JP2000101796A - Image reader - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the occurrence of moire due to a sampling position shift by reading a document image with high resolution. SOLUTION: One document is scanned and read twice and an optical flat plate is slanted in the respective read scans to shift sampling position by a distance of half as large as one pixel. A CCD 12 takes formed images in by R, G and B and supplies them to corresponding memory and register correction parts 16a, 16b and 16c. A counter 14 performs counting operation indicating a read position in a horizontal scanning direction. Gravity center correction quantity tables 15a, 15b and 15c supply gravity center correction quantities αR, αG and αB corresponding to read positions in the horizontal scanning direction to the respective memory and register correction parts 16a, 16b and 16c. The memory and register correction parts 16a, 16b and 16c correct sampling position deviation quantities corresponding to field angles according to the gravity center correction quantities.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an image reading apparatus for reading an image with high definition.

[0002]

2. Description of the Related Art As an image reading apparatus, a high-speed image reading apparatus constituting a digital copying machine in combination with a laser printer and a low-speed reading apparatus for taking an image into a PC or the like are known. A reading device of a digital copying machine needs a reading speed of at least about 160 mm / sec in order to shorten FCOT (time from when a copy button is pressed until the first copy is ejected). On the other hand, considering the office environment in the future, the reading scanner is connected to the network independently of the copying machine, and is required to have both functions as an input device of the copying machine and a function as a scanner terminal on the PC network. Such P
When high-definition reading is required as a scanner terminal on the C network, a resolution of about 1200 dpi is required.

However, when reading at 1200 dpi at a reading speed of 160 mm / sec,
The reading speed of the CCD sensor is as high as 115 MHz.
This is a reading speed not possible with current technology.
From such a background, when it is required to achieve both high-speed reading and high-definition reading with one machine, it is conceivable to provide a high-speed reading mode and a high-definition reading mode as a countermeasure and to switch according to the application. As a configuration of a machine corresponding to this mode switching, there is a method proposed for a long time in TV cameras or the like that shifts the sampling phase by 1/2 pixel (Japanese Patent Laid-Open No. 51-25914). By using this conventional technology, one CCD sensor equivalent to 600 dpi can be used.
Sampling equivalent to 200 dpi can be performed.

[0004]

By the way, in the above-mentioned prior art, an optical plate provided in an image forming optical path is mechanically tilted so that an image forming position with respect to a CCD sensor is reduced by 1 /.
It is shifted by two pixels. Here, when examining in detail the imaging characteristics when the optical plate is tilted,
It was found that the amount of phase shift caused by tilting the optical plate was slightly different between the image center and the image end. This phase shift amount is determined by the 7500 pixel CCD sensor (A3 short hand 60
When a displacement of 0.5 pixels at a sampling pitch of 600 dpi is performed at the center using the number of pixels necessary for reading at 0 dpi), the displacement amount is about 0.6 pixels in the peripheral portion. However, even if the sampling position shifts by such a degree, when periodic information such as a halftone dot is read, it becomes a source of moiré. This slight moiré is emphasized when image processing such as "digital filter", "picture character separation", and "black plate generation by removing under color" is performed, and there is a problem that the graininess of the image is deteriorated.

In particular, in recent years, high-quality printing and the like have become widespread, and the number of halftone dots higher than that of the conventional 175 line standard has been increasing. In addition, an inkjet printer is, for example, 720d
Dither output is performed at a dot density of pi. Assuming that these printed matter and the printer output image are read as a manuscript, conventionally, the manuscript had a sufficiently high definition.
Even at a reading density of dpi, moire occurs due to the close proximity between the sampling pitch and the halftone dot pitch.
For example, when a dot of 720 dpi is read,
Moire of 120 dpi (4.7 lines / mm) corresponding to two spatial frequencies is generated. The spatial frequency of this moiré is close to the spatial frequency of general character information, and is therefore a frequency that is easily emphasized. Even if the reading sampling density is increased to 1200 dpi by the above-described method in order to read such an original more clearly, if the sampling position shift remains as described above, 6
There is a problem that a sampling frequency component at 00 dpi remains and a moire component remains. Further, in recent years, when the conjugate length of the imaging lens is further shortened due to the downsizing of the image reading apparatus, the imaging angle of view of the imaging lens tends to increase, and the amount of phase shift between the center of the image and the image end is increased There is a problem that the difference between the two increases further.

The present invention has been made in view of the above circumstances, and provides an image reading apparatus capable of reading a document image with high resolution and reducing the occurrence of moire caused by uneven sampling phases. It is intended to be.

