JP3287722B2 - Image reading device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image reading apparatus for reading an image by an image sensor and converting the image into an electric signal.

[0002]

2. Description of the Related Art A conventional full-color reading system of this type includes a light source switching system, a prism decomposition system,
Although there are a filter switching method, an on-chip color filter method, and the like, the on-chip color filter method is considered to be optimal in terms of high-speed reading and color separation accuracy.

FIG. 11 shows a configuration example of a conventional color CCD linear image sensor of an on-chip color filter system.

In this conventional color CCD linear image sensor 1601, RGB on-wafer color filters are each provided with three CCD (charge coupled device) chips in parallel on the same wafer as shown in FIG. This is a three-line color CCD linear image sensor including sensors (chips) 1602 to 1604.

In FIG. 1, reference numeral 161 denotes a light receiving unit which performs photoelectric conversion according to the amount of incident light (only R is denoted by a reference numeral, and the same applies to G and B).
On the CCD sensor element of the light receiving section 161, R,
G and B color separation filters are arranged on-wafer.
At the head of the light receiving section 161, there is a light shield pixel section for arranging an aluminum mask on the light receiving section 161 to block incident light and always obtain an output in a dark state. 162, 1
63 is a transfer gate, is intended to be transferred to the CCD shift register 164, 165 in accordance with charges stored in the light receiving section 161 to the shift gate pulse phi _TG. The charges accumulated in the even-numbered pixels of the light receiving section 161 are transferred to the respective CCD shift registers 165 for the even-numbered pixels by the transfer gate 163, while the charges accumulated in the odd-numbered pixels of the light receiving section 161 are transferred to the odd-numbered pixels by the transfer gate 162. The data is transferred to each CCD shift register 164 for pixels.

The CCD shift registers 164 and 165 transfer the charge sent from the light receiving section 161 to the output section.
D transfer (complete transfer) and drive clock φ ₁ (φ _1R , φ
_{1FR, φ 1G, φ 1FG,} φ 1B, φ 1FB) and φ ₂ (φ _2R, φ
_{2FR, φ 2G, φ 2FG,} φ 2B, φ 2FB) by the two-phase driven.

Reference numeral 166 denotes an output gate for transferring electric charge to each CC.
From the D registers 164 and 165, the output capacitance units 167a, 1
67b. Reference numerals 167a and 167b denote output capacitance units for converting the transferred charges into voltages. 168a and 168b are two-stage source follower amplifiers for lowering the output impedance and preventing the output signal from being affected by noise.

The output capacitance units 167a and 167b and the source follower amplifiers 168a and 168b allow FDA (Floor).
aTakingDiffusionAmplifier).

OSAR, OSBR, OSAG, OSB
G, OSAB, OSBB are signal output terminals, φRAR, φ
RBR, φRAG, φRBG, φRAB, φRBB are reset pulse terminals, φ1R, φ1G, φ1B, φ2R,
φ2G and φ2B are CCD shift register clock terminals,
φTGR, φTGG, and φTGB are transfer gate clock terminals, and ODR, ODG, and ODB are source follower amplifier drain terminals.

In the color image sensor 1601 configured as described above, the light incident on the light receiving section 161 is converted into electric charge proportional to the amount of light, and this electric charge is converted by the shift gate pulse φ _TG into the CCD shift registers 165 and 164.
Are transferred to the even-numbered pixels and the odd-numbered pixels, respectively.
Is output to the FDA via the FDA, and the output capacitance unit 1 of the FDA
At 67a, 167b the charge output is converted to a voltage,
Next, a two-stage source follower amplifier 168aA167
b and each output terminal OSA, OSB.

[0011]

However, in the above conventional example, as shown in FIG. 12, the spectral transmittances of the respective color filters R, G, B and the spectral energy distribution of the light source shown in FIG. From the spectral transmittance of the infrared absorption filter of FIG.
The amount of charge generated from the photodiodes 1602, 1603, and 1604 is B-CCD <R-CCD <G-CCD
Finally, each CCD 1602, 1603, 16
B-CCD <R-CCD <G-
Sensitivity increases in the order of CCD, for example, R, G, B CC
D sensitivity is R: 2.1V / 1x. sec, G: 2.6V
/ 1x. sec, B: 0.86 V / 1x. sec.

Further, each CCD 1602, 1603, 16
The saturated output voltage of the normal CCD registers 164, 16
5 are equal because their sizes are the same.

When the three-line color CCD linear sensor is actually used, the configuration of the reading system is determined by the amount of light for obtaining an output voltage for obtaining a necessary S / N.

That is, the B-CCD 1604 having the lowest sensitivity.
Output voltage is determined by the amount of light at which the required S / N is obtained. For example, if the required S / N is 48 dB (256 levels) and the noise level from the CCD is 1 mV, the minimum output voltage is 256 mV.

Therefore, the output voltage of the B-CCD is 256 m
In the case of V, the output voltages of the R-CCD 1602 and the G-CCD 1603 are respectively 2.1 (V / 1 × sec) /0.86 (V / 1 × sec) × 256 mV ≒ 625 mV 2.6 (V / 1 × sec) ) /0.86 (V / 1 × .sec) × 256 mV ≒ 774 mV.

Here, assuming that the output voltage can be obtained at a certain accumulation time T ₁ (μsec), in order to increase the reading speed of the image reading apparatus, for example, a double reading speed is required. To obtain, the accumulation time T ₂ is T
₂ = T _1/2 becomes the output voltage of each CCD Each B-CD
D: 128 mV, R-CCD: 325 mV, G-CC
D: 383 mV and S / N of B-CCD 1604 is 4
8dB cannot be obtained.

In order to correct this, the amount of light must be doubled, but there is a problem such as a rise in the temperature of the apparatus, and it is not always necessary to double the amount. In order to construct the device, the amount of illumination light must be increased to such an extent that the problem of temperature rise does not occur, and the shortfall in this case must be reduced by using the S / N of the B-CCD. Therefore, a problem to be solved that causes deterioration of image quality occurs.

