JP2632871B2 - Image reading device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to an image reading apparatus that reads an image with a line sensor.

(Prior art)

An apparatus has been proposed in which a color original image is color-separated, photoelectrically read by an image sensor such as a CCD, and a color image is printed based on the read output.

In such an apparatus, for color separation reading of a color original image, blue (B), green (G), and red (R) filters for color separation are sequentially inserted into the optical path from the original to the image sensor. Then, the light image passing through each filter is sequentially read by an image sensor, and R, G, B filters are attached to each light receiving element of a line image sensor including a plurality of light receiving elements. An image reading apparatus that separates and reads the color image at a time is used.

However, in the former configuration, a three-color signal cannot be obtained by one-time image reading. For example, when performing a masking process or the like on the three-color signal, a memory for storing at least two-color signals for one screen. Had to be provided. Further, in the latter configuration, it is possible to obtain a three-color signal by one image reading, but since images of different colors are read by adjacent light receiving elements, crosstalk occurs between different color signals and the Occasionally, it was smeared.

〔Purpose〕

The present invention has been made in view of the above points, and an object of the present invention is to provide an image reading apparatus capable of performing good image reading without providing a large amount of memory.
More specifically, a plurality of line sensors arranged in parallel at predetermined intervals to read different lines on a document, and image signals output from at least one of the plurality of line sensors are read by the plurality of line sensors. Delay means for delaying and outputting the number of lines corresponding to the integer part of the number of pixels corresponding to the amount of positional shift, and for image signals of two adjacent lines output from at least one of the plurality of line sensors Interpolating means for performing an interpolation process to form an image signal of a virtual line between the two adjacent lines, and a coefficient for the interpolation processing by the interpolating means to a shift amount of the reading position of the plurality of line sensors. Setting means for setting according to the decimal part of the corresponding number of pixels, and to provide an image reading apparatus for correcting a shift of the reading position of the plurality of line sensors. .

〔Example〕

 Hereinafter, the present invention will be described using preferred embodiments.

FIG. 2 shows an example of a color document reading apparatus (hereinafter referred to as a color scanner) to which the present invention is applied. Document cover 900
9 placed on the platen glass 901
Color-separated image sensor for color-separating and reading 02 image information
903 is used, and reflected light from the original 902 exposed by the light source 904 is imaged on the image sensor 903 by the lens 908 via mirrors 905, 906, and 907. Reference numeral 912 denotes a sensor shift detection plate, and as shown in FIG. 7A, a black straight line 913 having an angle α with respect to the main scanning direction of the image sensor 903 is drawn on a uniform white background. It is arranged to be read. A light source unit 910 including a light source 904 and a mirror 905 and an optical unit 911 including mirrors 906 and 907 move at a relative speed of 2: 1. These optical units 910,
The step 911 is reciprocated by a stepping motor 909 at a constant speed in a direction indicated by an arrow (sub-scanning direction) perpendicular to the main scanning direction.

FIGS. 3A and 3B are front views of a color separation image sensor 903 (hereinafter, three-line CCD 903) used in the present embodiment, and a line sensor to which a red filter is applied as shown in FIG. 3A. Line sensor CCD with green filter on CCD1005
Line sensor CCD1003 with 1004 and blue filter
Are formed in parallel on the same chip 1002. The number of pixels of each CCD is 5000 pixels.
It can be read at ts / inch resolution. FIG. 3B is an enlarged view of the sensor section 1006 in FIG. 3A, in which the width of one pixel is 10 μm and the distance between the sensors is 180 μm.

Since one pixel of 400 dots / inch is 63.5 μm on the platen, the lens 908 reduces the document information to 1 / 6.35 and 3 lines C
Project onto CD903. Therefore, the distance between the CCDs of the three lines attached as shown in FIG. 2 is 180 μm, and the red reading line R and the green reading line G on the platen as shown in FIG.
And a distance a between each of the blue reading lines B, 1.14
3 mm. The three-line CCD 903 is driven to read the original in the sub-scanning direction at a resolution of 400 lines / inch, and 1.143 mm on the original table is shifted by 18 lines. As described above, according to the present embodiment, since a color image is read using a three-line CCD as shown in FIG. 3, each color signal can be obtained independently, and no color smearing occurs.