[0007]

In order to solve the above-mentioned problems, according to the first aspect of the present invention, photoelectric conversion means for converting light reflected from a document into an electric signal, and reflected light for the photoelectric conversion means are provided. An image forming position changing means for changing the image forming position in the main scanning direction, and a reading means for reading an original image formed on the photoelectric conversion means at a different image forming position on the original by the image forming position changing means. Correction means for performing gravity center correction based on image data at a first imaging position read by the reading means and image data at a second imaging position different therefrom; and Changing means for changing the correction amount according to the position in the main scanning direction. According to the present invention, one original is read at least twice, and at the time of the two readings, the image forming positions are shifted, and the image data at the first image forming position and the second image forming position different from the image data at the first image forming position are read. The center of gravity is corrected based on the image data at the image position. With the correction of the center of gravity, it is possible to correct a sampling position shift (imaging position shift) between the central portion and the peripheral portion caused by the angle of view. This makes it possible to read at a high resolution, and it is possible to prevent the occurrence of moire caused by uneven sampling phases.

According to the second aspect of the present invention, a photoelectric conversion means for converting reflected light from a document into an electric signal, and an optical path from the document to the photoelectric conversion means are provided. Has optical characteristics parallel to the optical axis,
An image forming lens for forming an image of the reflected light on the photoelectric conversion unit, an image forming position changing unit for changing an image forming position of the reflected light on the photoelectric conversion unit in a main scanning direction, and Reading means for reading a document image formed on the photoelectric conversion means at different image forming positions by the image forming position changing means. According to the present invention, one original is read at least twice, and at the time of the two readings, the image forming positions are shifted, and the image data at the first image forming position and the second image forming position different from the image data at the first image forming position are read. In the image data at the image position, the shift amount of the sampling phase of each image can be the same in the central portion and the peripheral portion without depending on the angle of view. This makes it possible to read at a high resolution, and it is possible to prevent the occurrence of moire caused by uneven sampling phases.

[0009]

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

[0010] 1. First embodiment 1-A. Configuration of First Embodiment FIG. 1 is a schematic diagram showing a partial configuration of an optical system of a document reading apparatus according to a first embodiment of the present invention. In the figure, the original reading apparatus according to the present embodiment uses a full-rate mirror unit / half-rate mirror unit to
The document image is scanned and formed on the D sensor. In the optical system of the image reading apparatus, a transparent optical flat plate 3 having both surfaces parallel and having a predetermined thickness is provided at the subsequent stage of the image forming lens 2 for forming and projecting a document image on the CCD sensor 1. I have. The optical flat plate 3 is mounted on a rotatable table 4 as shown in FIG. 2, and has magnetic members 5a and 5b at both ends. Electromagnets 6a, 6b having the ability to attract the magnetic body sufficiently by magnetic force are arranged near the magnetic bodies 5a, 5b. Electromagnets 6a,
6b, an exciting current is alternately applied to the magnetic material 5a,
5b are alternately sucked, and the optical flat plate 3 is set to be inclined at a predetermined angle and at a predetermined timing.

FIG. 3 is a conceptual diagram showing how the imaging state on the CCD sensor 1 changes depending on the inclination of the optical plate 3 described above. From the law of light refraction and the nature of the trigonometric function, the parallel shift amount Δ of the light beam when the optical plate 3 is inclined by the angle θ is calculated by the following equation. θ '= sin ^-1 (sin θ / n) Δ = d · (tan θ−tan θ ′) · cos θ

According to the above equation, when the refractive index n = 1.6,
By tilting the optical flat plate 3 having a thickness of d = 10 mm by about 0.07 degrees, the imaging position on the CCD sensor 1 becomes about 4.5 μm.
It can be seen that the shift is m. For example, if the pixel pitch is 9 μ
When the m CCD sensor 1 is used, the sampling position can be shifted by 0.5 pixel. In an image reading apparatus of a type in which the main scanning direction is electronically scanned using a line sensor and the sub-scanning direction is mechanically scanned, the sampling density in the sub-scanning direction is determined by the mechanical scanning during one line scanning by the line sensor. It can be changed by changing the amount of movement.

Further, as shown in FIG.
In reading of 00 dpi (23.6 dots / mm),
The reading position is moved by 1 / 23.6 mm during the main scanning reading of one line by the line sensor. By halving the moving amount, the sampling density in the sub-scanning direction is doubled. (See FIG. 4B).

