JP5950670B2 - Image reading apparatus, image processing apparatus, image processing method, and computer program - Google Patents

Description

  The present invention relates to a technique for optically reading an image formed on a read medium using a plurality of line sensors that receive light reflected by the read medium such as an original.

  2. Description of the Related Art Conventionally, image reading apparatuses that read a multicolor image on a read medium using an image sensor including a plurality of line sensors arranged at regular intervals so as to be parallel to each other have been widely used. As an image sensor of this type of image reading apparatus, there are three image sensors that receive light of three primary colors R (red), G (green), and B (blue) in order to read a multicolor image and perform color discrimination. In many cases, a three-color integrated sensor in which the line sensor is integrated into one chip is used. Usually, each line sensor constituting the image sensor is a solid-state imaging device including a plurality of photoelectric conversion elements arranged at a constant interval (pitch) along the main scanning direction of the image reading apparatus. When the image reading device reads a document image, the image sensor continuously moves a linear image of light reflected by the document while moving relative to the document in a sub-scanning direction substantially orthogonal to the main scanning direction. It is necessary to perform photoelectric conversion.

  The plurality of line sensors constituting the image sensor usually have the same longitudinal direction as the main scanning direction and are arranged at regular intervals along the sub-scanning direction. For this reason, when the image sensor moves relative to the read medium in the sub-scanning direction, it takes time to read the same area image of the read medium between the line sensors due to the gap (gap) of the line sensors. Difference occurs. For example, among the plurality of line sensors, a line sensor on one end side in the arrangement direction photoelectrically converts an optical image of a specific area of the read medium prior to the line sensor on the other end side in the arrangement direction. Therefore, a deviation (hereinafter also referred to as a “phase deviation”) regarding a reading position occurs between a plurality of image signals output in parallel from a plurality of line sensors at the same timing.

  Basically, such phase shift correction is performed in units of lines by a delay time corresponding to the interval (gap) of the line sensors for specific image signals among the plurality of image signals output in parallel from the plurality of line sensors. This can be realized by using a delay memory for delaying and an interpolation calculation unit for performing an interpolation calculation on the output of the delay memory. For example, in the case of an image reading device having a function of changing the relative speed (sub-scanning speed) of the image sensor in the sub-scanning direction with respect to the read medium in accordance with the reading magnification (variable magnification) for image reduction or enlargement, The phase shift is not an integral multiple of the delay time for one line, and a phase shift that is a fraction (decimal part) times the delay time for one line may remain uncorrected. In such a case, the fractional phase shift is corrected using interpolation calculation.

  Conventional techniques for correcting such a phase shift include, for example, JP-A-1-109966 (Patent Document 1), JP-A-9-270898 (Patent Document 2), and Japanese Patent No. 312377 (Patent Document 3). ), Japanese Patent No. 3230281 (Patent Document 4) and Japanese Patent No. 3379462 (Patent Document 5).

Japanese Patent Laid-Open No. 1-109966 (FIGS. 2, 5, 6, etc.) JP-A-9-270898 (FIGS. 5 to 7, paragraph 0008, paragraphs 0014 to 0016, etc.) Japanese Patent No. 311377 (FIGS. 3, 9, 29, paragraphs 0003-0009, paragraphs 0046-0047, paragraphs 0074-0076, etc.) Japanese Patent No. 3230281 (FIG. 3, paragraphs 0049-0056, etc.) Japanese Patent No. 3379462 (FIGS. 1-2, paragraphs 0015-0022, etc.)

  However, in the case of an image reading apparatus that uses a plurality of line sensors that respectively receive light of a plurality of colors in order to read a multicolor image, interpolation is individually performed on a plurality of image signals output in parallel from the plurality of line sensors. When processing is performed to correct the phase shift, color reproducibility may deteriorate.

  For example, when three line sensors such as an R line sensor, a G line sensor, and a B line sensor that receive light of three primary colors R, G, and B are used, the R line sensor and the output of the G line sensor When linear interpolation is performed on the output of the R line sensor and the B line sensor in order to correct the phase shift of the output of the B line sensor, since the linear interpolation is a smoothing process, the red component and the blue component of the read image Each MTF (Modulation Transfer Function) is degraded as compared with the MTF of the green component. As a result, the gradation balance of R, G, B may be lost and color reproducibility may be reduced.

  In particular, if there is an edge portion in which the gradation changes greatly in the original image, the gradation balance of R, G, B may be greatly lost due to the interpolation processing, and color misregistration may occur. In such a case, for example, not only does the image feel blurred, but a black and white document in the original image is erroneously recognized as a green or red-colored thin line. Problems arise. Further, when the image reading apparatus has a black character detection function for detecting a black character portion in the original image, the image reading apparatus changes the color of the black thin line to another color due to the MTF difference between the colors described above. There is also a risk of erroneous determination. For this reason, black character determination cannot be performed with high accuracy, and there is a problem that the quality of the black character portion and the thin line portion of the output image is deteriorated.

  In order to prevent the occurrence of such color misregistration, for example, when a thin line portion is detected from the original image, it is conceivable that the interpolation processing is not performed. However, if the interpolation process is not executed, the phase shift is not corrected, and there is a possibility that the tone balance of R, G, and B will be greatly lost for a color image.

  In view of the above, an object of the present invention is to provide an image reading apparatus, an image processing apparatus, an image processing method, and a computer program capable of suppressing color misregistration and realizing good color reproducibility even when phase deviation is corrected. Is to provide.

  An image reading apparatus according to a first aspect of the present invention is an image reading apparatus for reading a two-dimensional image from a read medium extending in a main scanning direction and a sub-scanning direction different from the main scanning direction. An image sensor including a plurality of lines of line sensors arranged at predetermined intervals along the scanning direction, and the image sensor is moved relative to the read medium in the sub-scanning direction and reflected on the surface of the read medium. A scanning drive unit that causes the image sensor to receive the received light, and the line sensors of the plurality of rows at regular time intervals within a scanning period in which the image sensor moves relative to the read medium in the sub-scanning direction. An image processing unit that generates image data based on a plurality of series of read pixel signals output in parallel, and the plurality of lines of line sensors respectively emit light of a plurality of different colors. The plurality of series of read pixel signals are photoelectrically converted and the plurality of series of read pixel signals are output in parallel, and the image processing unit is configured to output a reference series of read pixel signals of the plurality of series of read pixel signals and other series of read pixels. An interpolation position determination unit for determining a pixel interpolation position for correcting a phase shift between the signal and the plurality of read pixel signals of the plurality of series by performing an interpolation operation for each series, and the plurality of pixels at the pixel interpolation position An interpolation processing unit that calculates a gradation value for each of the colors, and whether or not there is an edge having a gradation change in the sub-scanning direction based on at least one read pixel signal of the plurality of read pixel signals An edge detecting unit that performs a shift when the interpolation processing unit deviates from the pixel interpolation position and the pixel interpolation position with respect to the read pixel signal of the reference sequence when it is determined that the edge exists. The interpolation calculation is performed using a weighting coefficient set with reference to the first interpolation position of any one of the positions, and for the read pixel signals of the other series, the first interpolation position and Is characterized in that the interpolation operation is executed using a weighting coefficient set with reference to different second interpolation positions.

An image processing apparatus according to a second aspect of the present invention is arranged at predetermined intervals along a sub-scanning direction with respect to a medium to be read that extends in a main scanning direction and a sub-scanning direction different from the main scanning direction. An image sensor including a plurality of lines of line sensors for photoelectrically converting light of a plurality of different colors and outputting a plurality of series of read pixel signals in parallel; and An image processing apparatus incorporated in an image reading apparatus including a scanning drive unit that causes the image sensor to receive light that is relatively moved in a scanning direction and reflected by the surface of the read medium. the image sensor Te is the plurality of series of read pixel signals from the line sensor of the plurality of rows every predetermined time interval in the scanning period is output in parallel to relative movement in the sub scanning direction An interpolation position determining unit that determines a pixel interpolation position for correcting a phase shift between a read pixel signal of a reference sequence and a read pixel signal of another sequence other than the reference sequence; and the read pixel signals of the plurality of sequences An interpolation processing unit that performs an interpolation operation for each series to calculate gradation values of each of the plurality of colors at the pixel interpolation position, and at least one series of read pixels among the plurality of series of read pixel signals An edge detection unit that determines the presence or absence of an edge having a gradation change in the sub-scanning direction based on a signal, and when the interpolation processing unit determines that the edge is present, the reference series of read pixels For the signal, the interpolation calculation is performed using a weighting coefficient set with reference to the first interpolation position of one of the pixel interpolation position and the position shifted from the pixel interpolation position. In addition, with respect to the read pixel signals of the other series, the interpolation calculation is performed using a weighting coefficient set with reference to a second interpolation position different from the first interpolation position. .