However, a color scanner generally outputs color separation signals of the same line on a document as RGB signals standardized by, for example, the NTSC system.
The reading position shift of 18 lines in the sub-scanning direction of 3,1004,1005 must be corrected. If this shift is not corrected well, for example, when a color image is reproduced using the RGB signals, a color shift may occur in the reproduced image, or unnecessary color reproduction may be performed in the edge portion of the image. There are inconveniences.

That is, in this embodiment, when scanning in the direction of the arrow in FIG. 4, reading of the R line by the CCD 1005 precedes, then reading of the G line by the CCD 1004 reaches the same position after 18 lines, and further reading of the B line by the CCD 1003 after 18 lines. The reading reaches the same position. Therefore, the read information of the R line is delayed by 36 lines, the read information of the G line is delayed by 18 lines, and the information of the three lines is aligned with the read information of the B line.

In addition, the deviation is corrected by using an interpolation method for a deviation equal to or less than a line unit.

FIG. 5 is a block diagram of a driving and signal processing circuit for the three-line CCD 903.

In FIG. 5, signals read by three lines of CCDs 1005, 1004, and 1003 are amplified by amplifiers 1104, 1105, and 1106, respectively.
The converters 1107, 1108, and 1109 convert the digital signal into 8-bit 256-step digital signals. Reference numeral 1113 denotes a clock generator which outputs two-phase clocks 1116 and 1117 for driving the three-line CCD 903, a pixel clock 1119, and an HSYNC signal 1118 which is a line synchronization signal.

Reference numeral 1110 denotes an n-line delay memory which outputs the R signal 1120 output by the CCD 1005 and converted as a digital signal by delaying n lines, and 1111 delays the G signal 1121 output by the CCD 1004 and converted as a digital signal by m lines. The output is an m-line delay memory.

Reference numeral 1114 denotes an interpolation calculator for correcting a shift in the pixel interval between the CCD 1005 and the CCD 1003, and reference numeral 1115 denotes an interpolation calculator for correcting a shift in the pixel interval between the CCD 1004 and the CCD 1003.

1125 is a control unit (CP
U), the pixel clock 1119 and the HSYNC signal 1118 are input from the clock generation unit 1113 to the CPU 1125, and the outputs of the A / D converters 1107, 1108, and 1109 are input.

Then, a sensor shift detecting plate 912 is provided by a three-line CCD 903.
, The interval between the three CCDs is detected based on the output obtained by reading the data, and the operation of the interpolation calculators 1114 and 1115 is controlled according to the detection result. The CPU 1125 also controls the operation of the entire document reading apparatus, such as turning on the light source 904 and driving the motor 909, in addition to this operation.

256-stage R signal 1123 output from interpolation calculator 1114
And the 256-level G signal 112 output from the interpolation calculator 1115
4 and the B signal 1122 in 256 stages from the A / D converter 1109 are input to the r, g, b inputs of the masking circuit 1112, respectively, and the NTSC in 256 stages is calculated by the following masking matrix operation.
R-G, B-signals of system R-NTSC1128, G-NTSC1129, B-N
Converted to TSC1130.

Ie Becomes

The masking circuit 1112 that performs this matrix operation is the sixth
The configuration is as shown in the figure.

That is, the R signal input to the r input is applied to the multipliers 201, 202, 20
3, and are multiplied by multiplication coefficients a ₁₁ , a ₂₁ , and a ₃₁ , respectively. The G signal input to the g input is a multiplier 204,
205 and 206 are input in parallel, and the multiplication coefficients a ₁₂ , a ₂₂ and a
_The B signal multiplied by ₃₂ is input to the multipliers 207, 208, and 209 in parallel. The multiplication coefficients a ₁₃ ,
It is multiplied with a _23, a _33. The outputs of multipliers 201, 204, and 207 are added by adder 210, and the outputs of multipliers 202, 205, and 208 are converters.
The addition is performed at 211, and the outputs of the multipliers 203, 206, and 209 are added at the adder 212, whereby R-NTSC, G-NTSC, and G-NTSC are formed from the adders 210, 211, and 212.

