JPS6259512B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an image processing device used in facsimiles and the like, which obtains an apparently high-quality output image from image information read by a low-resolution image reading device.

Conventionally, this type of image processing device reads pixels in a checkered pattern as shown in FIG. 1a, and outputs pixels marked with ◎ corresponding to the read pixels as shown in FIG. A method is known in which the pixel is restored using a signal and the output pixel marked with a circle corresponding to the skipped pixel is predicted from the pixel signals of the surrounding pixels. In this case, the predicted pixel x is determined by x=1/4 (a+b+c+d), for example. Here, a, b, c, and d are pixel signals of pixels surrounding the predicted pixel x, and x represents an estimated value of the predicted pixel x. However, in such a device, there is a one-to-one correspondence between the reading pixel and the output pixel marked with ◎, so in order to obtain high-quality output pixels, the resolution of the image reading device such as an image sensor must be high. .

Furthermore, as in Japanese Patent Application Laid-Open No. 128410/1982, a device was proposed that restores the output pixel corresponding to a read pixel from the pixel signal of the read pixel and predicts the output pixel between these output pixels from the pixel signal of the read pixel. has been done.
However, since this device predicts output pixels that do not correspond to the read pixels from the pixel signals of the read pixels, the intelligibility of the output image is low.

SUMMARY OF THE INVENTION It is an object of the present invention to provide an image processing device that eliminates the above-mentioned drawbacks and can obtain output images with high intelligibility and high quality using an image reading device with low resolution.

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 2 shows an image reading device according to an embodiment of the present invention, in which an image sensor 1 separates and reads scanning lines on an original 3 into pixels through a lens system 2, and the original 3 is fed and scanned in sub-scanning mode. will be held. When reading the letter "B" from the original document 3 with this apparatus, the letter "B" is converted into pixels as shown in FIG. In a normal image output device, the output pixel density corresponds one-to-one with the reading pixel density of an image reading device.
An output image as shown in FIG. 4 is obtained. However, in this embodiment, as shown in FIG.
1, 12, . . . are two output pixels 11-, respectively.
1, 11-2, 12-1, 12-2, . . . That is, the image output device outputs an image based on the image signal from the image sensor 1, but the number of output pixels of the image output device is twice the number of pixels read by the image reading device, and the image reading device uses a low resolution one. be able to. The image signal from the image sensor 1 is made to have apparently higher quality by the image processing circuit, that is, the number of pixels is doubled, and is then provided to the image output device. For example, when black information is input to the image sensor 1 from a pixel 11, two output pixels 11-
1 and 11-2 are both output as black. When gray information is input to the image sensor 1 from a pixel 12, the pixel 12- corresponding to that pixel 12
1, 12-2 are pixels 11, 1 on both sides of pixel 12
Compare the image signals of 3 and select the one with higher input image density,
That is, depending on which one is closer to black, one 12-1 is outputted as black and the other 12-2 is outputted as white. That is, by comparing the pixels 11 and 13 on both sides, pixel 1 with high density is selected.
The output pixel 12-1, which is closer to the output pixels 11-1 and 11-2 corresponding to 1, is black, and the other output pixel 1 is black.
2-2 is output as white. Image sensor 1
When gray information is input from pixel 13, the output pixels 13-1 and 13-2 corresponding to pixel 13 similarly compare the pixels 12 and 14 on both sides of that pixel 13, and depending on the result, output pixels 13-1 and 13-2 corresponding to pixel 13. -1 is output as black and the other 13-2 is output as white. When white information is input to the image sensor 1 from a pixel 14, pixels 14-1 and 14-2 corresponding to that pixel 14
Both are output as white.

