JPH01319367A - Color picture reader - Google Patents

Abstract

PURPOSE:To reproduce a read color picture with high accuracy by processing an object picture while processing it with shape and color separately and extracting the shape at a picture element density being twice that for the color picture discrimination. CONSTITUTION:A photodetection correction means 5 corrects a color to be discriminated in a way that the photodetection quantity (density) by the 1st photodetector 3a(i) and the photodetection quantity (density) by the 2nd photodetector 3b(i) are in the same degree and a shape extraction means 6 extracts the shape of a picture being a read object baled on the sensor photodetection state obtained by the correction. The processing by the shape extraction means 6 is applied based on the photodetection state of the 1st and 2nd photodetectors 3a(i), 3b(i). Thus, the shape information is obtained at a density twice that of the color picture information generated by the color picture information generating means 4 is obtained and implemented separately from the processing of the color picture discrimination, then even with the filter processing for the resolution correction relating to the shape, the basic data for discriminating the color picture is not changed. Thus, the color picture is reproduced with high accuracy.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a color image reading device applied to, for example, a two-color copying machine that targets two-color originals. A color image is created by using a sensor with two types of light-receiving characteristics whose relationship is determined in accordance with the image, and by performing color discrimination on a pixel-by-pixel basis based on the light-receiving needle of each sensor's light-receiving characteristics when scanning a document. The present invention relates to a color image reading device for obtaining information.

[Prior Art] Consideration of color reproducibility when copying a document such as a document where most of the image is occupied by a black image and there are also some red or blue underlines, marks, etc. Therefore, a so-called two-color copying machine that faithfully reproduces black and one other color is more suitable than a so-called full-color copying machine. By the way, since such a copying machine develops toner for each color in correspondence with each color image portion on the original, an image color image reading device is required to distinguish the image color on the original. However, conventional color image reading devices of this type are as follows.

The first is a first sensor that reads image information on a document using light transmitted or reflected by a member that selectively absorbs or reflects specific chromatic color components; a second image sensor that reads an image on the same document using light in a wavelength range that does not
This method compares the image signals from the first and second image sensors on the same part of the document and separates a specific chromatic color from other colors (black) (Japanese Patent Application Laid-Open No. 59-36478
(see publication).

The second method is to perform reading scans of the same document twice with different color characteristics by switching the color separation filters in one image sensor, and to obtain data through the image sensor in each scan. It compares image signals and separates specific chromatic color components (Japanese Patent Laid-Open No. 58-1
(See Publication No. 73963).

Each of the above uses sensors with two different light-receiving characteristics, and when scanning a document, color image information is obtained by performing color discrimination on a pixel-by-pixel basis based on the amount of light received by each sensor's light-receiving characteristics. This is what I did.

[Problems to be Solved by the Invention] The first method requires two image sensors and requires relatively high mechanical refinement of the optical system, which increases costs. In addition, the second
In the above method, two scans are required, resulting in a reduction in copying speed, and there is a problem in that the cost is high due to the necessity of accurately comparing the same pixels.

In order to improve these, the applicant has proposed a color image reading device using a two-color sensor (Japanese Patent Publication No. 62-25
4740), the resolution of the obtained image was not sufficient.

This is based on the following reasons.

Due to aberrations in the optical system that guides the reflected light from the original to the sensor, as shown in Figure 18, the density based on the amount of light received by the sensor in the isolated line image part B becomes a ladder image (where the lines are dense). A or solid image part (solid part)
The concentration in each part A is lower than that in
The density difference (difference between the peaks of the solid line and the broken line in the figure) due to different light receiving characteristics in , B is also the isolated line image part B.
The difference ΔD' at is smaller than the difference ΔD at the ladder image portion A, etc. (ΔD' - ΔD). From this, when accurately distinguishing an isolated line image as a color image, it is necessary to use a standard that takes into account the relatively small lli degree difference (difference in received light amount) for each light receiving characteristic. If the target is a ladder image or the like, there is a risk that the gap between the lines will be determined to be the color image. On the other hand, when accurately distinguishing color images of ladder originals, etc., it is done using a standard that takes into account the relatively large density difference ΔD in each light receiving characteristic, so it is necessary to use the same standard to identify isolated line images. In this case, there is a possibility that color image discrimination may not be performed. That is, it becomes difficult to discriminate the color images of both the isolated line image and the ladder image based on the same criteria. In contrast, conventionally, color image discrimination has been performed based on a standard that prioritizes isolated line images, assuming that the image is a document.