Each multiplication coefficient is obtained from the correspondence with the equation (1). Becomes

FIG. 7 shows the sensor displacement detection plate 912 shown in FIG.
FIG. 9 is a diagram showing a relationship between each of CCDs 1003 to 5 of CD903 and each output state of each of CCDs 1003 to 5.

As described above, reference numeral 912 denotes a sensor shift detection plate on which a black straight line 913 having an angle α with respect to the main scanning direction is drawn, 1003 denotes a CCD having a B filter, and 1004 denotes a CC having a G filter.
D and 1005 are CCDs having an R filter.

The sensor shift detecting plate 912 is placed at the same height as the platen glass surface 901 and its longitudinal direction is parallel to the main scanning direction.

 Now define as follows.

L: Width in the main scanning direction on the document table corresponding to the length of CCD1003 to 5 (unit: mm) P _BG : Black pixel detection interval between CCD1003 and CCD1004 (unit pixel) P _BR : Black pixel detection interval between CCD1003 and CCD1005 (unit pixel) C: CCD number of pixels G _BG: CCD 1003 as the spacing CCD1004 (unit Meri meters) G _BR: CCD 1003 as the spacing CCD1005 (unit Meri meter) these relationships are as follows.

G _BG : tanα × L × P _BG / C Equation (3) G _BR : tan α × L × P _BR / C Equation (4) If the width of the document read by the CCD is 1 (unit: millimeter), g _BG C = G _BG / l Equation (5) g _BR = G _BG / l Equation (6) g _BG : Interval between CCD1003 and CCD1004 (unit pixel) g _BR : Interval between CCD1003 and CCD1005 (unit pixel) L, C, tanα, l is a value that can be known in advance.
Therefore, if P _BG and P _BR are detected, three CCD1s placed in parallel
Pixel intervals g _BG and g _BR of 003 to 1005 can be known.

These P _BG and P _BR are detected by the CPU 1125 as described above.
That is, when the sensor shift detecting plate 912 is read by the three-line CCD 903, outputs from the CCDs 1003, 1004, and 1005 are as shown in FIGS. 7 (E), (F), and (G). Accordingly, the CPU 1125 monitors the outputs from the A / D converters 1107, 1108, and 1109, and determines at what pixel from the start of reading of the sensor shift detection plate 912 the output 914, 915, 916 corresponding to the black line 913 has occurred.
Detection is performed for each of the CCDs 1003, 1004, and 1005.

If it is determined that there is an output corresponding to the black line 913 at each of the x, y, and z pixels of the CCDs 1003, 1004, and 1005, P _BG and P _BR are obtained by the following calculation. .

P _BG = y−x Equation (7) P _BR = z−X Equation (8) FIG. 1 n-line delay memory 1110, m-line delay memory 11
FIG. 11 is a block diagram of 11 and interpolation calculators 1114 and 1115.

403 to 406 are selectors for selecting line image data stored in the delay memories 1110 and 1111. 407 to 410 multiply three image data output from the selectors 403 to 406 by a multiplication coefficient set by the CPU 1125.

411 and 412 are adders for adding the outputs of the multipliers 407 and 408 and the multipliers 409 and 410, respectively. The coefficients of the selectors 403 to 406 and the multipliers 407 to 410 are set by the CPU 1125.

In the interpolation calculations 1114 and 1115 of the present embodiment, the deviation of the pixel interval is corrected by linear interpolation represented by the following equation.

D (N _A + N _B ) = (1−N _B ) × D (N _A ) + N _B × D (N _A +
1) (Equation 9) D (n) = n-line pixel value N _A : integer part N _B : decimal part FIG. 8 is a processing flow chart of CPU 1125 for setting an interpolation calculation coefficient for correcting a line pixel shift. . In step 501, the optical units 910 and 911 are set at positions where the three-line CCD 903 reads the sensor shift detecting plate 912. Step 502 is
The sensor shift detection plate 912 is read by the three-line CCD 903, and the black pixel positions x, y, and z are detected as described above.
In step 503, using the equations (7) and (8) from the detected black pixel positions x, y, and z of each CCD, P _BG = (interval of black pixel detection positions of B sensor 102 and G sensor 103: unit pixel) P _BR = (interval of black pixel detection position between B sensor 102 and R sensor 104: unit pixel) is calculated.