FIG. 6 is a flowchart showing the operation of the image processing circuit in this embodiment. The image signal S(i) obtained by reading the i-th pixel of the document 3 with the image sensor 1 is quantized into three values of 0, 1, and 2 and is converted to NS.
(i). In this case, the image signal S(i) is 0≦S
(i) If <0.3, it is determined as white and output NS(i)
becomes 0, and if 0.3≦S(i)<0.6, it is determined as gray and NS(i)=1, and 0.6≦S(i)≦
If it is 1, it is determined to be black, and NS(i)=2.
When the quantized output NS(i) is 0, the image output device outputs both output pixels P(i, 1) and P(i, 2) corresponding to the pixel signal S(i) as 0, that is, white. let When NS(i)=2, the image output device makes the output pixels P(i, 1) and P(i, 2) both 2, that is, black. When NS(i)=1, the quantized output NS(i-1) for the pixel signals S(i-1) and S(i+1) of the (i-1) and (i+1)th pixels before and after it, Determine whether the product of NS(i+1) is 0 or not, and if it is 0, NS(i-
Determine whether 1) is 0. and NS(i-
1) When = 0, P(i, 1) = 0, P(i, 2)
= 2, and when NS(i-1)≠0, P(i,
1)=2 and P(i,2)=0. NS(i-
1), if the product of NS(i+1) is not 0, NS(i
-1), NS(i+1) and NS(i-1)≦
When NS(i+1), P(i, 1)=0, P(i,
2)=2, and when NS(i-1)>NS(i+1), P(i,1)=2 and P(i,2)=0. In addition, when NS(i)=1, directly NS(i-1),
The process may proceed to a step of comparing NS(i+1).

FIG. 7 shows the above-mentioned image processing circuit, in which the image signal from the image sensor 1 is quantized into three values by a three-value quantization circuit 21 and sequentially transferred to one-pixel memories 22 to 24 consisting of shift registers. I'm going to be done. Comparison circuit 25 includes 1 pixel memories 22 and 24
The output signal generation circuit 26 generates an output signal for two pixels when the image signal in the one-pixel memory 23 is not 0, and outputs it to the image output device via the gate 27. The detection circuit 28 is a one-pixel memory 23
It is detected whether the image signal is 0, 1, or 2. The selection circuit 29 keeps the gate 27 closed when the image signal of the 1-pixel memory 23 is 0 based on the output signal of the detection circuit 28, and closes the gate 27 when the image signal is 1.
One of the two output signals from the output signal generation circuit 26 is passed in accordance with the output of , and in the case of 2, both of the two output signals are passed.

A well-known image output device is used as the image output device in this embodiment, and the output image for the letter "B" in FIG. 3 is as shown in FIG. 8. It has been confirmed through experiments that the intelligibility of characters is generally significantly improved by performing such image processing. In particular, the print quality for Kanji information is superior to that of the past, and this is especially noticeable in terms of intelligibility.

The above embodiment is an example in which two output pixels correspond to one reading pixel, but by making one reading pixel correspond to three or more output pixels, the output form of this output pixel is changed to that one reading pixel and the output pixel. It can also be determined according to the image signals of adjacent read pixels. In another embodiment of the present invention, one reading pixel corresponds to four output pixels, and an image processing circuit as shown in FIG. 9 is used. That is, the image signal from the image reading device is transferred to the shift register 30 which serves as a pixel memory, and each pixel signal is sequentially transferred to the 5-value quantization circuit 31.
is quantized into five values. Furthermore, the shift register 30 has a capacity for 9 pixels or more, and as shown in FIG. If it is a pixel signal, the pixels surrounding the read pixel (i-1, j-
1), (i-1, j) (i-1, j+1), (i, j
-1), (i, j+1), (i+1, j-1), (i+
The pixel signals of the read pixels at addresses 1, j) and (i+1, j+1) are sent to the maximum value extraction circuit 32. The maximum value extraction circuit 32 compares these eight pixel signals with each other and extracts the maximum value signal.
The maximum value address designating circuit 33 is caused to designate one of the eight addresses A to H of the selection circuit 34 by one of the eight pixel signals. In this case, the maximum value extraction circuit 32 compares the eight pixel signals with each other, and when there are multiple maximum values, the pixel signal with the maximum value is read the earliest (the one sent earliest from the image reading device). ) selecting a pixel signal and thereby causing the maximum address designation circuit 33 to designate one of addresses A to H;
In addition, addresses A to H are the (i-
1,j-1) to (i+1,j+1). The output signal generation circuit 35 also outputs a pixel signal to be processed from the shift register 30 when the pixel signal is 0.
When it is not, it generates an output signal for 4 pixels and gate 3
6 to an image output device. Selection circuit 3
4 closes the gate 36 when the output of the 5-value quantization circuit 31 is 0, and closes the gate 36 when the output of the 5-value quantization circuit 31 is 1 to 4, respectively, as shown in FIG. 1 from gate 36
Output 4 pixels. Moreover, the selection circuit 34 selects at least one of the pixels a to d outputted by the output signal of the gate 36 to be black by the output of the quinary quantization circuit 31, which is the blackest among the pixels surrounding the pixel to be processed. The gate 36 is controlled in accordance with the address designation of the maximum address designation circuit 33 so as to be adjacent to the pixel.