Therefore, in the conventional image reading apparatus, the resolution of the obtained color image has to be low, for example, gaps between complicated image parts are determined to be color image parts.

On the other hand, in the technical field of this type of image processing, a technique for compensating for a decrease in resolution through filter processing is generally known. However, if such a technique is simply applied to the image reading device, the filtering process changes the data that is the basis of color discrimination, making accurate color image discrimination impossible. Therefore, such a resolution correction technique cannot be simply applied.

Therefore, an object of the present invention is to enable filter processing for resolution correction to be performed without affecting data that is the basis for color discrimination.

[Means for Solving the Problems] The present invention provides a sensor 1 with two types of light receiving characteristics whose relationship is determined corresponding to a target color image, and a sensor 1 that has two types of light receiving characteristics when scanning an original 2. Light reception fi(a,
The premise is a color image reading device that obtains color image information by performing color discrimination on a pixel-by-pixel basis based on b), and the technical means for solving the above problem in the color image reading device are as follows: As shown in FIG. 1, the sensor 1 arranges (i =1.2...,n), and the first and second photoreceptors (3a(i), 3b(i)) in each cell (C) of this sensor 1 are
A color image information creation means 4 for creating color image information by discriminating colors pixel by pixel using the light receiving unit 11C1j;
A light receiving correction means 5 corrects the received light so that the amount of light received by the first photoreceptor (3a(i)) and the second photoreceptor 3b( i)) Shape extraction means 6 for extracting the shape of the target image based on the light reception state obtained by the correction by the light reception correction means 5 based on the light reception state of "C", and the image information creation means 4
The color image correction means 7 corrects the color image information obtained by the shape extraction means 6 based on the shape information extracted by the shape extraction means 6.

[Effect J In the scanning process of the target document, the amount of light received by the first photoreceptor (3a(i)) and the second photoreceptor (3b(i)) for each unit cell from sensor 1 (image The color image information creation means 4 discriminates the target color image portion based on the color image area> and creates the corresponding color image information.On the other hand, the light reception correction means 5 uses the first Photoreceptor (3a(
i)) and the amount of light received at the second photoreceptor (3b(i)
)) is corrected so that the amount of light received (density) is approximately the same as that of the image sensor 10, and the shape extraction means 6 extracts the shape of the image to be read based on the sensor light reception state obtained by the correction.

Since the shape extraction means 6 performs processing based on the light receiving states of both the first and second photoreceptors (3a(i), 3b(i)), 1. Since shape information with twice the density of the color image information created by the F color scheme image information creation means 4 is obtained and is performed separately from the color image discrimination process, it is possible to obtain shape information with twice the density of the color image information created by the color scheme image information creation means 4, and because it is performed separately from the color image discrimination process, it is possible to Even if the filter processing is performed, the basic data for color Fitj image discrimination will not change.

The color image correction means 7 then processes the color image information creation means 4.
The color image information from is corrected based on the extracted shape. The correction by the color image correction means 7 is to correct the color image information as a common part with the extracted shape.

[Example] Embodiments of the present invention will be described below based on the drawings.

FIG. 2 is a diagram showing an example of the structure of a two-color copying machine to which a color image reading device according to the present invention is applied. In this example, color reproduction is performed using black and red.

In the figure, a light source 12 for exposure is placed below a platen 11 on which a target document 10 is placed, and light reflected from the surface of the document 10 via the platen 11 from the light source 12 is reflected by a mirror 14. , a lens system 16, a half mirror 18, and a "red" color separation filter 20 to the photosensitive drum 2.
When the platen 11 is moved in the direction of the arrow 13, the original is scanned by an optical system such as the light source 12. The light source 12 is, for example, about 3
000" halogen lamp is used.A charger 2 is installed around the photosensitive drum 22 to carry out an image process.
6. A first developer 1128 that performs development with black toner, a second developer 30 that performs development with red toner, a transfer corotron 32, and a peeling corotron 34 are arranged, and furthermore, a cleaning device 36 and a static eliminator are provided at the subsequent stage. A lamp 38 is arranged.