Therefore, a linear image sensor as shown in FIG. 16 has been proposed.

This is hereinafter referred to as Time Delay and
This is called an integration (TDI) method. The TDI has a plurality of lines of photoelectric conversion means of a linear image sensor, and sequentially combines output signals of the plurality of lines of photoelectric conversion means in synchronization with a reading speed of a scanner equipped with the line sensor, thereby obtaining photoelectric conversion means of the line sensor. It is possible to obtain an output signal which is several times as many as the number of lines of the conversion means.

In FIG. 16, reference numeral 1700 denotes the aforementioned TDI
It is a color CCD linear sensor that enables operation.
1, 1702 and 1703 are Red, Blue and Gre, respectively.
en CCD linear sensor units.

Reference numerals 1704a to 1704c denote linear photodiode arrays each having a red on-chip color filter.
To the output units 1718a and 1718c
CCD shift register for horizontal transfer to 18b
The reason why the CCD is divided into two CCDs 705a and 1705b and a photoresistor is to improve the reading speed of the color CCD linear image sensor.

Similarly, 1706a to 1706c, 170
Reference numerals 8a to 1708c denote linear photodiode arrays having blue and green on-chip color filters, respectively, 1707a and 1707b, 1709a and 1
Reference numeral 709b denotes a blue and green CCD shift register, and reference numerals 1719a and 1719b, and 1720b and 1720b denote blue and green CCD shift registers, respectively.

Reference numerals 1710 and 1712 denote shift gates SH1 and SH3 for temporarily storing charges generated in the linear photodiode array in synchronization with the reading speed of the scanner. Reference numerals 1711 and 1713 denote shift gates SH1 (17).
10), charges accumulated in the SH3 (1712) are transferred to the next-stage linear photodiode arrays 1704b and 1704c.
Shift gates SH2 and SH4 for transferring the charges from the shift gates SH1 (1710) and SH3 (1712) to the linear photodiode arrays 1704b and 1704c in order to combine the charges with the charges generated in.

Reference numeral 1714 denotes a linear photodiode 170
4c in synchronization with the reading speed of the scanner.
A shift gate SH5 for transferring data to the CD shift registers 1705a to 1705b;
(1714) through linear photodiode 1704
The electric charge transferred from c through the shift gates SG1 and SG2 of 1715 to 1716 is sequentially input to each pixel SG1.
The charge of the pixel corresponding to SG2 is transferred to the CCD shift register 1
705b and 1705a.

Reference numeral 1717 denotes a transfer gate TG1, which is a CCD shift register 1705a and 1705b, respectively.
It is for transferring electric charges between them.

Further, each of Blue and Green CCDs
The linear sensor units 1702 and 1703 are the same as the above-described Red CCD linear sensor unit 1701, and will not be described here.

Here, the color CCD linear sensor 1700
As described above, the charge integration direction and the charge transfer direction between the CCD and the shift register are performed only in the direction of the arrow.

FIG. 17 is a diagram showing a configuration example of a scanner 1800 equipped with the color CCD linear sensor 1700 of FIG.

The scanner 1800 is a scanner main body 1800
a and a document feeder 1800b.

Reference numeral 1810 denotes a platen glass on which an original is placed; 1805, a halogen lamp for exposing the original;
Denotes a first reflection mirror, which constitutes a mirror unit 1812.

Reference numeral 1803 denotes a second reflecting mirror, and 1804 denotes a third reflecting mirror, which constitutes a mirror unit 1813.

Reference numeral 1801 denotes a lens unit for reducing and forming an image of a reflected light from a document exposed and scanned by a halogen lamp 1805 on a color CCD linear sensor 1700. Reference numeral 1809 denotes a flow reading of the document using a document feeder 1800b. This is a platen glass for drift reading in the case of performing the following.

An original is mounted on a platen glass 1810, and the mirror units 1812 and 1813 are moved and scanned in the direction of arrow A (sub-scanning direction) by a stepping motor 1814 at a scanning speed of 2: 1 to read the original. In this case, the mirror units 1812 and 1813 start from the position indicated by the broken line.

The document scanner 1800b comprises the following.

Reference numeral 1806 denotes an input tray for originals;
Reference numeral 07 denotes a document pickup roller, reference numeral 1808 denotes a feed roller for feeding a document, and reference numeral 1811 denotes a discharge tray.

In this case, a document is placed on the input tray 1806 with its front side facing up. In the case of single-sided reading, the document is sent to a feed roller by a document tooth pick-up roller 1807 and fed by the feed roller according to the document reading timing. The reflected light image is condensed on the color CCD linear sensor 1700 through the mirror units 1812 and 1813 and the lens unit 1801 when it is conveyed in the direction of the dashed arrow and passes over the platen glass.

Next, in the case of reading a double-sided original, the original fed by the feed roller is transported in the direction of the solid line arrow, and the surface first passes through the reading position of the platen glass for flow reading and is read. Flip along the path,
The back side is read from a direction opposite to the reading direction when reading the front side, and the sheet is discharged to the sheet discharge tray 1811 in the same manner as when reading a single-sided original.

At this time, the color CCD linear sensor 170
The scanning direction of the formed image on 0 is the direction of arrow B when reading the front surface, and the direction of arrow C when reading the back surface.
In the case of the CD linear sensor 1700, since the integration direction of TDI is only a single direction, the color CCD linear sensor 170
Depending on the installation direction of 0, only one of the front and back sides can be read. Therefore, FIG.
The following color CCD linear sensors have been proposed.

Reference numeral 1300 denotes a color CCD linear image sensor, and 1301, 1302, and 1303 denote each Re.
d, Blue and Green CCD linear sensor units.

Numerals 1301a to 1301c denote linear photodiode arrays as photoelectric conversion means having red on-chip color filters, respectively.
03c, 1304a to 1304c are 1302a to 130
Like FIG. 2c, it is a linear photodiode array having blue and green on-chip color filters, respectively.