Next, two reading scans (with the sub-scanning density doubled as described above) are performed on one document, and
In each reading scan, by inclining the optical plate 3 to shift the sampling position by １ ／ pixel, the main scanning can be read at twice the original density (see FIG. 4C). ). At this time, the position reproducibility of the two mechanical scans is maintained very high by setting the same motor drive conditions for driving the scanning system each time.

1-B. However, since the imaging optical system has an angle of view, the optical flat plate 3 is tilted near the optical axis of the optical system when the optical flat plate 3 is tilted and at a peripheral portion away from the optical axis. May have different effects on the imaging characteristics. When this was verified by ray tracing or the like, it was found that the imaging characteristic changed slightly from the vicinity of the optical axis to the periphery with respect to an inclination of about 1 degree. On the other hand, it has been found that there is a difference in the shift amount of the sampling position between the vicinity of the optical axis of the imaging optical system and the peripheral portion.

FIG. 5 is a conceptual diagram showing a shift amount of an image forming position between a central portion and a peripheral portion in consideration of an angle of view. FIG. 6 is a conceptual diagram showing an example of the angle-of-view dependence of the imaging position due to the inclination of the optical plate 3. In FIG. 5, the position of an object outside the optical axis is represented by an inclination angle (angle of view) φ of the principal ray of the imaging light beam with respect to the optical axis, and the inclination of the optical plate 3 having a thickness d is represented by θ. , The amount of displacement of the sampling position on the line CCD sensor 1 is Δ = d · (tanψ−tan
ψ ') · cosθ. Where ψ '= sin ^-1
(Sinψ / n), ψ = θ + φ, and n is the refractive index of the optical flat plate 3.

For example, when the refractive index n = 1.6 and the thickness d = 1
The shift amount of the sampling position on the surface of the line CCD sensor 1 between the on-axis and off-axis when the 0 mm optical flat plate 3 is tilted by about 0.07 degrees, as shown in FIG.
In contrast to 4.5 μm, there is a difference of 5.3 μm in the peripheral portion. When the line CCD sensor 1 having a pixel pitch of 9 μm is used, the amount of displacement of the sampling position due to the inclination of the optical plate 3 is 0.5 pixels at the center, but about 0.6 pixels at the periphery. It corresponds to a shift. This is an amount based on the pixel (600 dpi) of the line CCD sensor 1, and corresponds to a displacement of the sampling position of about 0.2 pixel on an actual image obtained by performing sampling reading at 1200 dpi.

As described above, assuming that one original is read twice and the sampling position in the main scanning direction is changed during the two readings, as described above, Despite the high reproducibility, the shift between the two scans can exceed half the phase. In addition, an error may occur in the inclination amount of the optical plate 3. When these errors are added,
The above-described sampling position shift of 0.2 pixels between the central portion and the peripheral portion is added. For example, FIG. 7 shows what behavior occurs when the sampling position shift becomes 0.4 pixels due to the superposition of errors. FIG. 7 (a)
FIG. 7B shows read image data obtained when high-quality printing with 500 dots (19.7 dots / mm) of halftone dots is read without a sampling position shift, and FIG. 7B shows that the sampling position shift is 0.4 pixels. It is in a state where there was a minute.

As described above, if there is a sampling position shift, 600 dp is generated due to the periodicity of the sampling position shift.
i (23.6 dots / mm), and a sampling frequency of about 4 lines / mm
Moiré component occurs. Such a component having a low number of lines of 4 lines / mm is a frequency close to the spatial frequency of general character image information, and is a frequency component that is easily emphasized by image processing, and thus causes a problem in image quality. Therefore,
In the present embodiment, in order to correct the sampling position shift caused by reading the document twice, the amount of correction of the center of gravity is changed according to the angle of view.

1-C. Next, FIG. 8 is a block diagram showing an overall configuration including a correction circuit for correcting the angle-of-view dependency of the sampling position shift. In the figure, a clock generator 10 is a CCD
A clock supplied to the driver 11 and a clock synchronized with the scanning line are generated, and the CCD driver 1 generates the clock.
1. Supply to counter 14. The CCD driver 11
The CCD 12 is scanned based on the clock,
The image formed on 12 is captured for each of R, G, and B, and supplied to corresponding A / D converters 13a, 13b, and 13c. Each A / D converter 13a, 13b, 13c
Each of R, G, and B is converted into a digital signal, and supplied to the corresponding memory and registration correction units 16a, 16b, and 16c.