According to a third aspect of the present invention, there is provided an image processing method in which a read medium extending in a main scanning direction and a sub-scanning direction different from the main scanning direction is arranged at predetermined intervals in the sub-scanning direction. An image sensor including a plurality of lines of line sensors for photoelectrically converting light of a plurality of different colors and outputting a plurality of series of read pixel signals in parallel; and the image sensor with respect to the read medium in the sub-scanning direction An image processing method in an image reading apparatus including a scanning drive unit that causes the image sensor to receive light reflected on the surface of the read medium relative to the read medium. among the plurality of series of read pixel signals from the line sensor of the plurality of rows every predetermined time interval in the scanning period is output in parallel to relative movement in the sub-scanning direction, the reference system Determining a pixel interpolation position for correcting a phase shift between the read pixel signal of the other series and the read pixel signal of the other series other than the reference series, and at least one of the plurality of series of read pixel signals. Determining whether or not there is an edge having a gradation change in the sub-scanning direction based on the read pixel signal; and when it is determined that the edge is present, the pixel interpolation position with respect to the reference series read pixel signal And a color corresponding to the reference series at the pixel interpolation position by executing an interpolation operation using a weighting coefficient set with reference to the first interpolation position of any one of the positions shifted from the pixel interpolation position A second interpolation position different from the first interpolation position with respect to the read pixel signals of the other series when it is determined that the edge exists. Characterized in that it comprises a step of calculating the color tone values of corresponding to the other sequence in the pixel interpolation positions by performing the interpolation calculation using the weighting coefficient set as a reference.

According to a fourth aspect of the present invention, there is provided a computer program that is arranged at a predetermined interval in the sub-scanning direction with respect to a medium to be read that extends in a main scanning direction and a sub-scanning direction different from the main scanning direction. An image sensor including a plurality of lines of line sensors for photoelectrically converting light of a plurality of different colors and outputting a plurality of series of read pixel signals in parallel; and the image sensor in the sub-scanning direction with respect to the read medium A scanning drive unit that causes the image sensor to receive light that is relatively moved and reflected by the surface of the read medium, and constant within a scanning period in which the image sensor moves relative to the read medium in the sub-scanning direction. image with an image processing unit for generating image data from the line sensor of the plurality of rows for each time period on the basis of the read pixel signals of the plurality of streams outputted in parallel In reading apparatus, a computer program for executing the image processing in the image processing unit, the computer program is intended to be read and executed from the computer-readable recording medium by the image processing unit, the image The processing includes an interpolation position for determining a pixel interpolation position for correcting a phase shift between a read pixel signal of a reference series of the read pixel signals of the plurality of series and a read pixel signal of a series other than the reference series. Determination processing, edge detection processing for determining the presence or absence of an edge having a gradation change in the sub-scanning direction based on at least one series of read pixel signals among the plurality of series of read pixel signals, and the presence of the edges When the determination is made, the pixel interpolation position and the position shifted from the pixel interpolation position with respect to the reference series read pixel signal First interpolation for calculating a gradation value of a color corresponding to the reference series at the pixel interpolation position by executing an interpolation operation using a weighting coefficient set with the first interpolation position as one of the references Processing and interpolation using a weighting coefficient set with reference to a second interpolation position different from the first interpolation position for the read pixel signal of the other series when it is determined that the edge is present And a second interpolation process for calculating a gradation value of a color corresponding to the other series at the pixel interpolation position by performing an operation.

  According to the present invention, when it is determined that there is an edge, a reference sequence is obtained by performing an interpolation operation using a weighting coefficient set with reference to the first interpolation position for the read pixel signal of the reference sequence. On the other hand, for the read pixel signals of other series, a weighting coefficient set based on a second interpolation position different from the first interpolation position is used for the read pixel signals of other series. The tone value of the color corresponding to the other series is calculated by executing the interpolation calculation. Thereby, even if the phase shift of the read pixel signal is corrected, the color shift of the gradation value can be suppressed. Therefore, good color reproducibility can be realized.

1 is a diagram schematically illustrating a main configuration of an image reading apparatus according to a first embodiment of the present invention. 1 is a diagram schematically showing a configuration of a linear image sensor according to a first embodiment. (A), (B) is a figure which illustrates roughly the arrangement of a plurality of line sensors. 2 is a functional block diagram illustrating a schematic configuration of an image processing unit according to Embodiment 1. FIG. (A)-(C) are the figures which show the structural example of the line delay part of Embodiment 1. FIG. 3 is a functional block diagram illustrating a schematic configuration of an interpolation control unit according to the first embodiment. FIG. (A)-(C) are functional block diagrams which show schematic structure of the interpolation process part of Embodiment 1. FIG. It is a figure which shows the transition of the relative position of the line sensor in a scanning period. (A)-(C) is a figure which shows the phase shift between the outputs of a line delay part with respect to the pixel interpolation position of FIG. (A)-(C) is a figure for demonstrating the pixel interpolation method by a cubic function interpolation method. It is a figure which shows the transition of the relative position of the line sensor in a scanning period. (A)-(D) are graphs showing the relationship between the pixel position and the interpolation gradation value when it is assumed that the original image has the gradation edge of FIG. 11 and the interpolation control unit does not function. (A)-(C) is a figure for demonstrating the interpolation method of R component when the gradation edge of FIG. 11 exists. (A)-(C) is a figure for demonstrating the interpolation method of B component when the gradation edge of FIG. 11 exists. 12 is a graph showing interpolation gradation values of R component, G component, and B component generated when the gradation edge of FIG. 11 exists. It is a functional block diagram which shows schematic structure of the image process part of Embodiment 2 which concerns on this invention. FIG. 10 is a functional block diagram illustrating a schematic configuration of an interpolation control unit according to a second embodiment. It is a figure which shows the transition of the relative position of the line sensor in a scanning period. (A)-(D) are graphs showing the relationship between the pixel position and the interpolation gradation value when it is assumed that the original image has the gradation edge of FIG. 18 and the interpolation control unit does not function. It is a graph which shows the interpolation gradation value of R component, G component, and B component when the color shift of FIG.19 (D) is correct | amended. It is a figure which shows the transition of the relative position of the line sensor in a scanning period. (A)-(D) are graphs showing the relationship between the pixel position and the interpolation gradation value when it is assumed that the original image has the gradation edge of FIG. 21 and the interpolation control unit does not function. It is a graph which shows the interpolation gradation value of R component, G component, and B component when the color shift of FIG.22 (D) is correct | amended. It is a figure which shows the transition of the relative position of the line sensor in a scanning period. (A)-(D) are graphs showing the relationship between the pixel position and the interpolation gradation value when it is assumed that the original image has the gradation edge of FIG. 24 and the interpolation control unit does not function. FIG. 26 is a graph showing interpolation gradation values of an R component, a G component, and a B component when the color misregistration in FIG. 25 (D) is corrected. It is a figure which shows the transition of the relative position of the line sensor in a scanning period. (A)-(D) are graphs showing the relationship between the pixel position and the interpolation gradation value when it is assumed that the original image has the gradation edge of FIG. 27 and the interpolation control unit does not function. It is a graph which shows the interpolation gradation value of R component, G component, and B component when the color shift of FIG.28 (D) is correct | amended. It is a figure which shows the transition of the relative position of the line sensor in a scanning period. (A)-(D) are graphs showing the relationship between the pixel position and the interpolation gradation value when it is assumed that the original image has the gradation edge of FIG. 30 and the interpolation control unit does not function. It is a figure which shows the transition of the relative position of the line sensor in a scanning period. (A)-(D) are graphs showing the relationship between the pixel position and the interpolation gradation value when the original image has the gradation edge of FIG. 32 and the interpolation control unit does not function. It is a functional block diagram which shows schematic structure of the image process part of Embodiment 3 which concerns on this invention. FIG. 10 is a functional block diagram schematically showing a configuration of an interpolation control unit according to a third embodiment. It is a functional block diagram of an arithmetic unit showing a configuration when the functions of the image processing unit of the first to third embodiments are realized by a computer program. It is a flowchart which shows an example of the process sequence which concerns on Embodiment 4 which concerns on this invention. It is a flowchart which shows an example of the process sequence which concerns on Embodiment 5 which concerns on this invention. It is a flowchart which shows an example of the process sequence which concerns on Embodiment 6 which concerns on this invention.

  Hereinafter, various embodiments according to the present invention will be described with reference to the drawings. In the drawings, the same components are denoted by the same reference numerals, and detailed description thereof is omitted as appropriate so as not to overlap.