1305a-1305b, 1308a-1
Reference numeral 308b denotes a CCD shift register which is shift register means for horizontally transferring electric charges generated in the linear photodiode arrays 1301a to 1301c to the output units 109a to 109b and 130a to 130b which are reading means, and 1305a to 1305b. Are for forward (front) reading (in the direction of the solid line arrow in FIG. 18), 1308a to 1310
08b is the reverse direction) back side) for reading (in the direction of the dotted arrow in FIG. 18)
It is.

Also, 1305a to 1305b are 1303a
To the output units 109a to 1093c.
It is also a CCD shift register for horizontal transfer to b (for reverse reading of Blue).

Reference numerals 1306a to 1306b denote CCD shift registers for transferring electric charges generated in the blue linear photodiode arrays 1303a to 1303c to the output units 110a to 110b, and are for reading blue in the forward direction.

Further, it is a CCD shift register for simultaneously outputting the charges generated in the Green linear photodiode arrays 1308a to 1308c (for reading Green in the reverse direction).

Reference numerals 1307a to 1307b denote CCD shift registers for horizontally transferring electric charges generated in the linear photodiode arrays 1304a to 1304c of Green to the output units 111a to 111b.

Reference numerals 112, 113, and 114 denote first charge transfers for transferring charges generated in the linear photodiode array 1301a to the next-stage linear photodiode array 1301b and combining the charges with the charges generated in the linear photodiode array 1301b. Shift gate S
H1, SH2, and SH3, and the shift gates SH1 (112), SH2 (113), and SH3 (1
14) are sequentially operated to transfer the electric charges in the direction of the solid line arrow. However, when reading in the reverse direction, the shift gates SH1 to SH3 (11
The operation order of 2 to 114) is opposite to that in the forward reading.

That is, the shift gate SH3 (114) →
SH2 (113) → SH1 (112). 115-
117 is the linear photodiode array 1 in the same manner as described above.
Shift gates SH4 to SH6, which are first charge transfer means for transferring charges in a vertical direction for synthesizing charges generated between 301b and 1301c, respectively, and operate at the time of forward reading and reverse reading. The order is SH4 → SH
5 → SH6, SH6 → SH5 → SH4.

Reference numeral 118 denotes the linear photodiode array 1
The charge generated in 301c is transferred to the horizontal CCD shift register 1
A shift gate SH7 for transferring data to the scanning speeds 305a and 1305b in synchronization with the reading speed of the scanner;
Reference numeral 20 denotes switch gates SG1 and SG for sequentially transferring the charge of the linear photodiode array 1301c transferred by the shift gate SH7 (118) to the horizontal CCD shift registers 1305a and 1305b for each pixel.
2. The charge of the odd-numbered pixel is transferred to the horizontal CCD shift register 13050b by the switch gate SG1, and the charge of the even-numbered pixel is transferred to the horizontal CCD shift register 1305a by the switch gate SG2 (120).

Reference numerals 121 to 123 denote transfer gates TG1 to TG3 as second charge transfer means for transferring charges between the horizontal CCD shift registers 1305a and 1305b, and transfer charges between the linear photodiodes. As in the case of the transfer, the operation order is changed from TG1 → TG2 → TG3, TG3 →
The transfer direction can be switched between forward and reverse by changing the transfer direction as TG2 → TG1 (forward direction: solid arrow,
Reverse direction: dotted arrow).

Here, the horizontal CCD shift register 1305
Reference numerals a and 1305b denote two-phase driving, in which two shift register clocks φ1 and φ2 are alternately input as is generally known, and pulses are alternately input to the two CCD shift registers 105a and 105b. As a result, the potential of the VVD register changes and charges are sequentially transferred in the direction of the output section (here, the directions of 109a and 109b). In the transfer between the transfer gates TG1 to TG3 described above,
It is assumed that φ1 of the two shift register clocks is used.

Next, reference numeral 124 denotes a shift gate SH8.
The charge of the linear photodiode array 1301a is transferred to CC
The data is transferred to the D shift registers 1308a and 1308b in synchronization with the reading speed of the scanner. Reference numerals 125 and 126 denote switch gates SG1 and SG2 for sequentially transferring the charge of the shift gate SH8 to the CCD shift registers 1308a and 1308b for each pixel. The switch gate SG1 transfers the charges of the odd-numbered pixels to the CCD shift register 1308a. Then, the charges of the even pixels are transferred to the CCD shift register 1308b by the switch gate SG2.

Reference numerals 127 to 129 denote transfer gates TG
1 to TG3 for performing register transfer between the CCD shift registers 1308a and 1308b.
The operation is the same as described above.

FIG. 19 is a timing chart of the color CCD linear image sensor 1300 shown in FIG.

[0054] In reading the positive direction of (a), at the timing of T _1, SH7, SG1, SG2 becomes "H" linear photodiode array 1301c charge each switch gate SG1 (119), until SG2 (120) Transfer is, T SG1 is in the _second state "H" → "L", φ1 is "L" → "H", and the switch gate SG1 (11
The charge of 9) is transferred to the CCD shift register 1305a.

[0055] Next, T ₃ in φ1: "H" → "L ",
TG1: becomes "H", and charges are transferred from the CCD shift register 105a to the transfer gate TG1 (121).

[0056] Similarly, T ₄ in TG1: "H" →
"L", TG2: "L " → "H", in T ₅ TG
2: “H” → “L”, TG3: “L” → “H”,
The charge is transferred from the transfer gate TG1 (121) to TG2.
(122) and TG3 (123) are sequentially transferred.

At T ₆ , the switch gate SG 2 (1
20) is “H” → “L”, φ1 is again “L” → “H”
To the switch gate SG2 (120)
Is transferred to the CCD shift register 1305a, and the transfer gate TG3 (123) changes from "H" to "L".
The charge of (123) is transferred to the CCD shift register 1305b.

Also, the shift gate SH relating to the TDI operation
As shown in the figure, in the transfer from 1 to SH7, charges are transferred to the CCD shift registers 1305a and 1305b for the first time in the seventh cycle from SH1 to SH7, one stage at a time, one shift gate at a time, and output and read out.