The counter 14 performs a counting operation indicating a reading position in the main scanning direction based on a clock synchronized with a pixel output of the CCD driver 11 and a clock synchronized with a scanning line. This count result is obtained by correcting the center of gravity correction amount table (R-LUT, G-LU) every 256 pixels.
T, B-LUT) 15a, 15b, 15c. Each of the gravity center correction amount tables 15a, 15b, 15c
The center-of-gravity correction amount for the reading position in the main scanning direction is stored, and the corresponding center-of-gravity correction amounts α _R , α _G , and α _B according to the reading position in the main scanning direction are stored in the respective memory & registration correction units 16a and 16b. , 16c. The angle of view corresponds to the reading position in the main scanning direction. Each of the memory and registration correction units 16a, 16b, and 16c stores a centroid correction amount table 15
The sampling position shift amount corresponding to the angle of view is corrected based on the centroid correction amount from a, 15b, and 15c.

1-D. Memory & Registration Correction Unit Next, FIG. 9 is a block diagram illustrating a configuration of the memory & registration correction unit. Parts corresponding to those in FIG. 8 are denoted by the same reference numerals. Hereinafter, the configuration and operation of the memory and registration correction unit will be described. In the figure, a linear CCD sensor 1
The output from 2 is converted into a digital signal by the A / D converter 13 and is subjected to a preprocessing circuit (shading correction and registration correction for the line gap of the 3-line color CCD sensor).
After passing through 20, it is supplied to the illustrated changeover switch SW1. Here, in the first scan reading, the switch SW1 selects the path on the contact a side, and the read data is stored in the memory M0.

Next, in the second scan, the changeover switch SW1 shown selects the path on the contact b side. At this time, the reference clock driving the linear CCD sensor 12 is counted by the address counter 14, so that the read data at the position of the pixel being processed is read out from the memory M0 and, as described above, 256 pixels are read out. LU in which the correction amount is stored in increments
The center of gravity correction amount is read from T (correction table) 15. As the barycenter correction amount, the phase of the output pixel of the correction circuit 16 is
Two types of data are provided in accordance with a pixel located near the pixel position of the first scan (this is referred to as an odd-numbered pixel) and a pixel located near the pixel of the second scan (this is referred to as an even-numbered pixel). α _E and α _O are output. Switching circuits 21 and
2 generates control signals CS1 and CS2 of the changeover switches SW2 and SW3 in accordance with the signs of the two types of data α _E and α _O , and | α _E | , 1− | α _E |, | α _O |, 1− | α _O
| Is generated.

On the other hand, the switch SW2 has a pre-processing circuit 2
The data (D2 _{2i + 2} ) output from 0 and following the path on the contact b side of the switch SW1 and the data (D2 _2i ) one clock before output from the latch circuit 23 are supplied. Similarly, the switch SW3 has the output (D1 _{2i + 1} ) from the memory M0 and the _one ( ₁ ) output from the latch circuit 24.
The data before the clock (D1 _2i-1 ) is supplied. As a result, as a result of the two sampling scans, data of four spatially adjacent sampling points, D
Among the 1 _{2i -1} , D2 _2i , D1 _{2i + 1} , and D2 _{2i + 2} , those selected by the switches SW2 and SW3 are connected in parallel to the multiplier 2
5, 26, 27, 28.

In each of the multipliers 25, 26, 27, and 28, the above four data D1 _2i-1 , D2 _2i , D1 _{2i + 1} , D2 _{2i + 2}
, The weighting factors | α _E |, 1− | α
_The center of gravity is corrected by multiplying _E |, | α _O |, 1− | α _O |. Centroid corrected here is odd about the pixel, the output of a reference D1 _{2i + 1,} the center of gravity correction, whether shifted to the front side of the pixel (D2 _2i), or behind the pixel (D2 _{2i +2} ), the switching circuit 21
Is selected by switching the switch SW2 according to the control signal CS1 from the CPU. In addition, the pixel (D2 _2i ) is subjected to gravity center correction and the switch SW3 is switched by the control signal CS2 from the switching circuit 22 to select either D1 _{2i + 1} or D1 _2i-1. I do.

[0026] In the multiplier 25, | alpha _O | the multiplied data D2 _{2i + 2} or D2 _2i as the weight coefficient of the center of gravity correction, the multiplier 26, 1- | α _O | data as a weighting coefficient for the center of gravity correction D1 _{2i + 1} is multiplied. Further, these are added by the adder 29 and stored in the memory M1. Also, multiplier 2
7, the data D are set as | α _E |
1 _2i-1 or D1 _{2i + 1} is multiplied.
| Α _E | is multiplied by the data _D22i as a weighting coefficient for the correction of the center of gravity. Further, these are added by the adder 30 and stored in the memory M2. Since these two are output in parallel,
The video rate is the same as the video rate at the output of the pre-processing circuit 20.