As described above, according to the image processing device according to the present invention, the output pixel density is made larger than the reading pixel density, so it is possible to obtain a high-quality output image using an image reading device with a low resolution, and moreover, according to predetermined rules. Accordingly, if the pixel signal of a read pixel is quantized into n values (n is an integer of 3 or more) and the value is the maximum or the minimum, it corresponds to the read pixel (n-1)
All of the output pixels are set to be the same as the pixel signal of the reading pixel, and if the value obtained by quantizing the pixel signal of the reading pixel is an intermediate value, the output pixels of (n-1) corresponding to the reading pixel are Since the output form is determined based on the value of the pixel signal of the read pixel and the value of the pixel signal of the read pixel adjacent to this pixel, the intelligibility of the output image can be improved.

[Brief explanation of the drawing]

1A and 1B are diagrams for explaining a conventional image processing device, FIG. 2 is a schematic perspective view showing an image reading device in an embodiment of the present invention, and FIG. 3 is a diagram showing pixel decomposition of a read image. 4 is a diagram showing a normal output image, FIG. 5 is a diagram for explaining the same embodiment,
FIG. 6 is a flowchart of the same embodiment, and FIG. 7 is a block diagram showing an image processing circuit in the same embodiment.
FIG. 8 is a diagram showing an output image according to the same embodiment, and FIG.
The figure is a block diagram showing an image processing circuit in another embodiment of the present invention, and FIGS. 10 to 12 are diagrams for explaining the same embodiment. 11, 12, 13, 14...reading pixels, 11-
1, 11-2, 12-1, 12-2, 13-1,
13-2, 14-1, 14-2...Output pixel, 2
DESCRIPTION OF SYMBOLS 1... Three-value quantization circuit, 22-24... 1 pixel memory, 25... Comparison circuit, 26... Output signal generation circuit, 27... Gate, 28... Detection circuit, 29
...Selection circuit, 30...Shift register, 31...
...Five-value quantization circuit, 32...Maximum value extraction circuit, 3
3... Maximum value address designation circuit, 34... Selection circuit, 35... Output signal generation circuit, 36... Gate.

Claims (1)

[Claims] 1. A reading means that reads an image and converts it into an image signal, and converts the read image signal into an n value (n is an integer of 3 or more)
quantization means to quantize one reading pixel to 2
converting means for converting into (n-1) output pixels represented by a value to make the output pixel density larger than the reading pixel density, and the converting means converts the reading pixels according to a predetermined rule. When the quantized value of the image signal is the maximum or the minimum, all of the (n-1) output pixels corresponding to the reading pixel are made the same as the pixel signal of the reading pixel, and the pixel signal of the reading pixel is If the quantized value is an intermediate value, the output form of the (n-1) output pixels corresponding to the read pixel is the pixel signal value of the read pixel and the pixel signal of the read pixel adjacent to this pixel. An image processing device characterized in that the image processing device makes a decision based on a value of .