In the case of the reverse transfer in (b), the shift gate SH8 (1) is used instead of the shift gate SH7 (118).
24) operates, and shift gates SH1 to SH6 (112)
To 117), transfer gates TG1 to TG3 (12
1 to 123) are the same as those in the case of the forward transfer except that the operation timing of the forward transfer is the reverse of that in the case of the forward transfer in (a).

The color CCD linear image sensor 1 is described above.
00 Red CCD linear image sensor unit 1301
Has been described, Blue, Green CCD
In the linear image sensor units 1302 and 1303, the operation of the same symbol part is the same, so that the description is omitted here. Each signal is output from a CPU (not shown).

Incidentally, a plurality of linear sensor arrays are provided as described above, and the charge outputs photoelectrically converted by each linear sensor array are sequentially added in synchronization with the reading timing to output an output signal having a number of stages of the linear sensor array. Get color CC
The D linear image sensor has a structure that allows the integration direction of the TDI to be both forward and reverse to enable reading in the forward and reverse directions, and a structure that shares the CCD shift register and output unit between colors. The CCD output signal output from each output unit of the CCD linear image sensor outputs a different color signal at the time of forward-direction reading and at the time of reverse-direction reading, so that the output level differs.

[0062]

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and in the image reading apparatus according to the first aspect, a plurality of lines of photoelectric conversion means for converting light into an electric signal. And a photoelectric conversion unit disposed between the lines of the plurality of lines of photoelectric conversion means, wherein the charge generated by the photoelectric conversion means of each line is combined with the charge generated by the photoelectric conversion means of the adjacent line, and the photoelectric conversion of the adjacent line is performed. A first charge transfer unit that transfers the charges to the conversion unit; and a first and a second unit that are disposed with the plurality of lines of the photoelectric conversion units interposed therebetween and transfer the charges generated by the plurality of lines of the photoelectric conversion units to an output unit. A first shift register disposed between the first shift register and a photoelectric conversion unit of a line closest to the first shift register among the plurality of lines; A second charge transfer means for transferring charges to said first shift register means from the photoelectric conversion means line closest to the means, the second among the plurality of lines
The second shift register means is arranged between the photoelectric conversion means of the line closest to the second shift register means and the second shift register means, and the photoelectric conversion means of the line closest to the second shift register means is connected to the second shift register means. And a signal terminal for inputting a drive pulse of the first to third charge transfer means on the same wafer, and a drive input to the signal terminal is provided. The direction of charge transfer by the first charge transfer means is changed by a pulse, and the second and third charge transfer means are selectively driven. When reading in the forward direction, the first charge transfer is performed by the drive pulse. The second charge transfer means is driven by driving the charge transfer means in the positive direction, and the second charge transfer means is driven by the photoelectric conversion means on the line closest to the first shift register means. Read the charges by transferring to the output unit via the first shift register means,
At the time of reading in the reverse direction, the first charge transfer means is driven by the drive pulse to transfer the charges in the reverse direction, and the third charge transfer means is driven to form a line closest to the second shift register means. A linear image sensor configured to read out by transferring the electric charge synthesized by the photoelectric conversion unit to the output unit via the second shift register unit; and an electric charge output from the output unit of the linear image sensor. Correcting means for correcting the black or white level of the signal; and controlling the correcting means to correct the black or white level of the charge signal based on correction data corresponding to the charge transfer direction of the linear image sensor. And control means.

[0063]

FIG. 1 is a diagram showing the configuration of an image reading apparatus according to an embodiment of the present invention. Reference numeral 1300 denotes a TDI-type color CCD linear image sensor of the present embodiment, the detailed configuration of which is the same as that of FIG. 1301, 1302, and 1303 are Re, respectively.
d, Blue, and Green CCD linear sensor units, each having two lines of output signals. Reference numeral 101 denotes an analog signal processing unit, which is a sample and hold (S / H) circuit 10 that samples and holds two CCD output signals.
1a, 101b, a voltage control amplifier (VCA) 101 for amplifying a CCD output signal sampled and held by the S / H circuits 101a, 101b to a desired level.
c, 101d, and 8 after multiplexing the signals amplified to the desired level by the VCAs 101c and 101d.
A / D converter 101e for converting a bit into a digital signal
And one for each color, but only Red is shown here. An interline distance correction unit 102 corrects the distance of the color CCD linear sensor used in the present embodiment because the CCD linear sensors of each color are physically separated from each other by 24 lines on the same chip. Things.

FIG. 2 is a detailed block diagram of the line-to-line distance correction unit 102. 2M-bit FIFOs 102a and 102b respectively provided in R and B as inter-line correction memories
Consists of The reason why G has no FIFO is that the R and B signals to be read in advance based on the G signal are equal to the distance between lines (R is 24 × 2 = 48 lines, B is 24 lines).
This is because it needs to be delayed.

The two FIFOs 102a and 102b are controlled by reset signals WRST1 and WRST2 of a write address counter of the FIFO, a write / read enable signal ENB, and a write / read clock CLK (not shown in FIG. 3 described later).

FIG. 3 shows a timing chart of these control signals. The R signal is delayed by 48 lines, and the B signal is delayed by 24 lines.
After the line delay, the data is read out to the output in synchronization with the input of the G signal.

Reference numeral 104 denotes a shading correction unit which is a correction means, and an inter-line distance correction unit 102 corrects the inter-line distances of R, G, and B and corrects the same state as when the same line is read relatively. A signal is input.

The shading correction unit 104 corrects the variation of the black level (output level when no light is incident) signal of the CCD (corrects to the 00H level by removing the offset) and the standard white plate. White level shading (light amount unevenness in main scanning direction of optical system and CCD
Output pixel variation) (normalized to FFH level)
Is done.

The R, G, and B signals that have been subjected to shading correction are then input to a masking correction unit 105.

Masking correction is performed by standardizing R, G, and B spectral characteristics (for example, NTSC R, G, and B) in which the total spectral characteristics determined by the spectral sensitivity characteristics of the R, G, and B of the CCD and the spectral energy distribution of the light source are standardized. G, B characteristics), and the following matrix operation is executed. Here, Ri, Bi, Gi are input R, G, B signals, Ro,
Bo, Go are R, G, B signals after masking correction, a1
1 to a33 are masking coefficients.