In the above embodiment, the optical flat plate 3
The problem was the displacement of the sampling position between the vicinity of the optical axis (center part) and the peripheral part of the image away from the optical axis due to the inclination of the optical axis. There is color misregistration caused by chromatic aberration. Regarding this chromatic aberration of magnification, the manufacturing variation of the imaging lens 2 is also a major factor, and thus the amount of aberration differs for each device. To cope with such a situation, the ladder pattern is actually read, and based on the read result, the barycenter correction amount table of three colors is read, and the barycenter correction amounts of the three colors vary,
Variations in lateral chromatic aberration for each device can be corrected. Further, the variation in the amount of deflection of the optical flat plate 3 can be absorbed. Hereinafter, a procedure for obtaining a correction amount from data obtained by reading a ladder pattern will be described.

1-E. Correction of lateral chromatic aberration by reading ladder pattern Here, FIG. 10 is a conceptual diagram for explaining a ladder pattern provided on a platen glass. In the figure, a ladder reference plate 42 and a white reference plate 43 are provided outside a document placement area 41 on a platen glass 40 on which a document is placed. The ladder reference plate 42 and the white reference plate 43 are read when power is turned on, when maintenance is performed, or at predetermined intervals. First, a principle for obtaining a sampling phase difference from data obtained by reading two ladder patterns will be described. If the original image A (x) is a periodic data, the document image A (x) can be in the n-th order frequency omega _n and phase [Phi _n expressed by the following equation. A (x) = C + Σ n [a n · cos (ω n · x + Φ n)]

In order to obtain a phase component Φ _k corresponding to a certain frequency component ω _k from the above equation, the following is performed. First, the cos and the sin of the phase Φ _n are obtained by solving the relational expression obtained by convolving the cos and the sin in the same manner as the method of obtaining the Fourier component of the frequency. ∫A (x) · cos (ω _k · x) · dx = ∫C · cos (ω _k · x) · dx + ∫Σ _n [an · cos (ω _n · x + Φ _n )] · cos (ω _k · x ) · dx = 0 + ∫1 / 2 · Σ n [an · {cos (ω n · x + ω k · x + Φ n) + cos (ω n · x-ω k · x + Φ n)}] · dx = 0 + 1/2 · a _k · cos (Φ _k ) ∫ · dx → cosΦ _k = 2 · [∫A (x) · cos (ω _k · x) · dx] / [a _k · ∫dx] ≡2 · A COS _k / [a _k・ ∫dx]

[0030] ∫A (x) · sin (ω k · x) · dx = ∫C · sin (ω k · x) · dx + ∫Σ n [a n · cos (ω n · x + Φ n)] · sin ( ω _k · x) · dx = 0 + ∫1 / 2 · Σ _n [a _n · {sin (ω _n · x + ω _k · x + Φ _n ) −sin (ω _n · x−ω _k · x + Φ _n )}]・ Dx = 0 + 1/2 ・ a _k・ sin (Φ _k ) ∫ ・ dx → sinΦ _k = 2 ・ [∫A (x) ・ sin (ω _k・ x) ・ dx] / [a _k・ ∫dx ] ≡2 ・ A SIN _k / [a _k・ ∫dx]

Using the above relational expression, a difference between phase components corresponding to the same spatial frequency component of the periodic data of two colors is obtained.
Period data corresponding to read data of R and G, AR
(X) and AG (x), the spatial frequency component ω
The tan of the phase difference component corresponding to _k is as follows. Tan (ΦR _k −ΦG _k ) = sin (ΦR _k −ΦG _k ) / cos (ΦR _k −ΦG _k ) = {sin (ΦR _k ) ・ cos (ΦG _k ) −sin (ΦG _k ) ・ cos (ΦR _k )} / {cos (ΦR _k ) · cos (ΦG _k ) + sin (ΦG _k ) · sin (ΦR _k )} = {A SIN _{k k} · A COSG _k −A SING _k · A COSR _k } / {A COSR _k A COSG _k + A SING _k A SINR _k } By taking arctan in the above equation, the phase difference for each color can be obtained.