Step 504 is based on equations (3) to (6), and P _BG , P
_{From BR} , _gBG : (interval between B sensor 102 and G sensor 103: unit pixel) _gBR : (interval between B sensor 102 and R sensor 104: unit pixel) is calculated.

In step 505, g _BG : A _BG + B _BG (A _BG : Integer part B _BG : Decimal part) Equation (10) g _BR : A _BR + B _BR (A _BG : Integer part: B _BG decimal part) Equation (11) The selector and multiplier of FIG. 1 are set as follows.

The selector 406 is set so that line data of the _ABG line of the delay memory 1111 of the G signal flows to the multiplier 410.

Set a multiplier coefficient (1- _BBG ) in the multiplier 410.

The selector 405 is set so that the line data of the (A _BG +1) th line of the delay memory 1111 of the G signal flows to the multiplier 409.

• Set the multiplier 409 to the multiplier coefficient _BG .

The selector 404 is set so that the line data of the _ABR line of the delay memory 1110 of the R signal flows to the multiplier 408.

Set the multiplier coefficient (1-B _BR ) in the multiplier 408.

The selector 403 is set so that the line data of the (A _BR +1) th line of the delay memory 1110 of the G signal flows to the multiplier 409.

-Set a multiplier coefficient B _BR in the multiplier 407.

As described above, the interpolation circuits 1114 and 1115 that take in the output data from the CCD and form the interpolation data based on the data of the adjacent lines are provided. It is possible to form data of a typical line, and to set coefficients for this interpolation processing at the reading positions of three CCDs according to the amount of deviation.

This makes it possible to form virtual line data at positions where the displacements of the three CCD reading positions have been compensated, and therefore it is possible to electrically correct mechanical reading position deviations. Line R signal 1123, G signal 112
4, B signal 1122 can be input, and good matrix operation can be performed.

Second Embodiment Next, a second embodiment of the present embodiment will be described. Book second
In the embodiment, the optical units 910 and 911 are moved to improve the detection accuracy of the line sensor interval, and the pixel intervals read and detected a plurality of times are averaged while the reading position of the sensor shift detecting plate 912 is gradually changed.

FIG. 9 is a diagram showing a processing flow chart of the CPU 1125 in the second embodiment, which is a processing flow chart for setting an interpolation calculation coefficient for correcting a line pixel shift.

In step 601, the optical units 910 and 911 are
The position where D903 reads the sensor shift detection plate 912 is set. In step 602, the sensor shift detecting plate 912 is read by the three-line CCD 903, and the first black pixel position is read.
x ₁ , y ₁ , z ₁ are detected. Step 603 black pixel position x ₁ of CCD detected in, y _1, P _BG1 than z _1, to calculate the P _BR2.

In step 604, based on equations (3) to (6),
Calculate g _BG1 and g _BR1 .

Next, in step 605, the reading position of the sensor shift detecting plate 912 by the three-line CCD 912 is shifted by Δα.

In step 606, the number of times N for performing the processing of 602 to 605 is determined.
It is determined whether or not the process has been completed, and if not, the process returns to step 602 and sequentially proceeds to the first to Nth black pixel positions.
_{x 2, y 2, z 2} ~x N, y N, seek _{z N, g BG2, g BR2} ~g BGN, g
_Ask for _BRN .

In step 607, g _BG and g _BR obtained N times in step 604 are averaged.

AVR (g _BG ) = A _BG + B _BG [AVR (g _BG ): average value of g _BG A _BG : integer part B _BG : decimal part] Equation (12) AVR (g _BR ) = A _BR + B _BR [AVR ( g _BR ): average value of g _BR A _BR : integer part B _BR : decimal part] Expression (13) In step 608, based on the value obtained in step 607, the selector and multiplication in FIG. Set the container coefficient.

Third Embodiment Next, a third embodiment of the present invention will be described.