[0071]

[Outside 1]

The R, G, and B signals color-corrected by the masking correction unit 105 are subjected to image sharpening and smoothing processing by the filter processing unit 106 to adjust the image quality.

Here, for example, in the case of sharpening, a Laplacian filter using a matrix of about 3 × 3, 5 × 5, and 7 × 7 is used depending on the degree, and conversely, 3 × 3, 5 × 5, 7 A median filter using a matrix of about × 7 is used.

Each of the color signals subjected to the scaling process in the page memory unit 107 is input to the γ conversion unit 108, which performs γ correction corresponding to the output target. , And one table is composed of three 256-byte RAMs. In this embodiment, eight RAMs are used.

Next, each color signal whose image has been adjusted by the filter processing unit 106 is input to the page memory unit 107.

The page memory unit 107 has a page memory 107a of a maximum memory size of 66.3 Mbytes (A3 size) for each color, and its control is performed by an address controller 107b.

The page memory unit 107 includes a sub-scanning scaling unit 107c for performing scaling processing in the sub-scanning direction on image signals read from the page memory 107a according to the address controller 107b, and scaling processing in the main scanning direction. , And these are present in each color, but only R is shown here.

The sub-scanning magnification unit 107c performs interpolation calculation processing using 8-line input data, and the main-scanning magnification change unit performs interpolation calculation processing between eight peripheral pixels on the same line from a maximum of 800% to a minimum of 25%. A scaling process in the main-sub amount scanning direction in 1% units is executed.

Next, the analog processing section 101 will be described in detail. FIG. 4 is a detailed configuration diagram of the analog processing unit 101. The numbers in FIG. 4 that are the same as those in FIG. 1 are the same. The two CCD output signals are S / H
Sampled and held by the circuits 101a and 101b,
Only the signal change of D is separated. S / H circuit 101
The CCD signals in which only the signal change of the CCD is sampled and held in a and 101b are input to the VCAs 101c and 101d. In the VCAs 101c and 101d, the CCD output signal is amplified to a desired level, and
Output to By the way, VCA 101c and 101d
D / A converters 101h and 10h, each of which generates an output voltage level based on setting data of a CPU 401 as a control means.
The amplification factor can be changed according to the output voltage level 1i.

In the multiplexer 101f, the two CCD output signals are alternately multiplexed so as to be the same as the pixel arrangement of the CCD linear sensor and output as one CCD output signal.

The output signal of the multiplexer 101f is A /
Before being input to the D converter 101e, the buffer amplifier (B
VF) input to 101g, and A / D with low impedance
While outputting a CCD output signal to the converter 101e,
The offset level of the CCD output signal is adjusted according to the output voltage level of the D / A converter 101j so as to match the dynamic range of the A / D converter 101e.

The D / A converter 101j is the D / A converter 10
As in 1h and 101i, an output voltage level determined by the setting data set by the CPU 401 is generated to change the offset level of the buffer amplifier (BUF) 101g.

The detailed description of the BUF 101g is omitted here. Next, the A / D converter 101e converts the CCD output signal into an 8-bit digital signal and outputs it to the next-stage inter-line distance correction unit 102. Here, the CCD registers used to read out the charges of the R, G, and B CCD linear sensor units 1301, 1302, and 1303 of the color CCD linear image sensor 1300 and the output units thereof are R-CCD linear sensors in the case of reading in the forward direction. 130
1, CCD register 1305, output unit 109, G-CC
D linear sensor 1303, CCD register 1306, output unit 110, B-CCD linear sensor 1304, CCD
The register 1307 becomes the output unit 111. In the case of reading in the reverse direction, the R-CCD linear sensor 1301, CC
D register 1308, output unit 130, G-CCD linear sensor 1303, CCD register 1305, output unit 10
9, a B-CCD linear sensor 1304, a CCD register 1306, and an output unit 110. The analog signal processing unit 101 connected to each output unit outputs an output signal of a CCD linear sensor of a different color at the time of forward-direction reading and reverse-direction reading. Will be input. Therefore, the D / A converters 101h, 101i, and 10h that set the amplification level and offset level of VCA and BUF in the analog signal processing unit 101, respectively.
It is desirable that the setting data 1j have different values at the time of forward reading and the time of reverse reading.

Therefore, in this embodiment, the D / A converter 1
01h, 101i, and 101j are set data A1 /
A2, B1 / B2, and C1 / C2 are used, and two setting data are switched and used according to reading in the forward / reverse direction. Here, A1, B1, and C1 are setting data at the time of reading in the forward direction, and A2, B2, and C2 are setting data at the time of reading in the reverse direction. These setting data are backed up in a backup RAM 402 used together with the CPU 401, and sent to the above-mentioned D / A converters 101h, 101i, and 101j when the power of the image reading apparatus is turned on and when the image is read.
It is supplied from the backup RAM 402 through U401.

Next, the shading correction unit 104 will be described in detail.

The shading correction unit 104 shown in FIG.
The black correction circuits 1502a, 1502b, 150
2c, white level correction circuits 1503a, 1503b, 15
03c.

FIG. 6 shows a black correction circuit 1502a shown in FIG.
15 shows the configuration of 1502b and 1502c. Since these have the same configuration, the black correction circuit 1502a will be described as a representative. The black level output of each CCD sensor has a large variation between pixels when the amount of light input to the CCD sensor is small, and when this is output as it is and an image is output, for example,
When an image is printed by a printer, streaks and unevenness occur in the data portion of the image. Therefore, it is necessary to correct the output variation of the black portion, and the correction is performed by the circuit of FIG.

First, the forward reading mode is set (in the direction of the solid line in FIG. 17). Then, the output signal of the red linear image sensor 1301 is input to the black correction circuit 1502a.