In the above equation, A SIN _k = ∫AR (x) · sin (ω _k · x) · dx A SING _k = ∫AG (x) · sin (ω _k · x) · dx A COSR _k = ∫AR (x) ・ cos (ω _k・ x) ・ dx A COSG _k = ∫AG (x) ・ cos (ω _k・ x) ・ dx. Actually, the read digital data is discrete data, and thus the integration of the above equation is performed for each pixel.
Is replaced by

By dividing the phase difference data Φ _k by the spatial frequency ω _k , a color shift amount is obtained. When the amount of color misregistration is 4 μm, it is 12
This corresponds to 0.2 pixels of 00 dpi (about 48 dots / mm). By this calculation, the color misregistration amount with respect to the other two colors based on G among the three read colors is obtained. Since this amount is not affected by the change of the angle of the optical flat plate by about 0.1 degree, it can be obtained from one read scan data. This example is shown in FIG. In the illustrated example, R
In −G, the main scanning position is +0.12, 5 mm at 0 mm.
+0.10 at 0 mm, +0.01, 300 at 150 mm
It becomes -0.10 in mm. Similarly, in BG,
-0.22 at main scanning position of 0 mm, -0.1 at 50 mm
9, -0.03 at 150 mm, +0.19 at 300 mm
Becomes

Next, a sampling phase difference when the optical plate 3 is deflected is determined. In the same manner as described above, ladder pattern read data of the ladder reference plate 42 is collected for the G channel in two states where the optical flat plate 3 is inclined. For these two data, A1
(X) and A2 (x), the spatial frequency component ω
The tan of the phase difference component corresponding to k is: tanΦ _k == {A SIN1 _k · A COS2 _k −A SIN2 _k · A COS1
_k } / {A COS1 _k · A COS2 _k + A SIN2 _k · A SIN1 _k }. By taking this arctan, the phase difference Φ _k between the two sampling data can be obtained.

In the above equation, A SIN1 _k = ∫A1 (x) ・ sin (ω _k・ x) ・ dx A SIN2 _k = ∫A2 (x) ・ sin (ω _k・ x) ・ dx A COS1 _k a = ∫A1 (x) · cos ( ω k · x) · dx a COS2 k = ∫A2 (x) · cos (ω k · x) · dx. By dividing the phase difference data Φ _k by the spatial frequency ω _k , the pixel shift amount of the two samplings is obtained (sampling pitch on the original surface: 1200 dpi conversion).

Further, since a phase shift of 0.5 pixel is originally superimposed at the intended 600 dpi on the above-mentioned two sampling phase differences, a sampling position shift amount is obtained by subtracting this amount. This example is shown in FIG.
Shown in In the illustrated example, the main scanning position is 0 mm and 1.10
For pixels, 50mm for 1.00 pixels, 150mm for +
There is a shift of 1.10 pixels for 0.92 pixels and 300 mm.

In this case, the inclination amount is set to a small value in advance in consideration of the "property of the displacement of the sampling position due to the inclination of the optical plate 3 being larger in the peripheral portion than in the central portion", and correction is made in advance. The absolute value of the quantity is prevented from becoming too large.

By adding the shift correction amounts (the color shift amount and the sampling error amount) obtained in these steps, the correction amount of the centroid correction amount table 15 can be set. As an example, the correction amount α of the main scanning position 0 mm in this is
In the calculation process, first, one pixel, which is a predetermined phase shift, is subtracted from 1.10, which is the sampling error amount, to obtain 0.10 pixels. In correcting the center of gravity for the displacement of the sampling position for this pixel, the memory M0
The data of the first scan and the data of the second scan stored in the first scan data are different from each other in the sampling position in the direction away from each other.
The data of the second scan is corrected by 0.5 pixel for -0.5 pixel. These are the values of α _GO and α _GE without the chromatic aberration correction. On the other hand, α _RO, the value of α _RE, this α _GO, α
_GE plus chromatic aberration correction +0.12, α _RO = +
0.17, α _RE = + 0.12.

The main scanning position is 150 mm (near the optical axis).
In the case of the correction amount α, the data of the first scan and the data of the second scan deviate from each other in the sampling position in the approaching direction, so that the data of the first scan is
The data of the second scan is corrected by 0.04 pixels by +0.04 pixels. With respect to other data, as shown in FIG. 13, a correction amount can be obtained.

Based on the value of α thus obtained, a calculation is performed by the correction circuit 16 shown in FIG. 9 to correct the center of gravity. For example, when α _RO = + 0.17, the sign is positive, so that the output of the reference pixel is 1−0.17 = 0.83 times, 1
The center of gravity is corrected by multiplying the output after the pixel by 0.17 and adding the result. Here, as an example, the correction amounts of both (FIGS. 11 and 12) are simply added, but since the correction amounts are different between the even-numbered phase and the odd-numbered phase, the blurring effect becomes unbalanced. , 600 dpi pitch sampling period components may remain. In order to more strictly suppress the even-numbered phase and the odd-numbered phase, it is only necessary to perform correction separately for a component in which the even and odd correction coefficients have positive and negative symmetry and a component in which the even and odd are the same, as in the G component.