Embodiment 1 FIG.
FIG. 1 is a diagram schematically showing a main configuration of an image reading apparatus 1 according to Embodiment 1 of the present invention. The image reading apparatus 1 has a function of reading a color image formed on the back surface of a sheet-like read medium 2. The read medium 2 has a surface extending in each of the sub-scanning direction X and the main scanning direction Y, and has a thickness in the Z-axis direction orthogonal to both the sub-scanning direction X and the main scanning direction Y. It is. Examples of the medium 2 to be read include a document such as a paper medium. In the present embodiment, the sub-scanning direction X and the main scanning direction Y are orthogonal to each other, but the present invention is not limited to this.

  As shown in FIG. 1, the image reading apparatus 1 includes a translucent plate 10 on which a sheet-shaped read medium 2 is placed, and a surface of the read medium 2 via the translucent plate 10. An imaging sensor module 12 that detects a linear image of the light reflected by the read medium 2 by irradiating light to the read medium 2, and a scanning drive unit 11 that mechanically moves the imaging sensor module 12 along the sub-scanning direction X; An A / D converter (ADC) 13 for converting three series of analog electrical signals (read pixel signals) R1, G1, B1 output in parallel from the imaging sensor module 12 into digital signals R2, G2, B2, respectively; And an image processing unit 14 that performs image processing on the series of digital signals R2, G2, and B2 to generate image signals Rout, Gout, and Bout representing two-dimensional color image data.

  In addition, the image reading apparatus 1 includes a controller (main control unit) 15 that individually controls the operations of the scanning drive unit 11 and the image processing unit 14, and an operation panel 16. The operation panel 16 has, for example, an operation input device such as a key input device or a pointing device, and the user can input setting information or instruction information such as a scaling factor by operating the operation panel 16. It is. When the user designates the variable magnification, the image reading apparatus 1 can read an enlarged image or a reduced image of the image formed on the read medium 2 at a variable magnification other than the equal magnification (= 100%).

  The image reading apparatus 1 has not only the components 10 to 16 shown in FIG. 1 but also a configuration other than these components 10 to 16 (for example, a driver circuit that drives the light sources 21A and 21B). .

  The scanning drive unit 11 includes a drive mechanism (not shown) such as a guide mechanism that guides the image sensor module 12 along the sub-scanning direction X and a drive motor that drives the guide mechanism. The scanning drive unit 11 operates in response to a command from the controller 15 and can move the image sensor module 12 relative to the read medium 2 at a sub-scanning speed designated by the controller 15. The sub-scanning speed is set to a value corresponding to the variable magnification. During the scanning period in which the scanning drive unit 11 relatively moves the image sensor module 12 along the sub-scanning direction X, the image sensor module 12 continuously produces a linear image of the light reflected by the read medium 2 at regular time intervals. Thus, three series of read pixel signals R1, G1, and B1 can be output in parallel.

  As shown in FIG. 1, the imaging sensor module 12 includes a pair of light sources 21 </ b> A and 21 </ b> B that irradiate the back surface of the read medium 2 with a strip-shaped light along the main scanning direction Y via the translucent plate 10. It has a linear image sensor 23 and an equal magnification imaging optical system 22 for imaging the light reflected by the read medium 2 on the linear image sensor 23 at an equal magnification. Since the equal-magnification imaging optical system 22 is used, the image sensor module 12 of the present embodiment is a so-called contact image sensor (CIS: Contact Image Sensor). For the light sources 21A and 21B, for example, a module configured by arranging a plurality of LED (Light-Emitting Diode) elements along the main scanning direction Y may be used.

  The linear image sensor 23 has a longitudinal direction substantially the same as the main scanning direction Y, and has three rows of line-like shapes that respectively receive light of three primary colors of R (red), G (green), and B (blue). It has a solid-state image sensor, that is, a line sensor. The three line sensors R, G, and B photoelectrically convert the linear image of the reflected light from the read medium 2 and output three series of read pixel signals R1, G1, and B1, respectively.

  FIG. 2 is a diagram schematically showing a configuration of the linear image sensor 23 according to the first embodiment. As shown in FIG. 2, the linear image sensor 23 includes an R line sensor 23R having a color filter that transmits only red light (light having a wavelength of about 632 nm) and green light (light having a wavelength of about 530 nm). G line sensor 23G having a color filter that transmits only blue light, and B line sensor 23B having a color filter that transmits only blue light (light having a wavelength of about 470 nm). The R line sensor 23R has hundreds of light receiving elements 24R,..., 24R arranged along the main scanning direction Y as photoelectric conversion elements. Similarly, the G line sensor 23G has several hundred light receiving elements 24G,..., 24G arranged along the main scanning direction Y as photoelectric conversion elements, and the B line sensor 23B extends along the main scanning direction Y. .., 24B are arranged as photoelectric conversion elements. These R line sensor 23R, G line sensor 23G, and B line sensor 23B output read pixel signals R1, G1, and B1 having levels corresponding to the amounts of received light, respectively. As the light receiving elements 24R, 24G, and 24B, for example, a CCD (Charge-Coupled Device) or a CMOS (Complementary Metal-Oxide Semiconductor) image sensor can be used.

  3A and 3B are diagrams schematically illustrating the arrangement of the R line sensor 23R, the G line sensor 23G, and the B line sensor 23B. FIG. 3A is a diagram showing a structure in which the R line sensor 23R, the G line sensor 23G, and the B line sensor 23B are arranged along the sub-scanning direction X at an interval of one pixel pitch (= d). 3B is a diagram illustrating a structure in which the R line sensor 23R, the G line sensor 23G, and the B line sensor 23B are arranged along the sub-scanning direction X at intervals of 1/3 pixel pitch (= d / 3). is there.

  Next, the image processing unit 14 will be described. FIG. 4 is a functional block diagram illustrating a schematic configuration of the image processing unit 14 according to the first embodiment.

  As shown in FIG. 4, the image processing unit 14 includes a shading correction unit 31, line delay units 32 R, 32 G, and 32 B, interpolation processing units 33 R, 33 G, and 33 B, an edge detection unit 36, and an interpolation control unit 39. The edge detection unit 36 includes an R edge detection unit 37R, a G edge detection unit 37G, a B edge detection unit 37B, and an edge determination unit 38.

  The shading correction unit 31 corrects the characteristic variation of each light receiving element of the linear image sensor 23 for each line sensor. The shading correction unit 31 outputs a correction pixel signal R3 obtained by correcting the characteristic variation of each light receiving element 24R for the digital signal R2, and corrects the characteristic variation of each light receiving element 24G for the digital signal G2. The correction pixel signal G3 obtained in this way is output, and the correction pixel signal B3 obtained by correcting the characteristic variation of each light receiving element 24B for the digital signal B2 is output. These corrected pixel signals R3, G3, B3 are supplied to the line delay units 32R, 32G, 32B, respectively.

  The line delay units 32R, 32G, and 32B are delay memories having a function of correcting the phase shift between the correction pixel signals R3, G3, and B3 in units of lines by delaying the correction pixel signals R3, G3, and B3, respectively. The line delay unit 32R outputs four series of delayed pixel signals R41, R42, R43, and R44 for four consecutive lines to the interpolation processing unit 33R, and among these delayed pixel signals R41, R42, R43, and R44, the delayed pixel signal R42 and R43 are supplied to the R edge detector 37R. Further, the line delay unit 32G outputs four series of delayed pixel signals G41, G42, G43, and G44 for four consecutive lines to the interpolation processing unit 33G, and the delay among the delayed pixel signals G41, G42, G43, and G44. Pixel signals G42 and G43 are supplied to a G edge detector 37G. The line delay unit 32B outputs four series of delayed pixel signals B41, B42, B43, and B44 for four consecutive lines to the interpolation processing unit 33B, and among these delayed pixel signals B41, B42, B43, and B44, the delayed pixel signal B42 and B43 are supplied to the B edge detector 37B.