That is, prior to the copying operation, the reading output of the CCD sensors 1301, 1302, and 1303 is input to the present circuit as a black level image signal without turning on the halogen lamp 1805. Black level one line of the black level image signal inputted to the B _in the terminal RAM1601
CPU 401 sets data in rat 1608, selects A in selector 1602, closes gate 1603, and opens gate 1604. That is, the black level image signal is supplied to the selector 1602 and the gate 16
04 and input to the RAM 1601.

At this time, CS of the RAM 1601 is set to C
Set by the PU 401.

On the other hand, an address input to the RAM 1601 includes an address counter 1 initialized by HSYNC inversion.
The selector 160 receives the count output of the selector 605.
6 is selected, and the black level image signal for one line from the gate 1605 is stored in the RAM 1601 in accordance with the address value from the address counter 1605 (the above is the black reference value capturing mode, and the black level image signal stored in the RAM 1601 is set). The image signal is black level data).

However, since the black level data taken into the RAM 1601 in this manner is very minute, it is easily affected by noise generated in the three-line CCD sensor 1301, the analog signal processing unit 101, and the like. If the data is used as the black correction data, the image of the black portion may become rough with much noise, which is not preferable.

Therefore, the black level RAM 160 shown in FIG.
The CPU 401 executes the calculation shown in the flowchart of FIG. 7 on the black level data captured in 1 to remove the influence of noise.

First, in step S601, the black level R
The CPU 401 closes the gates 1603 and 1604 with respect to the latch 1608 so as to take in the black level data stored at the addresses B ₁ to B _{4678 of the AM 1601} into the work register of the CPU 401.
09 is opened, and data is set so that B of the selector 1606 is selected. Thereby, the black level RAM 16
01 is accessed by the address from the address bus of CPU 401, the black level data (B ₁₎ through the work register and the data bus of the CPU401 ~ (B ₄₆₇₈₎ is read.

Next, in step S602, the black data is read into the RAM 1601 again. That is,
The CPU 401 applies a gate 1609 to the latch 1608.
Is closed, and gates 1603 and 1604 are opened. Further, data is set so that A of the selector 1606 is selected. As a result, the black level RAM 1601 stores the address counter 16
At step 05, the black level data is written to the RAM 1601.

In step S603, the black level RAM 16
The black level data (B ₁ ′) to (B ₄₆₇₈ ′) stored from address B ₁ to B _{4678 of No.} 01 are stored in step S 60.
The data is read into the work register of the CPU 401 in the same manner as in 2.

[0097] At step S604, the added data comrades at the same address of the black level data taken in the work register _{(B 1) ··· (B 4678} ) and _{(B 1 ') ··· (B} 4678'), Divide by the number of data 2 and temporarily store the result in the work RAM. Then, in step S605,
Close the gate 1603 and 1609, open the gate 1604, also data sets rats 1608 to select the B of the selector 1602 and 1606, again, the address B ₁ of the black level RAM1601 to B _4678, the data of working RAM Select 1602 and Gate 1
The data is written according to the address data from the CPU 401 via the 604. In this way, by averaging the black level data, the black level data from which noise has been removed is set in the RAM 1601.

In this embodiment, the corrected black level data is obtained by calculating the average of two different levels of the taken black level data. However, the accuracy of the corrected data can be further improved by further increasing the number of averages. Can be raised.

Next, when the reverse reading mode is set (in the direction of the dotted line in FIG. 17), the blue correction circuit 1502a supplies the blue
Of the linear image sensor 1302 is input.

Here, as in the case of the red image sensor 1301, black level data is fetched.
The RAM 1610 is selected by the data set in the latch 1608 by the CPU 401, the processing is performed in the same manner as in the flowchart of FIG.
610 is set.

After the fetching and correcting operation of the black level data as described above, the reading of the original image is executed. When reading a document image, the CPU 401
4 and 1609 closed, gate 1603 opened, and
Latch 1608 is selected so that A of selector 1606 is selected.
And the RAM 1601 or 1610 is set to the data read mode. Thereby, the black level R
Black level data AM1601 or 1610 is read for each pixel in accordance with the address value from the address counter 1605 through the gate circuit 1603, in synchronization with the input of an image signal to B _in the terminal to the B input of the subtracter 1607 Is entered.

Here, the black level RAM 1601 is selected and used for reading in the forward direction, and the black level RAM 1610 is selected and used for reading in the reverse direction.

Therefore, the output of the black correction circuit shown in FIG. 6 corresponds to the black level data DK (i) (i is a pixel address) read from the black level RAM 1601 or 1610 and the red (R) signal for the black level data DK (i). _in (i) -DK (i)
= R _out (i) for each pixel (black correction mode). Similarly, for green (G) and blue (B), black correction circuits 1502b and 1502 having the same configuration as in FIG.
At c, similar control is performed.

The respective color data subjected to the black level correction as described above enter white level correction (shading correction) units 1503a, 1503b and 1503c, respectively, and the white level correction (shading correction) is performed by the halogen lamp 1.
Based on the white data output from the CCD sensors 1301, 1302, and 1303 when the uniform white plate 1818 is illuminated by 805, the sensitivity variations of the illumination system, the optical system, and the sensor are corrected.

FIG. 8 shows a configuration example of the white level correction circuit. The basic circuit configuration is the same as that of the black correction circuit shown in FIG. 6. However, in the above-described black correction, correction is performed by a subtractor 1607, whereas in white correction, a multiplier is used instead of a subtractor. Is used.

That is, in this embodiment, as the initial value setting mode, a standard white paper (for example, designated copy paper) on the platen glass 1810 is placed and the reading level R _D1 of the standard white paper and the reading level R of the white plate 1504 are set. _S is measured by each CCD sensor for each pixel in accordance with the flowchart of FIG. 9, and is backed up in a backup memory (not shown) of the CPU 401.

[0107] Additionally, blank actually used, for example, placing a copy sheet on the platen glass, Figure a blank reading level R _D2 each pixel by the color CCD sensor 9
Is measured in accordance with the flowchart of FIG.

However, the white level correction units 1503a and 1503
03b and 1503c have the same configuration, operation, and the like, and the white level correction unit 1503a will be described as a representative.