Further, a one-dimensional image sensor 24 is arranged behind the half mirror 18, and a light output device 40 that performs exposure processing based on "red" image information extracted from the detection signal from the image sensor 24 is arranged. First developing machine 28
and the second developing device 30. With such a configuration, in the process of rotation of the photosensitive drum 22, toner images developed in MR to be "black" and "red" are transferred onto the paper 45 sent out from the cassette 43, and transferred to the conveying device 42 and the fixing device. 44 to a discharge tray 46 (one bus, two color copying).

The detailed configuration of the image sensor 24 (two-color sensor) is shown in FIG. 3, for example.

This is, for example, a COD (W coupling device) or the like, and has a structure in which the light-receiving elements 25 are arranged one-dimensionally (-dimensional image sensor), and each light-receiving element 2
On the light receiving surface of 5, green filters 27 having spectral transmission characteristics as shown in FIG. 4(a) and yellow filters 29 having spectral transmission characteristics as shown in FIG. 4(b) are arranged alternately. . Specifically, for example, 400SP1
．． 4608 dots of light receiving elements 25 are arranged at (15,75 dots/am), and a gelatin filter having the above spectral transmission characteristics is deposited on the light receiving surface of each light receiving element 25. And green filter 2
The light-receiving element 25 corresponding to 7 and the light-receiving element 25 corresponding to the yellow filter 29 form a pair as light-receiving bodies having different light-receiving characteristics, and are handled by the subsequent control system as a unit cell C corresponding to one pixel. There is.

FIG. 5 is a block diagram showing an example of the basic configuration of a color image reading device according to the present invention.

In the same figure, 52 is an output circuit that digitizes an image signal of a level corresponding to the amount of light received by each light receiving element 25 of the image sensor 24, and from this output circuit 52, a signal is sent to the light receiving element 2 in which the green filter 27 is arranged.
Green image data corresponding to the image signals from the light receiving elements 25 and yellow image data corresponding to the image signals from the light receiving elements 25 in which the yellow filter 29 is arranged are alternately output in accordance with the arrangement of each light receiving element. ing. 54 is a normalization circuit that converts each image data from the output circuit 52 into density data; 56;
converts the density data from the normalization circuit 54 into yellow density data "Yell" corresponding to the yellow image data and green density data "Gre" corresponding to the green image data.
This is a signal separation circuit that distributes signals to 131. Reference numeral 58 denotes a positional deviation correction circuit for correcting the positional deviation caused by the correspondence of two light-receiving elements 25 to one pixel, and 60 a line ballast that holds one line of yellow density data Yalt from the signal separation circuit 56. A memory 62 is green density data areaan that has been corrected by the positional deviation correction circuit 58.
This is a color discrimination circuit that discriminates whether a corresponding pixel is red (chromatic color) based on the yellow density data YELL passed through the line buffer memory 60.

The basic functions realized by the color discrimination circuit 62 are as follows.

In the red region of the target image, the difference in the spectral transmission characteristics of the green filter 27 and yellow filter 29 (the fourth
(see figure), there will be a difference in the concentration data based on the output. Specifically, the green density data Green becomes larger than the yellow density data Yall. Focusing on the difference in detected density in the red region, the green density Green-yellow density Yell≧ΔD
a...(1) ΔDa: When the red density threshold is reached, the pixel corresponding to the unit cell can be determined to be "red". In addition, since sufficient density cannot be obtained in the background part (white part) of the original even through each of the above filters (there is a large amount of light ω), green density Green≦Δob Yellow density Y (311≦ΔDb . . . 2)Δ
Db: When the white density threshold is reached, the pixel can be determined to be "white" (background area), and furthermore, pixels other than the above conditions can be determined to be a black area.