5A to 5C are diagrams illustrating configuration examples of the line delay units 32R, 32G, and 32B. As shown in FIG. 5 (A), line delay unit 32R includes a line memory _40R 1 _{~40R N} to N lines (N is an integer of five or more) holds the corrected pixel signal R3 of partial transfer, these line memories 40R ₁ ~40R _N from four lines from the delayed signal _IR 1 _{~IR N} output in parallel delayed pixel signals R41, R42, and a R43, the selector selects the R44 41R. Line memory _40R 1 _{~40R N,} are connected in series, each of the line memories _40R 1 _{~40R N} is stored capable of retaining corrected pixel signal R3 of one line in the main scanning direction Y of the R line sensor 23R Have capacity. The selector 41R operates in response to control signals CR1 supplied from the controller 15, in accordance with the given delay [delta] _R in the control signal CR1, in the output _IR 1 _{~IR N} line memories _40R 1 _{~40R N} The delayed pixel signals R41, R42, R43, and R44 for four lines are selected. Similarly, as illustrated in FIG. 5B, the line delay unit 32G holds line memories 40G _{1 to} 40G _N that store and transfer correction pixel signals G3 for _N lines, and these line memories 40G _{1 to} 40G _N. delayed signal _IG 1 _~IG delayed pixel signals for four lines from among the _N output in parallel from the G41, G42, selects the G43, G44 and a selector 41G. The selector 41G operates in response to control signals CG1 supplied from the controller 15, in accordance with the given delay [delta] _G in the control signal CG1, in the output _IG 1 _{~IG N} line memories _40G 1 to 40 g _N To 4 lines of delayed pixel signals G41, G42, G43, and G44 are selected. Further, as shown in FIG. 5C, the line delay unit 32B holds the line of corrected pixel signals B3 for _N lines 40B _{1 to} 40B _N and the line memories 40B ₁ to 40B _N. delayed signal _IB 1 _~IB delayed pixel signals for four lines from among the _N B41 outputted in parallel, B42, B43, and a selector 41B for selecting B44. The selector 41B operates in accordance with a control signal CB1 supplied from the controller 15, in accordance with the given delay [delta] _B by the control signal CB1, in the output _IB 1 _{~IB N} line memory _40B 1 _{~40B N} The delayed pixel signals B41, B42, B43, and B44 for four lines are selected.

The phase shift between the correction pixel signals R3, G3, and B3 is the arrangement interval (distance) of the R line sensor 23R, G line sensor 23G, and B line sensor 23B in the sub-scanning direction X, and the linear image sensor 23 with respect to the read medium 2. Due to the scan interval. The scanning interval of the linear image sensor 23 corresponds to the distance that the linear image sensor 23 moves in the sub-scanning direction X while capturing an image for two lines, and can change according to the magnification. For example, the controller 15 changes the delay times δ _R , δ _G , and δ _B so that the phase shift of the correction pixel signals R3 and B3 of the other series with respect to the series (reference series) of the correction pixel signal G3 can be corrected in line units. A value corresponding to the magnification can be set. Each of the line delay units 32R, 32G, and 32B can correct a phase shift that is an integral multiple of the delay time for one line. ) Is not provided. Such correction of the fractional phase shift is performed by the interpolation processing units 33R, 33G, and 33B.

  In the present embodiment, the line delay unit 32R outputs the delayed pixel signals R41, R42, R43, and R44 for four lines in parallel, but the present invention is not limited to this. The same applies to the line delay units 32G and 32B. Each of the line delay units 32R, 32G, and 32B outputs a delayed pixel signal for five or more lines according to the number of data points to be used for the interpolation calculation performed by the subsequent interpolation processing units 33R, 33G, and 33B. You may output in parallel.

  The R edge detection unit 37R determines the presence / absence of a red edge having a large gradation change in the sub-scanning direction X from the delayed pixel signals R42 and R43 for two consecutive lines, and determines the determination signal EDR indicating the determination result as an edge determination. To the unit 38. For example, if the amount of gradation change between adjacent pixels (lines) in the sub-scanning direction X exceeds a threshold value, it is determined that there is a red edge, otherwise there is no red edge. Can be determined. Similarly, the G edge detection unit 37G determines the presence or absence of a green edge having a large gradation change in the sub-scanning direction X from the delayed pixel signals G42 and G43 for two continuous lines, and a determination signal indicating the determination result The EDG is output to the edge determination unit 38. Further, the B edge detection unit 37B determines the presence or absence of a blue edge having a large gradation change in the sub-scanning direction X from the delayed pixel signals B42 and B43 for two continuous lines, and generates a determination signal EDB indicating the determination result. The data is output to the edge determination unit 38.

  The edge determination unit 38 is based on the determination signals EDR, EDG, and EDB, and in the sub-scanning direction X in the original image formed on the read medium 2 according to the combination of the presence / absence of the red edge, the green edge, and the blue edge. The presence or absence of an edge is determined. For example, the edge determination unit 38 determines the presence / absence of an edge, such as an edge of a monochrome monochrome image, that is assumed to lose the gradation balance of R, G, B due to the interpolation processing by the interpolation processing units 33R, 33G, 33B. The edge determination signal ED indicating the determination result is output to the interpolation control unit 39.

  The interpolation control unit 39 determines a pixel interpolation position common to R, G, and B for correcting the above-described fractional phase shift in accordance with the edge determination signal ED, and R, G, and B based on the pixel interpolation position. Interpolation control signals PHR, PHG, and PHB representing the interpolation center positions of G and B are supplied to the interpolation processing units 33R, 33G, and 33B, respectively. The interpolation processing units 33R, 33G, and 33B perform interpolation calculation using weighting coefficients on the output sequences of the line delay units 32R, 32G, and 32B in accordance with the interpolation control signals PHR, PHG, and PHB. R, G, and B gradation values at the pixel interpolation position are generated. Then, image signals Rout, Gout, and Bout representing the R, G, and B gradation values are output to the outside.

FIG. 6 is a functional block diagram showing a schematic configuration of the interpolation control unit 39 of the present embodiment. As shown in FIG. 6, the interpolation control unit 39 includes an interpolation position determination unit 43 that determines parameters r, g, and b indicating pixel interpolation positions in accordance with the control signal CL supplied from the controller 15, and an edge determination signal. An interpolation position control unit 44 for calculating phase correction amounts (hereinafter also referred to as “correction amounts”) Δr, Δg, Δb for correcting the parameters r, g, b according to ED, and adders 45R, 45G. , 45B. The adder 45R adds the correction amount Δr to the parameter r and outputs an interpolation control signal PHR indicating the addition result. The adder 45G adds the correction amount Δg to the parameter g and performs interpolation control indicating the addition result. The signal PHG is output, and the adder 45B outputs the interpolation control signal PHB indicating the addition result by adding the correction amount Δb to the parameter b. Therefore, when the values of the interpolation control signals PHR, PHG, and PHB are expressed by qr, qg, and qb, the following expression is established.
qr = r + Δr
qg = g + Δg
qb = b + Δb

  The interpolation position determination unit 43 determines parameters r, g, and b indicating pixel interpolation positions for correcting this phase shift according to the fractional phase shift specified by the control signal CL. The fractional phase shift may be calculated by the interpolation position determination unit 43 instead of being calculated by the controller 15. As will be described later, the interpolation control signals PHR, PHG, and PHB are signals that control the value of the interpolation coefficient (weighting coefficient) used for the interpolation calculation in the interpolation processing units 33R, 33G, and 33B. When the edge determination signal ED indicating no edge is input, the interpolation position control unit 44 sets the correction amounts Δr, Δg, and Δb to reference values (= 0). On the other hand, when the edge determination signal ED indicating the presence of an edge is input, the interpolation position control unit 44 sets the correction amounts Δr, Δg, Δb to values other than the reference values, and the interpolation processing units 33R, 33G, 33B Change the weighting factor. As a result, it is possible to reduce a gradation value shift (color shift) between R, G, and B in the same pixel.

7A to 7C are functional block diagrams illustrating a schematic configuration of the interpolation processing units 33R, 33G, and 33B. As shown in FIG. 7A, the interpolation processing unit 33R includes an interpolation coefficient calculation unit 48R that calculates interpolation coefficients (weighting coefficients) KR1 to KR4 according to the interpolation control signal PHR, and an output R41 of the line delay unit 32R. To R44, an interpolation calculation unit 47R that executes interpolation calculation (filtering) using the interpolation coefficients KR1 to KR4. The interpolation coefficient calculation unit 48R calculates interpolation coefficients KR1 to KR4 according to a predetermined interpolation method. A method for calculating the interpolation coefficients KR1 to KR4 will be described later. The interpolation calculation unit 47R can calculate the R gradation value QR at the pixel interpolation position by executing an interpolation calculation according to the following equation.
QR = (KR1 × R41 + KR2 × R42 + KR3 × R43 + KR4 × R44) / SR

  Here, SR = KR1 + KR2 + KR3 + KR4.

Similarly, the interpolation processing unit 33G includes an interpolation calculation unit 47G and an interpolation coefficient calculation unit 48G, as shown in FIG. 7B. The interpolation calculation unit 47G calculates interpolation coefficients (weighting coefficients) KG1 to KG4 using a predetermined interpolation method according to the interpolation control signal PHG. Also, the interpolation coefficient calculation unit 48G can calculate the G gradation value QG at the pixel interpolation position by executing an interpolation calculation according to the following equation.
QG = (KG1 × G41 + KG2 × G42 + KG3 × G43 + KG4 × G44) / SG

  Here, SG = KG1 + KG2 + KG3 + KG4.