FIG. 9 is a flowchart showing an example of a procedure for obtaining a measured whiteboard brightness value.

In this case, similarly to the black level correction unit, the forward reading mode is first set, and the white level correction unit 1503 is set.
The output signal of the red linear image sensor 1301 whose black level has been corrected is input to a.

In step S801, the halogen lamp 18
05 is turned on, and in step S802, the CPU 401
Waits for one second and waits for the halogen lamp 1805 to rise.

At step S803, the CCD sensor 13
The image data of the standard white plate 1818 output from 01 is stored in the correction RAM 1801 for one line via the selector 1802 and the gate 1804.

Next, in step S804, the halogen lamp 104 is turned off, and in step S805, the CPU 40
1 temporarily stores the white image data (R _Si ) = (W _i ) i = 1 to 4678 of the standard white plate stored in the correction RAM 1801 in the working RAM via the gate 1809.

In step S806, the white image data (R _Si ) of the standard white plate stored in the working RAM
= (W _i ) i = 1 to 4678, image center 2339
256 pixels before and after the pixel (2083 pixels to 2595
The average value of the output levels of a total of 512 pixels (pixels), that is, R _s = {(W ₂₀₈₃ ) + (W ₂₀₈₄ ) +... + (W ₂₅₉₅ )} / 512 is obtained and backed up to the backup RAM.

The read data R _S of the standard white plate is obtained by obtaining a pixel having the peak value of the entire image instead of the average value of the output levels of the pixels before and after the center of the image, and calculating the average value of the output levels before and after the pixel. Then, this may be used as the read data R _S of the standard white plate. In this case, the read data R _D2 standard blank read data R _D1 blank obtained by the following, using the region of the same area of the image used to calculate the read data R _S of the standard white plate.

Further, in step S807, the mirror unit 1 is moved to below the standard white paper on the platen glass 1810.
The halogen lamp 1805 is turned on again in step S808.

In step S809, the CPU 401 counts for one second. In step S801, the CCD sensor 1
The uniform white level standard white paper white image data output from 301 is stored in the correction RAM 1801 for one line via the selector 1802 and the gate 1804.

In step S811, the halogen lamp 18
05 turned off, at step S812, CPU 401 stores the white image data (W ₁₎ ~ address W ₁ to _W-4678 correction RAM1801 the (W ₄₆₇₈₎ into the working RAM.

Next, in step S812, the standard blank white image data (W
₁₎ ~ (W ₄₆₇₈₎ from 2083 pixel ～2595 average value of the output level of the pixel R _D1, i.e., the _{R D1 = {(W 2083)} + (W 2084) + ··· + (W 2595)} / 512 Calculate and store in backup RAM.

Next, in step S814, each working R
Brightness measurements of white plate from the image data R _S of the image data R _D and the standard white plate blank stored in AM, i.e., stored in the working RAM seeking C = R _S / R _D, back up.

Here, the output measurement value of the blank paper is represented by N (= type
255), and as described above, the white plate is a so-called gray having a lower reflection density than white paper, but since it is extremely difficult to reproduce a complete gray, the white plate has a certain color or has some discoloration. In this case, if this is used as a white plate and white correction is performed, the gray balance of the read image data will be lost. Therefore, assuming that the output balance of each of the R, G, and B CCD sensors when reading the white plate has been lost, a white plate brightness correction factor K (= type1) is set for each color as a coefficient for correcting the lost balance. By doing so, the above-described gray balance correction can be performed, and the output value N of the white paper and the whiteboard brightness correction rate K are also stored in the working RAM, and the whiteboard brightness measurement value C is stored.
Back up with.

Next, when the reverse reading mode is set in the same manner as in the black level correction section, the white level correction section 1503a causes the blue linear image sensor 1302 whose black level has been corrected.
Are output, and blank image data R _D ′, standard white plate image data R _S ′, white plate brightness measurement value C ′,
The white board brightness correction factor K 'is obtained and stored in the working RAM for backup.

Next, in the actual correction, first, when reading in the forward direction, prior to the copying operation or the reading operation, the exposure lamp 1805 is turned on to expose the white plate 1818, and the white image of uniform white level output from each CCD is output. Data is stored in the correction RAM 1801 for one line via the selector 1802 and the gate 1804.

Next, the correction coefficient of each pixel is calculated according to the flowchart of FIG.

That is, in step S1001, the correction RA
White image data stored to the address W ₁ to _W-4678 of M1801 to incorporate into the working registers of the CPU 401, CPU 401 closes the gate 1803 and 1804 to the latch 1808, and the gate 1809
Is opened, and data is set so as to select B of the selector 1806.

Thus, the correction RAM 1801 is stored in the CPU
Accessed by an address from an address bus 401,
White image data (W ₁ ) to (W ₄₆₇₈ ) are read into the work register of the CPU 401 via the gate 1809 and the data bus.

Next, in step S1002, the working RAM is used for the white image data (W ₁ ) to (W ₄₆₇₈ ) at the addresses W _{1 to} W ₄₆₇₈ which have worked on the work register.
The shading correction coefficient E _i , that is, E _i = (N × C × K) / W _i , using the assumed output value N, the white board brightness correction rate K, and the white board brightness measurement value C of the blank paper backed up by (I = 1 to 4678) is calculated for each pixel.

Then, in step S1003, the gate 1
803 and 1809 are closed, the gate 1804 is opened, and data is set in the latch 1808 so as to select B of the selectors 1802 and 1806.
The data of the working RAM is _written to the addresses W _{1 to} W _{4678 of} 1801 via the selector 1802 and the gate 1804 from the CPU 401 according to the address.

[0129] At the time the original image reading, in synchronization with the input image data D _i input into B _in the terminal, the correction coefficient E _i are input to the multiplier 1807 via a gate 1803 from the correction RAM1801, D ₀ = D _i An operation of × E _i is performed for each pixel and output as corrected data.

Similarly, when reading in the reverse direction, the correction RAM
1802 is selected and correction is performed.