Under these circumstances, the color discrimination circuit 62 inputs the green density data arean and the yellow density data Yell and discriminates "red" according to the above equation (1).
2) The function of determining "white" (background area) according to the formula and determining "black" for other conditions has been realized.

At this time, since the color of one pixel is discriminated based on the amount of light received by the two light-receiving elements, the color image data obtained as a result of color discrimination is 172 of the detection density of the image sensor 24, that is, 200SP1. becomes.

Red 81 degree threshold ΔOa in the above equations (1) and (2)
Although the white density threshold value ΔDb is determined experimentally, for example, as shown in FIG.
1ΔQa is set to about 0.1, and the white density threshold ΔDb is set to about 0.2. In Figure 6, the density corresponding to the reflectance of white is "0", the density corresponding to 1/10 of that is "1.0", and the density corresponding to 1/100 of that is "1.0". It is set as "2.0".

Further, the functions realized by the correction circuit 58 are as follows.

This positional deviation correction circuit 58 is necessary because the color discrimination circuit 62 corresponds based on signals from two light-receiving elements treated as unit cells, that is, density information at different positions in a strict sense. This is because the color of one pixel is determined. A specific problem is that in the edge portion of a black image, the condition of equation (1) above may be satisfied and the edge portion may be determined to be a red image area, even though it is a black image portion. For example, in one unit cell, if the green density sampling position is a completely black image area and the yellow density sampling position is in the middle of the density change from black to white, the rMrjl difference will occur and the above (1) will occur. The formula holds true. Therefore, in order to eliminate such problems,
The positional deviation correction circuit 58 specifically averages the green density data of adjacent unit cells to create new green density data.

70 is yellow density data Ye from the signal separation circuit 56
When ll is a red image, green density data a
72 is a serialization circuit that alternately arranges and serializes the yellow density data "Yell" that has passed through the data correction circuit 70 and the green density data "Green" from the signal separation circuit 56. It is a circuit.

The specific configuration of the data correction circuit 70 and serialization circuit 72 is shown in FIG. 7, for example. Here, the data correction circuit 70 is constituted by an operational TiROM, and in this operational ROM, the yellow density data Ye11 from the signal separation circuit 56 for the red image is determined to be on the same level as the green density data Green, as shown in FIG. Thus, new yellow density data Yetbi is obtained by executing the calculation Yell'=(1+α)Yell on the input yellow density data Ye11. For example, the coefficient α is set to “1, 2”. Note that this correction coefficient α is determined based on the color to be discriminated, the filter characteristics to be provided in each light-receiving element 25, and the like. Further, the serialization circuit 72 has a 7-lip 70 knob 73 that temporarily holds the green density data Green from the signal separation circuit 58, and the green density data Green from the flip 70 knob 73 and the data correction circuit (operation ROM) 70. The multiplexer 74 switches its output in synchronization with the clock signal from the counter 75 in response to the yellow density data YELL° from the counter 75. The period of the clock signal from the counter is 172 of the output period of each density data from the flip 70 knob 73 and the data correction circuit 70, and the corrected yellow density data Yell° and green density data ereen are arranged alternately. This creates new serialized concentration data. Since this new density data is created by serializing the density data of each color mentioned above, it is 4003P similar to the detection density of the image sensor 24.
1. bawl.

Furthermore, 64 is a resolution correction circuit that performs predetermined filter processing on the density data from the serialization circuit 72, and 66 is a binarization circuit that creates binary image data based on the correction data from the resolution correction circuit 64. The resolution correction circuit 64 and the binarization circuit 66 constitute a shape extraction means.