In the third embodiment, a black line drawn in parallel with the main scanning line direction in the sensor shift detection 912 is defined by three lines CCD9.
The sensor interval is obtained based on the position where the peak of the read black signal is given while shifting the reading position of 03, and thereby the line sensor interval shift is interpolated.

 FIG. 10 and FIG. 11 are views showing a third embodiment.

FIG. 10 shows a black straight line 9 drawn on the sensor shift detection plate 912.
FIG. 14 is a diagram showing a relationship between 14 and a CCD interval.

Reading position and CCD1004 where CCD1003 gives maximum black output
There reading position interval gives the maximum black output .DELTA.l _BG (Units Units millimeters), and CCD1003 detects the reading position distance reading position and CCD1005 give maximum black output gives the maximum black output .DELTA.l _BR (unit mm), the following equation To calculate the CCD intervals g _BG and g _BR .

g _BG = Δl _BG / l _e formula _{(14) g BR = Δl BR} / l e formula (15) g _BG: CCD1003 the spacing CCD1004 (unit pixel) g _BR: CCD 1003 and CCD1005 interval (unit pixel) le _e : Width of document read by each CCD in the sub-scanning direction (unit: millimeter) FIG. 11 is a processing flowchart of the CPU 1125 in the third embodiment. That is, in steps 701 to 703, the aforementioned Δl _BG and Δl _BR are obtained from the maximum black output of each CCD. And step 70
In step 4, g _BG and g _BR are calculated according to equations (14) and (15).
Further, in step 705, the coefficients of the selector and the multiplier shown in FIG. 1 are set as in step 505 described above.

In each of the above-described embodiments, the interpolation operation for interpolating the line pixel shift is a linear interpolation operation. However, other interpolation methods relating to the interpolation of image data, for example, the cubic. A convolution method or the like can also be used.

In this embodiment, three CCDs having R, G, and B filters are used.
Has been described, the present invention is not limited to this. For example, an image sensor having an R filter and an image sensor not having any The present invention is also applicable to a configuration for performing reading.

〔effect〕

As described above, according to the present invention, in a configuration in which different lines on a document are read by a plurality of line sensors arranged in parallel at predetermined intervals, a plurality of line sensors can be provided without providing a large amount of memory. Of the reading position can be satisfactorily corrected not only for the integer part of the number of pixels corresponding to the deviation amount, but also for the decimal part, whereby it is possible to suppress, for example, the occurrence of color deviation in reading a color image.

[Brief description of the drawings]

FIG. 1 is a block diagram of a delay memory and an interpolation calculator, and FIG.
3A and 3B show a three-line sensor, FIG. 4 shows a reading position of the three-line sensor, and FIG. 5 shows signal processing. FIG. 6 is a block diagram of the masking circuit, FIG. 7 is a diagram showing the relationship between the sensor shift detection plate and the CCD,
FIG. 8 is a flowchart showing the processing procedure of the CPU, and FIG.
FIG. 10 is a flowchart showing a processing procedure in the second embodiment of the CPU, FIG. 10 is a view showing the relationship between the sensor shift detecting plate and the CCD in the third embodiment, and FIG. 11 is a processing in the third embodiment of the CPU. It is a flowchart which shows a procedure, 903 is 3
Line CCDs 1003 to 1005 are CCDs, 1110 and 1111 are n-line delay memories, 1114 and 1115 are interpolation calculators, and 1112 is a masking circuit.

Claims (1)

(57) [Claims] A plurality of line sensors arranged in parallel at predetermined intervals to read different lines on a document; and a plurality of line sensors for outputting image signals output from at least one of the plurality of line sensors. Delay means for delaying the output by the number of lines corresponding to the integer part of the number of pixels corresponding to the shift amount of the reading position, and image signals for two adjacent lines output from at least one of the plurality of line sensors And an interpolation unit that performs an interpolation process on the image data to form an image signal of a virtual line between the two adjacent lines, and calculates a coefficient for the interpolation process by the interpolation unit by shifting a reading position of the plurality of line sensors. Setting means for setting according to a fractional part of the number of pixels corresponding to the amount, wherein a shift of a reading position of the plurality of line sensors is corrected. Place.