As described above, in this embodiment, a stable output signal can be obtained by performing appropriate shading correction of the output signal regardless of the reading direction, that is, the addition direction of the electric charges. Was.

[0132]

As described above, according to the present invention, it is possible to appropriately correct the black level or the white level according to the charge transfer direction in the linear image sensor.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an image reading apparatus according to an embodiment of the present invention.

FIG. 2 is a configuration block diagram of an interline distance correction unit according to the embodiment of the present invention.

FIG. 3 is a timing chart of line distance correction according to the embodiment of the present invention.

FIG. 4 is a configuration block diagram of an analog signal processing unit according to the embodiment of the present invention.

FIG. 5 is a configuration block diagram of a shading correction unit according to the embodiment of the present invention.

FIG. 6 is a block diagram illustrating a configuration of a black correction circuit according to an embodiment of the present invention.

FIG. 7 is a flowchart illustrating the operation of the embodiment of the present invention.

FIG. 8 is a configuration block diagram of a white level correction circuit according to the embodiment of the present invention.

FIG. 9 is a flowchart illustrating the operation of the embodiment of the present invention.

FIG. 10 is a flowchart illustrating the operation of the embodiment of the present invention.

FIG. 11 is a configuration diagram of a conventional linear image sensor.

FIG. 12 is a diagram illustrating a spectral transmittance of a color filter.

FIG. 13 is a diagram showing a spectral energy distribution of a light source.

FIG. 14 is a diagram showing a spectral transmittance of an infrared absorption filter.

FIG. 15 is a diagram illustrating spectral characteristics of a conventional linear image sensor.

FIG. 16 is a configuration diagram of a TDI-type linear image sensor.

FIG. 17 is a configuration diagram of an image reading apparatus using a TDI-type linear image sensor.

FIG. 18 is a configuration diagram of a TDI-type linear image sensor capable of transporting charges in forward and reverse directions.

FIG. 19 is a timing chart of a TDI-type linear image sensor capable of transferring charges in forward and reverse directions.

[Explanation of symbols]

 104 Shading correction unit 1300 CCD linear image sensor

──────────────────────────────────────────────────続 き Continuation of front page (72) Inventor Takahashi Tadashi ▼ 3-30-2 Shimomaruko, Ota-ku, Tokyo Inside Canon Inc. (56) References JP 4-133571 (JP, A) JP JP-A-2-226966 (JP, A) JP-A-63-199570 (JP, A) JP-A-7-221929 (JP, A) JP-A-7-221920 (JP, A) JP-A-5-167843 (JP JP-A-60-123159 (JP, A) JP-A-61-198859 (JP, A) JP-A-1-264058 (JP, A) JP-A-62-122377 (JP, A) (58) Field surveyed (Int.Cl. ⁷ , DB name) H04N 1/024-1/207 H04N 1/40-1/409 H04N 1/46 H04N 1/60 H04N 5/30-5/335 G06T 1/00

Claims (6)

(57) [Claims] 1. A plurality of lines of photoelectric conversion means (1301a, 1301b, 1301c) for converting light into an electric signal.
And arranged between the lines of the plurality of lines of photoelectric conversion means, and the photoelectric conversion means of the adjacent line for combining the charge generated by the photoelectric conversion means of each line with the charge generated by the photoelectric conversion means of the adjacent line. The first charge transfer means (112, 113, 114, 115, 11) for transferring to the conversion means.
6, 117), and first and second shift register means (1305) interposed between the plurality of lines of photoelectric conversion means and transferring charges generated by the plurality of lines of photoelectric conversion means to an output unit.
a, 1305b, 1308a, 1308b) and a photoelectric conversion unit (1301c) of a line closest to the first shift register unit in the plurality of lines, and the first shift register unit, Second charge transfer means (118, 119, 1) for transferring charges from the photoelectric conversion means (1301c) on the line closest to the first shift register means to the first shift register means.
20), the second shift register means being disposed between the photoelectric conversion means (1301a) of the line closest to the second shift register means in the plurality of lines and the second shift register means; Charge transfer means (124, 125, 1) for transferring charges from the photoelectric conversion means (1301a) of the line closest to the second shift register means to the second shift register means.
26) and a signal terminal for inputting a drive pulse of the first to third charge transfer means on the same wafer, wherein the first charge transfer means is driven by the drive pulse input to the signal terminal. And selectively drives the second and third charge transfer means, and drives the first charge transfer means with the drive pulse to move the charge in the positive direction when reading in the forward direction. And the second charge transfer means is driven to drive the photoelectric conversion means (1301) of the line closest to the first shift register means.
c) reading by transferring the combined electric charge to the output unit via the first shift register means, and in the case of reading in the reverse direction, driving the first charge transfer means by the drive pulse to transfer the electric charge. The photoelectric conversion means (1301) of the line closest to the second shift register means by transferring the charge in the reverse direction and driving the third charge transfer means.
a) a linear image sensor (1300) configured to read out by transferring the electric charge synthesized in a) to the output unit via the second shift register means, and output from the output unit of the linear image sensor; Correcting means (104) for correcting a black level or a white level of the charge signal; and correcting the black level or the white level of the charge signal based on correction data corresponding to a charge transfer direction of the linear image sensor. An image reading apparatus comprising: a control unit (401) for controlling a unit. 2. The first and second shift register means include odd pixel shift register means (1305b, 1308a) for transferring odd pixel charges and even pixel shift register means (1305b, 1308a) for transferring even pixel charges. 1305a,
1308b) is provided in each case.
The image reading device according to claim 1. 3. The image reading apparatus according to claim 1, wherein a color filter for color-separating light incident on the photoelectric conversion unit is formed on the photoelectric conversion unit. 4. An image reading apparatus according to claim 1, wherein said shift register means is a CCD shift register. 5. The image reading apparatus according to claim 1, further comprising a moving unit that moves a light image of a document relative to the linear image sensor. 6. The linear image sensor is a TDI
(Time Delay and Integrati
The image reading apparatus according to claim 1, wherein the image reading apparatus is of an on-type.