In the resolution correction circuit 64, for example, the center pixel (
Only e) has a positive value, and the surrounding pixels ((a) (bHc) (d
)) takes a negative value or zero, and a certain pixel (e)
The filter characteristic is such that the density difference between the pixel and the adjacent pixel ((a), (b), (c), and (d)) is emphasized. Note that the yellow density data that has undergone the above correction process and is subject to processing by the resolution correction circuit 64 has variations due to changes in hue, saturation, and brightness of the red image, so the above-mentioned - type of correction is usually not performed. Not completely consistent with green density data. For this reason, if the filter coefficient of the center pixel (e) is increased to emphasize the difference in density between adjacent pixels too much, the variation in yellow density will affect the image, making data correction ineffective. Appropriately, the filter coefficient of the center pixel (e) is about "2 to 3". The MTF characteristic (M
oduration Transfer Function
on), the high spatial frequency range is emphasized as shown in FIG. 9(b). Here, ωO is the Nyquist frequency, and specifically, the shape data above is 4008P1.
Because of the roar, ωo −5,0 (Iρ/hazy).

Resolution correction circuit 6 that serves as a digital filter as described above
The specific configuration of 4 is shown in FIG. 10, for example.

In the figure, 53 and 55 are FIFO-configured buffers that hold one line of density data from the serialization circuit 72, and are connected in series. 518-51
quD type flip 70 tube (D-FF), D-F
The F51a and 51b directly receive the density data from the serialization circuit 72, the D-FFs 51c, 51d, and 51e receive the density data via the buffer 53, and the D-FFs 51f and 51a receive the density data via the buffer 55 at the same timing. It is configured to shift up sequentially.

This matrix-like D-FF 51a to 51G corresponds to the pixel matrix shown in FIG. 9(a). That is, pixel (a) is connected to D-FF51b, and pixel (b) is connected to D-FF5.
1Q, pixel (C) corresponds to D-FF 51c, pixel (d) corresponds to D-FF 51e, and pixel (e) corresponds to D-FF 51d.

57 is an adder that adds the density data in the D-FFs 51d and 510, that is, the density data of pixels (a) and (b); 59 is an adder that adds the density data in the D-FFs 51c and 51e, that is, the density data of pixels ((C) and (d) are adders that add the concentration data. 61 and 63 are respectively calculation ROMs, and the calculation ROM 61 multiplies the 1m degree value (a+b) added by the adder 57 by 1 by a constant -, and the calculation ROM 63 has a function of doubling the concentration value (cod) added by the adder 59 by a constant -.

65 is an operational WROM that multiplies the density data in the D-FF 51d by a constant m, 67 is an adder that adds the output values from each operational ROM 61 and 63, and 69 is an adder that adds the output value from this adder 67 and the above operational ROM 65. This is an adder that adds the output values. The output value (eo) of this final stage adder 69 is (e')=m(e)-kl (a+b)-k2 (
cod), and m=3. 1 = 2-0.5 and t
- In case 2, the output value (eo) becomes the correction value (resolution correction value) of the center pixel density (e) with the filter characteristics shown in FIG. 9(a).

On the other hand, the binarization circuit 66 is specifically composed of a comparator or an arithmetic ROM regarding density, and its output corresponds to the shape (image information) of the target image.

Further, in FIG. 5, 68 is an AND circuit, which is common to pixels that are determined to be red image portions or black image portions by the color discrimination circuit 62 and pixels extracted as shape portions by the binarization circuit 66. The portion is output from the AND circuit 68 as red image data or black image data. Note that the output pixel density (20
08P1. ) is the output pixel density (
400SPf, ) is 172, so the output cycle of the binarization circuit 66 is 1 of the output cycle of the color discrimination circuit 62.
It is set to 72.

Regarding the color discrimination system, the MTF characteristics in the color image reading device having the above configuration are, for example, as shown in FIG. 11, the MTF characteristics of the optical system (see FIG. As a result, the MTF of the entire color discrimination system also decreases in the high spatial frequency range ((C) of the same figure) as a result of the decrease in the MTF in the high spatial frequency range ((B) in the same figure).

Regarding the shape extraction system, for example, as shown in Fig. 12, although the MTF characteristics of the optical system similarly decrease in the high spatial frequency range (Fig. 12(a)), the MFT characteristics of the resolution correction is improved over a wide range of spatial frequencies ((b) in the same figure corresponds to Fig. 9(b)), so the MTF of the entire shape extraction system remains at a relatively high level up to the Nyquist frequency ω0. It turns out. Therefore, the difference in detected density between the ladder image or solid image and the isolated line image becomes small, and a high-resolution image can be extracted using a single standard (same value).

