and FIG. 5-2 shows an example of a Laplacian filter. FIG. 6 is a block diagram showing details of the additional data setting unit, FIG. 7 is a block diagram showing details of the error diffusion circuit, and FIG. 8 is a diagram showing coefficient values in the dither matrix circuit. 9, FIG. 9 is a diagram showing a diffusion matrix and a weight coefficient value of an error diffusion circuit, FIG. 10 is a block diagram showing a second embodiment of the present invention, and FIG. 11 is a diagram for determining a coefficient α. Explanation diagram, 10 is an input sensor unit, 11 is an A / D converter, 12 is a correction circuit, 13
Is an edge detection circuit, 14 is an additional data setting device, 15 is an adder, 16 is an error diffusion circuit, and 17 is a multi-value printer.

Claims (1)

(57) [Claims] What is claimed is: 1. An image processing apparatus for performing multi-level processing of input image data into multi-level data of three or more levels by an error diffusion method, comprising: input means for inputting image data; dither signal generating means for generating a dither signal; An edge amount detecting means for detecting an edge amount from the image data input by the input means; and a dither signal generated by the dither signal generating means such that the smaller the edge amount detected by the edge amount detecting means is, the larger the dither signal becomes. Control means for controlling the size of the image data; addition means for adding the dither signal whose size is controlled by the control means to the image data input by the input means; and Processing means for performing multi-level processing on the image data to which the controlled dither signal has been added into multi-level data of three or more levels; An image processing apparatus, comprising: error data adding means for adding error data generated during multi-level processing in a stage to newly input image data.