JPS59123367A - Picture reading system - Google Patents

Abstract

PURPOSE:To obtain excellent picture quality inhibiting Moire by selecting a signal having a larger level out of outputs from two sensors where a picture reading position is provided relatively with a shift by a minute amount in the scanning direction as a reading signal so as to suppress moire. CONSTITUTION:A scanner section 1 consists of a light source 3 irradiating the surface of an original 2, reflecting mirrors 7, 8 making the reflected light incident to two line sensors 5, 6 via a lens 4, and a light shield plate 9. The sensors 5, 6 where lots of image pickup elements are arranged in a prescribed pitch in the direction of main scanning sample an original picture formed via the lens in a prescribed period. The picture reading position of the sensors 5, 6 is shifted relatively by a minute amount DELTA in the direction of main scanning, respectively. Signals (s), (t) are outputted from the sensors 5, 6, amplified by amplifiers 15, 16 and inputted to a comparator 10 and then a selecting circuit 13 via delay circuits 11, 12.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field] The present invention relates to an image reading method that reads an image by dividing it into pixels.

[Prior art]

BACKGROUND ART In digital copying machines such as Fakunmi and Riya, a line sensor is generally used to scan an image and sample it at a constant cycle, thereby decomposing the image into pixels and reading the image. However, when a periodic image is read using this method and produced in 11), there is a problem in that a strong moiré (striped pattern) occurs.

One way to prevent this moiré is to
(There is a random dither method that randomly changes the binarization threshold of the number. However, when this random dip method is applied to non-periodic halftone paintings or line drawings, the image quality of the 14 raw image deteriorates considerably. It has disadvantages and also complicates the configuration of the reading device.As a modification of the random dither method,
A method of controlling the variance of random numbers used as a value threshold according to the properties of the tWt image (Special Publication No. 57 = '6077
No. 2). This method can improve the image quality compared to the random dither method, but 1. This has the disadvantage that the reading device becomes more complicated due to the need to determine the properties of the bottom-removed image.

〔the purpose〕

The purpose of the present invention is to suppress moiré regardless of the image attributes (binary, halftone, halftone) without complicating the reading device, and to obtain an image that is better than the random dither method. We will provide a JC method.

〔Example〕

The 1lIII image reading method invented by Scanner uses two sensors that sample at a constant cycle while scanning one image per image, and relatively shifts the image and one source position by a minute amount in the scanning direction. Among the signals for the corresponding pixel with the valley sensor output level 2, the signal with the higher level is selected as the reading signal at t1. Below is Figure 1.゛Referring to Figure 13,
An example of non-invention (explained in total 71.111).

FIG. 1 is a schematic diagram of an image reading device that is unique to the invention.

J is a scanner section, and a light source 3 illuminates the surface of the document 2.
, a reflecting mirror 7.8 for allowing reflected light from the original 2 to enter the two line sensors 5.6 via the lens 4, and a light shielding plate 9.

The line sensor 5.6 has a large number of image sensors arranged at a constant pitch in the main scanning direction, and images are formed through a lens.
Original B, one image is sampled at a constant period. The image reading positions of the line sensors 5 and 6 are relatively flat by a minute amount Δ in the main scanning direction (the (arrangement direction) of the image pickup elements on the line sensor). In this embodiment, , Lips selected by ΔNia.

p is the sampling period of the line sensor 5.6 (the arrangement pitch of the imaging elements). Note that the sub-scanning is performed by moving the document 2.

From Line Sen-4115, 6, ll! is synchronized with a common clock (main scanning clock) provided from a part not shown. No. 48 (an analog signal with a level proportional to the pixel ferocity), s, and t are serially output in II cumulative units. These signals s and t are amplified by amplifiers 15 and 16 and input to a comparison circuit 10, and further input to a selection circuit 13 via delay circuits 11 and 12. This delay circuit l], 1
2 is for compensating for the operation delay of the comparison circuit 10, and if the comparison circuit 10 is sufficiently fast and its operation delay can be ignored, the delay circuits 11 and 12 can be omitted. The comparator circuit 10 compares the levels of the signals s and t, and if s)t, then 0''
If S < t, output "1". Selection circuit 13
If the output of the comparator circuit 10 is O"l, then the delay circuit selects the delayed signal S and outputs it as the read signal ■, and if the output of the comparator circuit 10 is 11111, then the delay circuit
The signal t that has passed through 2 is output as a read signal V. 1
4 is a binarization circuit, which converts the 1-value signal V into 2 values at a certain threshold value.
Convert and output.

Next, the occurrence of moiré will be explained with reference to FIG.

Now, suppose that the ball's own ball pattern on a certain Tsutomori line is a periodic black and white pattern as shown in FIG. 2(a).

Each imaging probe of one line sensor 5 is positioned as shown in FIG. 2(b), and each imaging probe of the other line sensor 6 is positioned as shown in FIG. 2(C). If it is assumed that
It becomes as shown in e). That is, even though a monochrome pattern with a constant period is being read, the signals s and t have a period of f. It's wavy. This is moiré. Line sensor 5.
Since the sampling period according to No. 6 is the same, the moiré period is the same for both signals S and t, but the phases are different. When Δ=tan, moiré with opposite phases occurs in the signals s and t. In Figure 8g2, the position in the main scanning direction is taken as the horizontal axis, so the signals Si, t for the corresponding pixel l are expressed as shifted by Δ, but the signals S1, tl for the corresponding pixel l are expressed by the line sensor. 5.6 will be output at the same time. That is, the signals s and t are outputted from the line sensor 56 at the timings shown in FIGS. 3(a) and (1).

This will be explained below with reference to FIG.

Since the pixel l is Sat, the signal S is selected as the read signal V tL. Pixel! +1, s (t, so the signal t is mF
=Selected as signal V. Similarly, pixel i''-2
, 1+3, the signal t is selected, and ! Signal S is selected from +4 to i+7. Therefore, the signal V is
The result is as shown in CI, and the binarized signal is shown in Figure 3(d).
Moiré is completely suppressed.

In this example, a line sensor with a scanning function was used as the sensor, but it is also possible to use a sensor with a single image sensor and perform main scanning by moving the sensor or the sensor, or by combining it with optical means. good.

[Efficacy] More than that? As is clear from the description, the present invention can suppress moiré and degrade the image quality of all types of binary, halftone, and 114-point images with a simple device configuration. It is possible to read the data without any trouble.

[Brief explanation of the drawing]

FIG. 1 is a 4'1q diagram of an image reading device according to the present invention, and FIG. 2 is an image pattern, a line sensor element arrangement,
FIG. 3 is a cross-sectional diagram of the line sensor output signal, the read signal, and its binary signal. 5.6... Line sensor, 10... Comparison circuit, 13
...Selection circuit. Representative Patent Attorney Makoto Suzuki (',,'7”,=”
,":;+chi. Figure 3 Li,,+It+2Lsu3L+4a,+5L child 6 is also +7
(Li.

Claims (2)

[Claims] (1) Two sensors for sampling at a constant cycle while scanning an image are installed /C1'1. The position for reading both openings is shifted in the forward direction by a slight amount of 41:1, and the valley sensor outputs signals for both corresponding systems, and the g signal with the higher level is output. , 1lIll image with %(di) selected as the I'C signal,
JJ'e method. (2) II between sensors! Poem 6' ``Claim 1 JJt A self-published image reading system, characterized in that the maximum deviation of the image acquisition position is approximately half of the sampling period.