In addition, the interpolation processing unit 33B includes an interpolation calculation unit 47B and an interpolation coefficient calculation unit 48B as shown in FIG. 7C. The interpolation calculation unit 47B calculates interpolation coefficients (weighting coefficients) KB1 to KB4 using a predetermined interpolation method according to the interpolation control signal PHB. Further, the interpolation coefficient calculation unit 48B can calculate the B gradation value QB at the pixel interpolation position by executing an interpolation calculation according to the following equation.
QB = (KB1 × B41 + KB2 × B42 + KB3 × B43 + KB4 × B44) / SB

  Here, SB = KB1 + KB2 + KB3 + KB4.

  Next, the operation of the image processing unit 14 having the above configuration will be described below. In the following description, as shown in FIG. 3B, the R line sensor 23R, the G line sensor 23G, and the B line sensor 23B (hereinafter simply referred to as “line sensor 23R,” at a 1/3 pixel pitch (= d / 3). 23G, 23B ") is used, and the image reading apparatus 1 uses the original image of the read medium 2 at a scanning interval of 1 pixel pitch (= d). Shall operate to read

FIG. 8 is a diagram illustrating transition of the relative positions of the line sensors 23R, 23G, and 23B in a scanning period in which the linear image sensor 23 relatively moves in the sub-scanning direction X when reading an image from the read medium 2. In FIG. 8, the horizontal axis represents the elapsed time t, and the vertical axis represents the relative positions of the line sensors 23R, 23G, and 23B in the sub-scanning direction X. 8, time indicating timing for linear image sensor 23 reads an image on a pixel _{pitch, T -2, T -1, T} 0, T +1, T +2, indicated by _{T +3.} These time _{T -2, T -1, T 0} , T +1, T +2, the center position of each of the _{T +3} corresponding light-receiving element 23G (image reading position) P (n-2), P (n-1 ), P (n), P (n + 1), P (n + 2), and P (n + 3). K (n−1), K (n), and K (n + 1) in FIG. 8 indicate pixel interpolation positions.

  As shown in FIG. 8, the shift in the image acquisition position between the line sensors 23R, 23G, and 23B at each time does not become an integral multiple of the image reading pitch, but a shift that is a fraction (decimal part) times occurs. For this reason, it is necessary to correct the phase shift between the correction pixel signals R3, G3, and B3 by interpolation processing. As shown in FIG. 8, when the pixel interpolation position K (n) is set as the position of interest, the interpolation calculation units 47R, 47G, and 47B have positions P (n−1) of four points in the vicinity of the position of interest K (n). , P (n), P (n + 1), and P (n + 2) are used as reference pixels to execute the interpolation calculation, and the gradation values QR, QB, and QG at the pixel interpolation position K (n) are obtained. Calculate to correct phase shift.

9A to 9C are diagrams illustrating phase shifts between the outputs of the line delay units 32R, 32G, and 32B with respect to the pixel interpolation position K (n) in FIG. FIG. 9A shows output values R _n−1 , R _n , R _n , R _n , R _n , R ₍ R) of the line delay unit 32R corresponding to the image reading positions P (n−1), P (n), P (n + 1), and P (n + 2), respectively. FIG. 9B is a diagram illustrating phase states of a series of R _{n + 1} and R _{n + 2} , and FIG. 9B corresponds to image reading positions P (n−1), P (n), P (n + 1), and P (n + 2), respectively. FIG. 9C is a diagram showing a phase state of a series of output values G _n−1 , G _n , G _{n + 1} , and G _{n + 2} of the line delay unit 32G that performs image reading positions P (n−1) and P It is a figure which shows the phase state of the series of the output value _Bn-1 , _Bn , _{Bn + 1} , _{Bn + 2} of the line delay part 32B corresponding to (n), P (n + 1), and P (n + 2), respectively. 9A, 9B, and 9C show interpolation gradation values QR, QG, and QB at the pixel interpolation position K (n).

As shown in FIG. 9B, when the G component pixel interpolation position is set between the G component values G _n and G _{n + 1} , as shown in FIG. 9A, the R component pixel is set. The interpolation position is set to an internal dividing point that is separated from the position of the output value R _{n by} a distance of 0.833 and separated from the position of the output value R _{n + 1 by} a distance of 0.167, as shown in FIG. 9C. The pixel interpolation position of the B component is set to an internal dividing point separated from the position of the output value B _{n + 1 by} a distance of 0.167 and separated from the position of the output value B _{n + 2 by} a distance of 0.833.

  Examples of the interpolation calculation method include a known nearest neighbor method, a linear interpolation method, and a cubic function interpolation method. In the present embodiment, an interpolation calculation using a cubic function interpolation method will be described as an example, but other methods can also be used.

FIGS. 10A to 10C are diagrams for explaining a pixel interpolation method by a cubic function interpolation method. 10A is a diagram showing the positions and phase states of the interpolated gradation values Q _n−2 , Q _n−1 , Q _n , Q _{n + 1} , and Q _{n + 2} , and FIG. 10B is a line delay unit. FIG. 10C is a diagram showing the positions and phase states of output values P _n−1 , P _n , P _{n + 1} , and P _{n + 2} of any one of 32R, 32G, and 32B. FIG. 10C is an interpolation function used in the interpolation calculation. It is a graph which shows h (x). The horizontal axis of FIG. 10 (C) represents the distance x from the position of gray scale value Q _n to be interpolated (pixel interpolation position) K (n). The interpolation function h (x) is an even function that is distributed symmetrically with respect to the pixel interpolation position K (n) as a center and attenuates as the distance from the center increases. Therefore, h (−x) = h (x) is established. In this pixel interpolation method, weighting factors h (s + 1), h (s), h are added to reference values P _n−1 , P _n , P _{n + 1} , and P _{n + 2} at four points around the pixel interpolation position K (n). (1-s), multiplied by h a (2-s) respectively (weighting) can calculate the interpolated gray level values Q _n by normalizing the sum of the multiplication result. That is, the interpolated gradation value Q _n is calculated according to the following equation (1).

When no edge is detected, the weighting factors h (s + 1), h (s), h (1-s), and h (2-s) are set with reference to the pixel interpolation position K (n). More specifically, the weighting coefficients h (s + 1), h (s), h (1-s), and h (2-s) are reference pixel interpolation positions as shown in FIG. The distances s + 1, s, 1-s, and 2-s from the pixel interpolation position K (n) of the reference values P _n−1 , P _n , P _{n + 1} , and P _{n + 2} near four points of K (n) are input. Is calculated as the output of the interpolation function h (x). As the position of the reference value used for the interpolation calculation is further away from the pixel interpolation position K (n), the value of the weighting coefficient becomes smaller.

  The interpolation calculation units 47R, 47G, and 47B can generate a two-dimensional image in which the phase shift is corrected by executing the above-described interpolation calculation for each line. However, the conventional technique has a problem in that color balance occurs due to a large loss in the R, G, B gradation balance depending on the position of the pixel interpolation position K (n).

  FIG. 11 is a diagram illustrating the transition of the relative positions of the line sensors 23R, 23G, and 23B in the scanning period, as in FIG. The original image formed on the read medium 2 in FIG. 11 has an edge (gradation edge) Eg0 having a large gradation change in the sub-scanning direction X. The position of the gradation edge Eg0 substantially coincides with the pixel interpolation position K (n). Assuming that the interpolation control unit 39 does not function in such a case, there is a risk that the gradation balance of R, G, and B is greatly lost due to the above-described interpolation calculation and color misregistration occurs.

12A to 12D are graphs showing the relationship between the pixel position and the interpolation gradation value when it is assumed that the original image has the gradation edge Eg0 of FIG. 11 and the interpolation control unit 39 does not function. It is. As shown in FIGS. 12A to 12D, the gradation value sharply changes from 255 to zero at the gradation edge Eg0. FIG. 12A is a graph showing R component interpolation gradation values QR _n−1 , QR _n , QR _{n + 1} , QR _{n + 2} , and FIG. 12B is a G component interpolation gradation value QG _{n−. 1} , QG _n , QG _{n + 1} , and QG _{n + 2} , and FIG. 12C is a graph showing B component interpolation gradation values QB _n−1 , QB _n , QB _{n + 1} , and QB _{n + 2} . FIG. 12D is a graph showing the interpolated gradation values QR _n , QG _n , and QB _n of FIGS. 12A, 12B, and 12C.