In the color image reading device configured as described above, the discrimination standard for the red image to be read corresponds to a relatively low density level suitable for isolated line images, and red image discrimination is performed in a wider range than actually. (2008P1.)
. Then, the color image data has twice the density (4008P1.
) and performs AND processing on the shape signal extracted at a high resolution by resolution correction and the color signal in a relatively wide range, resulting in high resolution red image data (4003Pl). That is, as shown in FIG. 13, by performing an AND operation on a somewhat blurred color signal and a vivid shape signal for a red original image, an output signal corresponding to a high-resolution color image is obtained.

In the above example, the extraction of the red image part was explained.
The same applies to extraction of the black image portion.

In addition, in the two-color copying machine shown in FIG. 2, the black image portion is extracted by the "red" color separation filter 20, so the above-mentioned color image reading device basically has a black image extraction function. However, for example, when reproducing a black image part in red, a color inversion function is required, and an extraction function for the black image is required.
A function for extracting the flesh color image (black, red) is also required when writing with the light output device 40 using the above. When extracting the black image part in this way, since the green density and yellow density are originally approximately the same with respect to the black image, the discrimination color information in the color discrimination circuit 62 ('black' in this case) is used.
The correction coefficient α in the data correction circuit 70 is α based on
= 0, or by switching, the image density data is directly serialized without being processed by the data correction circuit 70 (see FIG. 17).

Note that when image writing is performed using the light output device 40 based on the color image data obtained as described above, density conversion is required to match the printing density of the light output device 40.

Furthermore, reading of other color images, for example, reading of a blue image portion, can be realized by changing the filter characteristics of the image sensor 24 and performing similar processing.

For example, the yellow filter is changed to a cyan filter that has spectral transmission characteristics as shown in Figure 14, and the light source is a three-wavelength type fluorescence that includes blue and green components and has spectral characteristics as shown in Figure 15. (In a two-color copying machine, the "red" color separation filter 20 is used as the "blue" color separation filter 20.)
will be changed to that of ). In this case, the color discrimination condition based on green density data Green and cyan density data Cyan is, for example, as shown in FIG. 16, for a blue image, green density Green - cyan density Cyan≧Δ
Dc...(3) ΔDC: Blue density threshold, green density G for the background part (white image)
reen≦∆Db Cyan density Yellow≦∆Db ... (4) ∆Db
= white density threshold. Then, conditions other than the above (3) and (4) correspond to a black image. Here, as in the case of red image discrimination, the blue density threshold ΔDc is set to about 0.1, and the white density threshold ΔDb is set to about 0.2.

The data correction circuit 70 also includes cyan density data Cya.
n will be input. And data correction circuit 70
New cyan density data Cyan is obtained by the calculation Cyan' = (1+a) Cyan. In addition,
The correction coefficient α is set to α=1.0, for example.

As mentioned above, when distinguishing red images, yellow filters and green filters were used, and when discriminating blue images, cyan filters and green filters were used.
The combination of filters is not limited to these, and can be arbitrarily set as long as the light-receiving characteristics of the colors to be discriminated are different. For example, one light-receiving characteristic can be set to a desired spectral characteristic using a filter, etc., and the other can be set as desired. Regarding the above, a continuous spectrum characteristic may be obtained without providing a filter or the like. However, a filter that only contains spectral components of the color to be extracted, for example, a red filter for a red image,
For blue images, it is better to avoid using a blue filter because the density output for the relevant color component will be extremely reduced, making it difficult for the data correction circuit 70 to perform density correction.