For example, assuming a case where the main scanning of R is 0 mm, the chromatic aberration is corrected to the plus side by 0.12 pixels even and odd, and the sampling is corrected to ± 0.05 pixel, D ″ R1 _{2i + 1} = 0.95 D'R1 _{2i + 1} +0.05 D'R2 _{2i + 2} = 0.95 (0.88 DR1 _{2i + 1} +0.12 DR2 _{2i + 2} ) +0.05. (0.88 DR2 _{2i + 2 +} 0.12 • DR1 _{2i + 3} ) = 0.836 • DR1 _{2i + 1 +} 0.158 • DR2 _{2i + 2 +} 0.006 • DR1 _{2i + 3}

[0042] _{D''R1 2i = 0.95 · D'} R2 2i +0.05 · D'R1 2i-1 = 0.95 · (0.88 · DR2 2i +0.12 · DR1 2i + 1) +0.05 · (0.88DR1 2i _{-1 +} 0.12.DR2 _2i ) = 0.842.DR2 _{2i + 1 +} 0.114.DR1 _{2i + 1 +} 0.044.DR2 _{2i + 2} It is also possible to correct the effect without making the effect unbalanced.

2. Second Embodiment As described in the first embodiment, when a normal imaging lens 2 is used in the imaging optical system, the amount of image position shift Δ on the line CCD sensor 1 shown in FIG. = D
(Tanψ−tanψ ′) · cos θ. Also,
When two readings are performed on one document, if the sampling position in the main scanning direction is changed by changing the amount of inclination of the optical plate 3, the image forming position shift amount due to an error in mechanical scanning or the amount of tilt. An error is superimposed on Δ. If the sampling position shift becomes 0.4 pixels due to the superposition of such errors, for example, 600 dpi (23.6 dots / m2)
m), a moire component of about 4 lines / mm was generated due to the difference from the spatial frequency of the original. In the first embodiment described above, the shift of the sampling position due to the change of the angle of view is corrected by using a correction circuit that corrects the center of gravity based on the image data at different image formation positions, so that the occurrence of moiré is generated. Was reduced. On the other hand, in the second embodiment, the occurrence of moire is reduced by devising an imaging optical system.

FIG. 14 is a schematic diagram showing a partial configuration of an optical system of a document reading apparatus according to the second embodiment. In the optical system according to the second embodiment shown in this figure, the sampling position shift amount when the optical plate 3 is tilted is made equal at all angles of view.
A telecentric optical system in which the outgoing light is parallel light is employed. Also in the present embodiment, the drive system shown in FIG. 2 is used to adjust the amount of tilt of the optical plate 3, whereby the main scanning of the sampling position is performed in the first and second original readings. The direction is shifted by 画素 pixel.

Here, the telecentric optical system means an optical system in which either the entrance pupil or the exit pupil exists at infinity. This telecentric optical system is arranged by placing an aperture stop at an image space, an object space focal plane, or a position conjugate to them, or a collimator that converts the light emitted from the convex lens system into parallel light on the exit side of the convex lens system. This can be realized by using a combination lens in which lenses are arranged.

The imaging lens 2 'shown in FIG. 14 has optical characteristics set so that the principal ray of the emitted light is parallel to the optical axis. Therefore, the light beam emitted from the imaging lens 2 ′ is guided to the subsequent optical flat plate 3 under the same angle condition in the entire image length range. As a result, when the imaging light beam emitted from the imaging lens 2 passes through the optical flat plate 3, it is possible to make the imaging position shift amount in the entire area in the main scanning direction constant and uniform.

Further, even when the optical flat plate 3 is provided with the function of an infrared cut filter in which an interference film whose spectral characteristic has an angle dependence in principle is also provided, the spectral characteristic in the entire region in the main scanning direction is obtained. Can be constant and uniform. This is the same in the first embodiment. Thus, according to the second embodiment,
Since the amount of displacement of the imaging position in the entire area in the main scanning direction is constant, it is possible to eliminate the difference in the sampling position between the central portion and the peripheral portion on the line CCD sensor 1. As a result, it is possible to greatly reduce the moire caused by the sampling position shift without using an electrical correction circuit.