As shown in FIG. 12 (A), the reference value _R n located on both sides of the gradation edge _EG0, right reference value _{R n + 1} of _{R n + 1} is closer and gradation edge pixel interpolation position K (n) It is in the vicinity of Eg0. For this reason, the value of the weighting coefficient to be multiplied by the reference value R _{n + 1} on the right side is larger than the value of the weighting coefficient to be multiplied on the reference value R _n on the left side, and the interpolated gradation value QR _n is near zero. As shown in FIG. 12B, the reference values G _n and G _{n + 1} located on both sides of the gradation edge Eg0 are substantially equidistant from the pixel interpolation position K (n). For this reason, the values of the weighting coefficients to be multiplied by these reference values G _n and G _{n + 1} are almost equal, and the interpolation gradation value QG _n is about 128. Then, as shown in FIG. 12 (C), the reference value _B n located on both sides of the gradation edge _EG0, reference values _{B n} on the left of _{B n + 1} is closer to the pixel interpolation position K (n). Therefore, the values of the weighted coefficients to be multiplied to the left of the reference value B _n is greater than the value of the weighted coefficients to be multiplied to the right of the reference values B _{n + 1,} the interpolation gray-scale value QB _n becomes around 255. Accordingly, as shown in FIG. 12D, a shift (color shift) occurs between the interpolated gradation values QR _n , QG _n , and QB _n , and the gradation balance of R, G, and B is greatly lost. For interpolation gradation value QB _n of B is relatively large, so that the color shift bluish occurs.

  On the other hand, in this embodiment, since the interpolation control unit 39 functions, color misregistration can be suppressed.

The R edge detection unit 37R, the G edge detection unit 37G, and the B edge detection unit 37B calculate the absolute value of the difference between the reference values before and after the pixel interpolation position K (n), and the absolute value is calculated based on a predetermined threshold value. If it is larger, it is determined that there is an edge, and the values V _EDR , V _EDG , and V _EDB of the determination signals EDR, EDG, and EDB are set to 1 (active level). Specifically, as shown below, the values V _EDR , V _EDG , and V _EDB of the determination signals EDR, EDG, and _EDB are set.
If | R _n −R _{n + 1} |> Thr, V _EDR = 1,
If | G _n −G _{n + 1} |> Thg, V _EDG = 1,
When | B _n −B _{n + 1} |> Thb, V _EDB = 1.

On the other hand, when the absolute value is equal to or less than the predetermined threshold value, the values V _EDR , V _EDG , and V _EDB of the determination signals EDR, EDG, and EDB are set to zero. The threshold values Thr, Thg, Thb may be set individually according to the characteristics of each color. For example, the G component contains a lot of luminance components, and the gradation change portion of the original image is easily erroneously recognized as an edge. For this reason, the threshold value Thg may be set to a smaller value than the other threshold values Thr and Thb.

  In this embodiment, the R edge detection unit 37R, the G edge detection unit 37G, and the B edge detection unit 37B all determine the presence or absence of a color edge based on the delayed pixel signals for two lines. However, the present invention is not limited to this. The configurations of the R edge detection unit 37R, the G edge detection unit 37G, and the B edge detection unit 37B may be appropriately changed so as to determine the presence / absence of a color edge based on delayed pixel signals for more than three lines.

The edge determination unit 38 determines that there is an edge and determines an edge determination signal ED when at least one of the values V _EDR , V _EDG , and V _EDB of the determination signals EDR, EDG, and EDB is 1 (active level). Is set to 1; otherwise, the value of the edge determination signal ED is set to zero. In addition, when the value of the edge determination signal ED is 1, the interpolation position control unit 44 in FIG. 6 calculates correction amounts Δr, Δg, and Δb that reduce color misregistration, and the value of the edge determination signal ED is zero. In this case, the correction amounts Δr, Δg, Δb are all set to zero.

  The edge determination unit 38 determines that there is an edge and sets the value of the edge determination signal ED to 1 only when all of the determination signals EDR, EDG, and EDB are 1 (active level). Good.

When the value of the edge determination signal ED is 1, the interpolation position control unit 44 can calculate the correction amounts Δr, Δg, Δb based on the above-described cubic function interpolation method (FIG. 10). The interpolation position control unit 44 compares one of the interpolation gradation values QR _n and QB _n generated from the delayed pixel signals R41 to R44 and B41 to B44 as compared with the case where the correction amounts Δr, Δg, and Δb are zero values. both the correction amount Δr to approach the interpolated tone values QG _n generated from the delayed pixel signal G41~G44, Δg, sets the [Delta] b.

FIGS. 13A to 13C are diagrams for explaining an interpolation method for the R component when the gradation edge Eg0 of FIG. 11 exists. FIG. 13A is a diagram showing the positions and phase states of the interpolated gradation values QR _n−2 , QR _n−1 , QR _n , QR _{n + 1} , and QR _{n + 2} , and FIG. 13B is a line delay unit. FIG. 13C is a diagram showing the positions and phase states of the output values R _n−1 , R _n , R _{n + 1} , and R _{n + 2} of 32R, and FIG. 13C shows an interpolation function h (x used in the interpolation calculation for the R component -Δr). As shown in FIG. 13C, the interpolation processing unit 33R has four reference values R _n−1 in the vicinity with the interpolation position shifted from the pixel interpolation position by the correction amount Δr (<0) as a reference (center). , R _n , R _{n + 1} , R _{n + 2} , an interpolation calculation is performed to generate an interpolation gradation value QR _n . At this time, the interpolation coefficient calculation unit 48R in FIG. 7A outputs the output values h (2-qr) and h (1-qr) of the interpolation function h (x−Δr) as the weighting coefficients KR1, KR2, KR3, and KR4. , H (qr), h (qr + 1) are calculated (qr = r + Δr).

On the other hand, FIGS. 14A to 14C are diagrams for explaining the interpolation method of the B component when the gradation edge Eg0 of FIG. 11 exists. FIG. 14A is a diagram illustrating the positions and phase states of the interpolated gradation values QB _n−2 , QB _n−1 , QB _n , QB _{n + 1} , and QB _{n + 2} , and FIG. 14B is a line delay unit. FIG. 14C is a diagram showing the positions and phase states of the output values B _n−1 , B _n , B _{n + 1} , and B _{n + 2} of 32B. FIG. 14C shows an interpolation function h (x used in the interpolation calculation for the B component. -Δb). As shown in FIG. 14C, the interpolation processing unit 33B has four reference values B _n−1 in the vicinity with the interpolation position shifted by the correction amount Δb (> 0) from the pixel interpolation position as a reference (center). , B _n , B _{n + 1} , B _{n + 2} , an interpolation calculation is performed to generate an interpolation gradation value QB _n . At this time, the interpolation coefficient calculation unit 48B in FIG. 7C outputs the output values h (2-qb) and h (1-qb) of the interpolation function h (x−Δb) as the weighting coefficients KB1, KB2, KB3, and KB4. , H (qb), h (qb + 1) are calculated (qb = b + Δb).

The interpolation coefficient calculation unit 48G in FIG. 7B sets the G component interpolation amount Δg to zero, so that the interpolation processing unit 33G refers to the four neighboring points centered on the normal pixel interpolation position. An interpolation operation is performed on the values G _n−1 , G _n , G _{n + 1} , and G _{n + 2} to generate an interpolation gradation value QG _n .

FIG. 15 shows the interpolation gradation values QR _n , QG _n , and QB _n of the R component, the G component, and the B component that are generated when the gradation edge Eg0 of FIG. 11 exists and the interpolation control unit 39 functions. It is a graph. As shown in FIG. 15, it can be seen that the interpolation gradation values QR _n and QB _n are closer to the interpolation gradation value QG _n as compared with the case of FIG. Thereby, compared with the case of FIG. 12D, the gradation balance of R, G, and B is improved, and the color shift is reduced.

  The reason why the correction amount Δg for the G component is set to zero is that the G component includes more luminance components than the R component and the B component, and therefore the interpolation center position for the G component deviates from the pixel interpolation position. This is because the change in the output image is conspicuous if the position is set to a different position. However, by shifting the G component interpolation position from the pixel interpolation position, it is also possible to change the brightness of the edge portion to make the edge portion clearer.

As described above, in the image reading apparatus 1 according to the first embodiment, when the edge detection unit 36 determines that there is an edge, the interpolation processing unit 33G outputs pixels to the outputs G41 to G44 of the line delay unit 32G. and performing interpolation calculation using a weighting factor KG1~KG4 the interpolation position K (n) is set as a reference for calculating the gradation value QG _n of the G component. On the other hand, the interpolation processing unit 33R, with respect to the outputs R41 to R44 of the line delay unit 32R, as shown in FIGS. 13A to 13C, the interpolation position shifted by the correction amount Δr from the pixel interpolation position. The interpolation calculation using the weighting coefficients h (2-qr), h (1-qr), h (qr), h (qr + 1) (= KR1 to KR4) set with reference to The adjustment value QR _n is calculated. Further, the interpolation processing unit 33B uses the interpolation position shifted from the pixel interpolation position by the correction amount Δb as a reference with respect to the outputs B41 to B44 of the line delay unit 32B, as shown in FIGS. 14 (A) to (C). B component gradation value by executing an interpolation operation using weighting coefficients h (2-qb), h (1-qb), h (qb), h (qb + 1) (= KB1 to KB4) set as QB _n is calculated. Therefore, as shown in FIG. 15, even when the phase shift is corrected by the interpolation calculation by the interpolation processing units 33R, 33G, and 33B, the color shift between the gradation values QR _n , QG _n , and QB _n can be suppressed. it can. Therefore, it is possible to obtain an effect that a good color reproducibility can be realized by improving the gradation balance of R, G, and B and suppressing color shift.