In the above embodiment, the color image reading device is applied to a so-called two-color copying machine, but the present invention is not limited to this, and can similarly be applied to other image processing devices that require color-matching image discrimination. It can be applied to

[Effects of the Invention] As explained above, according to the present invention, the shape and color of the target image are processed separately, so resolution compensation processing is performed during shape extraction without affecting color information. Furthermore, since the shape extraction described above is extracted with twice the pixel density as color image discrimination, the modified color image information obtained by modifying the color image information based on the shape information is more The resolution will be high. Therefore, more vivid reproduction of the read color image is possible.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of the present invention, FIG. 2 is a diagram showing an example of the basic configuration of a two-color copying machine to which the color image reading device according to the present invention is applied, and FIG. 3 is a block diagram showing the configuration of the image sensor. 4 is a diagram showing an example of spectral transmission characteristics of an image sensor, FIG. 5 is a block diagram showing an example of a color image reading device according to the present invention, and FIG. 6 is a diagram showing an example of color discrimination. FIG. 7 is a block diagram showing a specific configuration example of the data correction circuit and serialization circuit, FIG. 8 is a diagram showing an example of density correction in the data correction circuit, and FIG. 9 is a diagram showing examples of conditions. Diagram showing the filter characteristics and MTF characteristics of the resolution correction circuit, No. 10
The figure is a block diagram showing a specific configuration example of the resolution correction circuit, Figures 11 and 12 are diagrams showing the MTF characteristics of the device shown in Figure 5, and Figure 13 is a diagram showing the MTF characteristics of the device shown in Figure 5. 14 is a diagram showing the spectral transmission characteristics of other filters. FIG. 15 is a diagram showing an example of the spectral characteristics of the light source. FIG. 16 is a diagram showing an example of the conditions for color discrimination. FIG. 17 is a block diagram showing another example of a color image reading device, and FIG. 18 is a diagram showing an example of detected density states in a ladder image portion and an isolated line image portion. [Explanation of symbols] 1...Sensor 2...Document 3 a (i) (i-1, 2,..., n)...First
Photoreceptor 3b (i) (i-1, 2, -, n) = second
Photoreceptor 4...Color image information creation means 5...Light receiving correction means 6...Shape extraction means 7...Color image correction means 24...Image sensor 25...Light receiving element 27... Green filter 29...Yellow filter 52...Output circuit 54...Normalization circuit 56...Signal separation circuit 58...Position shift correction circuit 60...Line buffer 7 62...Color discrimination circuit 64...Resolution correction circuit 66...Binarization circuit 70...Data correction circuit 72...Serialization circuit C...Unit cell patent applicant Fuji Xerox Co., Ltd. Agent
Patent attorney Tomohiro Nakamura (3 others) Figure 4 Wavelength (nm) Wavelength (nm) Figure 6 Green sensor concentration Figure 7 Figure 8 Green sensor concentration Figure 9 One O Spatial frequency (Ip/IIIn) Figure 11 Spatial frequency Cl2p/Kan) Spatial frequency (I2p/Kaku) Spatial frequency (j2p/Kaku) Figure 12 Spatial frequency (Jlp/am) Spatial frequency (Alp/■) Spatial frequency (J2p/m) Figure 14 Figure 15 Wavelength (rum) Figure 13 (Shape signal) Figure 16 0.2 0.4 0.6 0.8 1.0 1.2
1.4 1.6 Edge sensor concentration Figure 18 A B

Claims (1)

[Claims] A sensor (1) having two types of light receiving characteristics whose relationship is determined corresponding to a target color image is used, and each light receiving characteristic of the sensor (1) is used when scanning an original (2). The amount of light received for each characteristic (a
, b), which obtains color image information by performing color discrimination on a pixel-by-pixel basis. array of photoreceptors {3a(i), 3b(i)} as a unit cell (C) (i=1, 2,..., n)
This sensor (1) has a structure in which color discrimination is performed in pixel units based on the amount of light received by the first and second photoreceptors {3a(i), 3b(i)} in each unit cell (C) of this sensor (1). a color image information creation means (4) that creates color image information using the color image information; } Light reception correction means (5) that performs light reception correction so that the amount of light received at
and the first photoreceptor {3a(i)} and the second photoreceptor {3b(
i) The above-mentioned light reception correction means (5) based on the light reception state in
) Shape extracting means (6) extracting the shape of the target image based on the light reception state obtained by the correction in (6); and Shape extracting the color image information obtained by the color image information creation means (4). A color image reading device comprising a color image correction means (7) for correcting the shape based on the shape information extracted by the means (6).