[0048]

As described above, according to the present invention, one original is read at least twice, and at the time of the two readings, the respective image forming positions are shifted to form the first image forming position. Is performed based on the image data at the second image forming position and the image data at the second image forming position different from this, the sampling position shift (image position shift) between the central portion and the peripheral portion caused by the angle of view is reduced. Since the correction is performed, it is possible to obtain an advantage that reading can be performed at a high resolution and occurrence of moiré due to a sampling position shift can be reduced. According to another aspect of the present invention, since an optical system that does not cause a shift in the sampling position between the central portion and the peripheral portion caused by the angle of view is used, similarly, in this case, it is possible to read at a high resolution. This has the advantage that the occurrence of moire caused by the sampling position shift can be reduced.

[Brief description of the drawings]

FIG. 1 is a schematic diagram illustrating a partial configuration of an optical system of a document reading apparatus according to a first embodiment of the present invention.

FIG. 2 is a conceptual diagram for explaining a drive system for tilting an optical flat plate.

FIG. 3 is a conceptual diagram showing an amount of movement of an imaging position on a CCD sensor due to an inclination of an optical plate.

FIG. 4 is a conceptual diagram for explaining a scanning direction and sampling points.

FIG. 5 is a conceptual diagram showing a shift amount of an image forming position between a central portion and a peripheral portion when an angle of view is considered.

FIG. 6 is a conceptual diagram showing an example of the angle-of-view dependence of an imaging position due to the inclination of an optical plate.

FIG. 7 is a conceptual diagram for explaining image quality deterioration due to a sampling position shift.

FIG. 8 is a block diagram showing an overall configuration including a correction circuit for correcting the angle-of-view dependency of a sampling position shift.

FIG. 9 is a block diagram illustrating a configuration of a memory and registration correction unit.

FIG. 10 is a conceptual diagram illustrating a ladder pattern provided on a platen glass.

FIG. 11 is a diagram illustrating an example of a pixel shift due to chromatic aberration.

FIG. 12 is a diagram illustrating an example of a sampling position shift due to an inclination of an optical flat plate.

FIG. 13 is a diagram illustrating an example of a correction amount for a main scanning position.

FIG. 14 is a schematic diagram illustrating a partial configuration of an optical system of a document reading apparatus according to a second embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 CCD sensor 2, 2 'imaging lens 3 Optical plate 10 Clock generator 11 CCD driver 12 CCD 14 Counter 13a, 13b, 13c A / D converter 16a, 16b, 16c Memory & cash register correction part

 ──────────────────────────────────────────────────の Continued on the front page (72) Inventor Yuichi Ichikawa 2274 Hongo, Ebina-shi, Kanagawa Fuji Xerox Co., Ltd.

Claims (6)

[Claims] A photoelectric conversion unit that converts light reflected from the document into an electric signal; an image forming position changing unit that changes an image formation position of the reflected light with respect to the photoelectric conversion unit in a main scanning direction; Reading means for reading a document image formed on the photoelectric conversion means at a different image formation position by the image formation position change means; image data at a first image formation position read by the reading means; A correction unit configured to perform barycenter correction based on image data at a different second imaging position; and a changing unit configured to change a correction amount by the correction unit according to a position in the main scanning direction. Image reading device. 2. A photoelectric conversion means for converting reflected light from a document into an electric signal, and provided in an optical path from the document to the photoelectric conversion means, wherein a principal ray of emitted light is parallel to an optical axis. Having an optical characteristic, an imaging lens for imaging the reflected light on the photoelectric conversion unit, and an imaging position changing unit for changing an imaging position of the reflected light on the photoelectric conversion unit in a main scanning direction, An image reading apparatus, comprising: reading means for reading a document image formed on the photoelectric conversion means at different image forming positions by the image forming position changing means on the document. 3. The image forming position changing means is inserted into an optical path between the document and the photoelectric conversion means, and an optical flat plate having two parallel surfaces, and an angle changing means for changing an inclination angle of the optical flat plate. 3. The method according to claim 1, further comprising:
The image reading device according to claim 1. 4. The image reading apparatus according to claim 3, wherein the optical flat plate has a function of cutting off infrared rays. 5. The image forming position changing means includes:
5. The image according to claim 1, wherein a position shifted by １ ／ pixel in the main scanning direction from the first image forming position is set as a second image forming position. 6. Reader. 6. The image reading apparatus according to claim 1, wherein the amount of correction by the change unit is set based on a result of reading a reference pattern arranged along a main scanning direction.