  The edge detection unit 36 determines the presence / absence of a color edge for each series from a plurality of series of signals obtained from the plurality of line sensors 23R, 23G, and 23B, and determines the final edge based on the determination result. Since the presence or absence is determined, the edge determination accuracy can be increased. Further, since the threshold values Thr, Thg, Thb can be individually set according to the characteristics for each color, edge determination according to the reading situation and purpose can be performed. For example, the threshold Thg can be adjusted to facilitate the edge determination of a G image that includes a large amount of luminance components.

  In addition, even when the linear image sensor 23 having a structure in which the line sensors 23R, 23G, and 23B are arranged at one pixel pitch as shown in FIG. 3A is used, the line sensor as shown in FIG. It is possible to obtain the same effect as when the linear image sensor 23 having a structure in which the sensors 23R, 23G, and 23B are arranged at a 1/3 pixel pitch is used.

Embodiment 2. FIG.
Next, a second embodiment according to the present invention will be described. FIG. 16 is a functional block diagram illustrating a schematic configuration of the image processing unit 14D according to the second embodiment. The configuration of the image reading apparatus of the second embodiment is the same as that of the image reading apparatus 1 (FIG. 1) of the first embodiment except for the image processing unit 14D of FIG.

  As shown in FIG. 16, the image processing unit 14D is similar to the image processing unit 14 of the first embodiment, and includes a shading correction unit 31, line delay units 32R, 32G, and 32B, and interpolation processing units 33R, 33G, and 33B. And an edge detector 36. The image processing unit 14D of the present embodiment further includes an edge position detection unit 49 and an interpolation control unit 39D.

  The edge position detection unit 49 detects the position of the edge when it is determined that there is an edge in accordance with the determination signals EDR, EDG, EDB and the edge determination signal ED supplied from the edge detection unit 36. Edge position information SEP indicating the detection result is given to the interpolation control unit 39D. The interpolation control unit 39D has a function of controlling the interpolation position used in the interpolation calculation according to the edge detection position.

  FIG. 17 is a functional block diagram showing a schematic configuration of the interpolation control unit 39D of the present embodiment. As shown in FIG. 17, the interpolation control unit 39D, like the interpolation control unit 39 in the first embodiment, interpolates positions for determining parameters r, g, and b indicating pixel interpolation positions in accordance with the control signal CL. A determination unit 43 is included. The interpolation control unit 39D further includes an interpolation position control unit 44D that calculates correction amounts Δr, Δg, and Δb for correcting the parameters r, g, and b according to the edge position information SEP, and adders 45R, 45G, 45B. The adder 45R adds the correction amount Δr to the parameter r and outputs an interpolation control signal PHR indicating the addition result. Similarly, the adder 45G adds the correction amount Δg to the parameter g and outputs an interpolation control signal PHG indicating the addition result, and the adder 45B adds the correction amount Δb to the parameter b and outputs the addition result. The interpolation control signal PHB shown is output.

  Next, the operation of the image processing unit 14D having the above configuration will be described below. In the following description, a linear image sensor 23 having a structure in which line sensors 23R, 23G, and 23B are arranged at a 1/3 pixel pitch (= d / 3) as shown in FIG. 3B is used. The image reading apparatus according to the present embodiment is assumed to operate so as to read the original image of the read medium 2 at a scanning interval of one pixel pitch (= d).

First, when the gradation edge Eg0 that substantially coincides with the pixel interpolation position K (n) is formed as shown in FIG. 11, the following three inequalities hold simultaneously.
| R _n −R _{n + 1} |> Thr,
| G _n −G _{n + 1} |> Thg, and
| B _n −B _{n + 1} |> Thb.

Therefore, the values V _EDR , V _EDG , and V _EDB of the determination signals EDR, EDG, and _EDB are as follows.
V _EDR = 1, V _EDG = 1, and V _EDB = 1.

The value V _SEP of the edge position information SEP is set to “7”. If it is assumed that the interpolation control unit 39D does not function at this time, color misregistration occurs as shown in FIG. 12D, and the gradation balance of R, G, and B is greatly lost. On the other hand, as shown in FIGS. 14A and 14B, the interpolation calculation is performed using the weighting coefficient based on the interpolation position that is shifted from the pixel interpolation position by the correction amounts Δr and Δb. The gradation balance of B can be improved. At this time, the absolute value | Δr | of the R component correction amount Δr is considered to be substantially equal to the absolute value | Δb | of the B component correction amount Δb.

Next, FIG. 18 is a diagram illustrating the transition of the relative position of the line sensors 23R, 23G, and 23B in the scanning period, as in FIG. The original image formed on the read medium 2 in FIG. 18 has a gradation edge Eg1 having a large gradation change in the sub-scanning direction X. The gradation edge Eg1 is, G line sensor 23G at time _{T +1} is formed between the upper end of the lower end and the B line sensor 23B of the G line sensor 23G when passing through the image reading position P (n + 1). At this time, the following inequality holds simultaneously.
| R _n −R _{n + 1} | <Thr,
| G _n −G _{n + 1} | <Thg and | B _n −B _{n + 1} |> Thb.

Therefore, the values of the determination signals EDR, EDG, and EDB are as follows.
V _EDR = 0, V _EDG = 0, and V _EDB = 1.

In this case, the value V _SEP of the edge position information SEP is set to “1”.

19A to 19D are graphs showing the relationship between the pixel position and the interpolation gradation value when it is assumed that the original image has the gradation edge Eg1 of FIG. 18 and the interpolation control unit 39D does not function. It is. As shown in FIGS. 19A to 19D, the gradation value sharply changes from 255 to zero at the gradation edge Eg1. FIG. 19A is a graph showing R component interpolation gradation values QR _n−1 , QR _n , QR _{n + 1} , QR _{n + 2} , and FIG. 19B is a G component interpolation gradation value QG _{n−. 1} , QG _n , QG _{n + 1} , QG _{n + 2} , and FIG. 19C is a graph showing B component interpolation gradation values QB _n−1 , QB _n , QB _{n + 1} , QB _{n + 2} . FIG. 19D is a graph showing the interpolated gradation values QR _n , QG _n , and QB _n of FIGS. 19A, 19B, and 19C.

As shown in FIG. 19 (A), the reference value _R n located on both sides of the gradation edge _Eg1, the reference value _{R n} of the left of the _{R n + 1} is in the vicinity of the pixel interpolation position K (n). Therefore, the interpolation gray-scale value QB _n is slightly smaller than 255. On the other hand, as shown in FIGS. 19A and 19B, the interpolation gradation values QR _n and QG _n of the R component and the G component are approximately 255. Accordingly, as shown in FIG. 19D, a shift (color shift) occurs between the interpolation gradation values QR _n and QG _n and the interpolation gradation value QB _n, and the gradations of R, G, and B are generated. Balance is lost. In this case, the interpolation position controller 44D, by calculating the correction amount Δb to approach the interpolated tone values QB _n as shown in Figure 20 to the interpolation gray-scale value QR _n, R, G, floor B Tonal balance can be improved. At this time, the absolute value | Δr | of the R component correction amount Δr is considered to be smaller than the absolute value | Δb | of the correction amount Δb, and the absolute value | Δr | is substantially equal to zero.

Next, FIG. 21 is a diagram illustrating the transition of the relative positions of the line sensors 23R, 23G, and 23B in the scanning period, as in FIG. The original image formed on the read medium 2 in FIG. 21 has a gradation edge Eg2 having a large gradation change in the sub-scanning direction X. The gradation edge Eg2 is, G line sensor 23G at time _{T +1} is formed between the lower end of the upper end and the R line sensor 23R of G line sensor 23G when passing through the image reading position P (n + 1). At this time, the following inequality holds simultaneously.
| R _n −R _{n + 1} | <Thr,
| G _n −G _{n + 1} |> Thg and | B _n −B _{n + 1} |> Thb.

  Therefore, the values of the determination signals EDR, EDG, and EDB are as follows.

V _EDR = 0, V _EDG = 1, and V _EDB = 1.
In this case, the value V _SEP of the edge position information SEP is set to “3”.

22A to 22D are graphs showing the relationship between the pixel position and the interpolation gradation value when it is assumed that the original image has the gradation edge Eg2 of FIG. 21 and the interpolation control unit 39D does not function. It is. As shown in FIGS. 22A to 22D, the gradation value sharply changes from 255 to zero at the gradation edge Eg2. 22A is a graph showing R component interpolation gradation values QR _n−1 , QR _n , QR _{n + 1} , and QR _{n + 2} , and FIG. 22B is a G component interpolation gradation value QG _{n−. 1} , QG _n , QG _{n + 1} , QG _{n + 2} , and FIG. 22C is a graph showing B component interpolated gradation values QB _n−1 , QB _n , QB _{n + 1} , QB _{n + 2} . FIG. 22D is a graph showing the interpolated gradation values QR _n , QG _n , and QB _n of FIGS. 22A, 22B, and 22C.

As shown in FIG. 22 (A), the reference value _{R n + 1, R} n + reference value _{R n + 1} of the left of the _two located on either side of the gradation edge Eg2, since in the vicinity of the pixel interpolation position K (n), The interpolated gradation value QR _n is around 255. Further, as shown in FIG. 22 (C), the reference value _B n located on both sides of the gradation edge _Eg2, reference values _{B n} on the left of _{B n + 1} is in the vicinity of the pixel interpolation position K (n) Therefore, the interpolation gray-scale value QB _n is the value slightly smaller than 255. In contrast, as shown in FIG. 22 (B), the interpolation gray-scale value QB _n of the G component is around 128. Therefore, as shown in FIG. 22D, a shift (color shift) occurs between the interpolation gradation values QR _n and QB _n and the interpolation gradation value QG _n, and the gradations of R, G, and B are generated. Balance is lost. In this case, the interpolation position control unit 44D calculates the correction amounts Δr and Δb so that the interpolation gradation values QR _n and QB _n approach the interpolation gradation value QR _n as shown in FIG. The gradation balance of G and B can be improved. At this time, it is considered that the absolute value | Δr | of the correction amount Δr is larger than the absolute value | Δb | of the correction amount Δb.

Next, FIG. 24 is a diagram illustrating the transition of the relative positions of the line sensors 23R, 23G, and 23B in the scanning period, as in FIG. The original image formed on the read medium 2 in FIG. 24 has a gradation edge Eg3 having a large gradation change in the sub-scanning direction X. The gradation edge Eg3 is, G line sensor 23G at time T ₀ is formed between the upper end of the lower end and the B line sensor 23B of the G line sensor 23G when passing through the image reading position P (n). At this time, the following inequality holds simultaneously.
| R _n −R _{n + 1} |> Thr,
| G _n −G _{n + 1} |> Thg and | B _n −B _{n + 1} | <Thb.

Therefore, the values of the determination signals EDR, EDG, and EDB are as follows.
V _EDR = 1, V _EDG = 1, and V _EDB = 0.

In this case, the value V _SEP of the edge position information SEP is set to “6”.

25A to 25D are graphs showing the relationship between the pixel position and the interpolation gradation value when it is assumed that the original image has the gradation edge Eg3 of FIG. 24 and the interpolation control unit 39D does not function. It is. As shown in FIGS. 25A to 25D, the gradation value sharply changes from 255 to zero at the gradation edge Eg3. FIG. 25A is a graph showing R component interpolation gradation values QR _n−1 , QR _n , QR _{n + 1} , and QR _{n + 2} , and FIG. 25B is a G component interpolation gradation value QG _{n−. 1} , QG _n , QG _{n + 1} , QG _{n + 2} , and FIG. 25C is a graph showing B component interpolated gradation values QB _n−1 , QB _n , QB _{n + 1} , QB _{n + 2} . FIG. 25D is a graph showing the interpolated gradation values QR _n , QG _n , and QB _n of FIGS. 25A, 25B, and 25C.

As shown in FIG. 25 (A), the reference value _R n located on both sides of the gradation edge _Eg3, the right of the reference value _{R n + 1} of _{R n + 1} is closer to the pixel interpolation position K (n), the interpolation Floor The adjustment value QR _n is a value slightly larger than zero. Further, the vicinity of 25 as shown (C), the right side of the reference values _{B n} of the reference values _{B n-1, B n} positioned at both sides of the gradation edge Eg3, the pixel interpolation position K (n) Therefore, the interpolation gradation value QB _n is almost zero. In contrast, as shown in FIG. 25 (B), the interpolation gray-scale value QB _n of the G component is around 128. Therefore, as shown in FIG. 25D, a shift (color shift) occurs between the interpolation gradation values QR _n and QB _n and the interpolation gradation value QG _n, and the gradations of R, G, and B are generated. Balance is lost. In this case, the interpolation position control unit 44D calculates the correction amounts Δr and Δb so that the interpolation gradation values QR _n and QB _n approach the interpolation gradation value QR _n as shown in FIG. The gradation balance of G and B can be improved. At this time, the absolute value | Δr | of the correction amount Δr is considered to be smaller than the absolute value | Δb | of the correction amount Δb.

Next, FIG. 27 is a diagram illustrating the transition of the relative positions of the line sensors 23R, 23G, and 23B in the scanning period, as in FIG. The original image formed on the read medium 2 in FIG. 27 has a gradation edge Eg4 having a large gradation change in the sub-scanning direction X. The gradation edge Eg4 is, G line sensor 23G at time T ₀ is formed between the lower end of the upper end and the R line sensor 23R of G line sensor 23G when passing through the image reading position P (n). At this time, the following inequality holds simultaneously.
| R _n −R _{n + 1} |> Thr,
| G _n −G _{n + 1} | <Thg and | B _n −B _{n + 1} | <Thb.

Therefore, the values of the determination signals EDR, EDG, and EDB are as follows.
V _EDR = 1, V _EDG = 0, and V _EDB = 0.

In this case, the value V _SEP of the edge position information SEP is set to “4”.

28A to 28D are graphs showing the relationship between the pixel position and the interpolation gradation value when it is assumed that the original image has the gradation edge Eg4 of FIG. 27 and the interpolation control unit 39D does not function. It is. As shown in FIGS. 28A to 28D, the gradation value sharply changes from 255 to zero at the gradation edge Eg4. FIG. 28A is a graph showing R component interpolation gradation values QR _n−1 , QR _n , QR _{n + 1} , and QR _{n + 2} , and FIG. 28B is a G component interpolation gradation value QG _{n−. 1} , QG _n , QG _{n + 1} , and QG _{n + 2} , and FIG. 28C is a graph showing B component interpolation gradation values QB _n−1 , QB _n , QB _{n + 1} , and QB _{n + 2} . FIG. 28D is a graph showing the interpolated gradation values QR _n , QG _n , and QB _n in FIGS. 19A, 19B, and 19C.

As shown in FIG. 28 (A), the reference value _R n located on both sides of the gradation edge _Eg4, the right of the reference value _{R n + 1} of _{R n + 1,} since in the vicinity of the pixel interpolation position K (n), The interpolated gradation value QR _n is a value slightly larger than zero. On the other hand, as shown in FIGS. 22B and 22C, the interpolation gradation values QG _n and QB _n are almost zero. Therefore, as shown in FIG. 28D, a shift (color shift) occurs between the interpolation gradation values QB _n and QG _n and the interpolation gradation value QR _n, and the R, G, and B gradations are generated. Balance is slightly lost. In this case, the interpolation position control unit 44D calculates the correction amount Δr so that the interpolation gradation value QR _n approaches the interpolation gradation value QG _n as shown in FIG. Tonal balance can be improved. At this time, the absolute value | Δr | of the correction amount Δr is larger than the absolute value | Δb | of the correction amount Δb, and the absolute value | Δb | is considered to be almost zero.

Next, FIG. 30 is a diagram illustrating the transition of the relative positions of the line sensors 23R, 23G, and 23B in the scanning period, as in FIG. The original image formed on the read medium 2 in FIG. 30 has a gradation edge Eg5 having a large gradation change in the sub-scanning direction X. The gradation edge Eg5 is, G line sensor 23G at time _{T 0} is formed at the upper end near the G line sensor 23G when passing through the image reading position P (n). At this time, the following inequality holds simultaneously.
| R _n −R _{n + 1} | <Thr,
| G _n −G _{n + 1} | <Thg and | B _n −B _{n + 1} | <Thb.

Therefore, the values of the determination signals EDR, EDG, and EDB are as follows.
V _EDR = 0, V _EDG = 0, and V _EDB = 0.

In this case, the value V _SEP of the edge position information SEP is set to “0”.

31A to 31D are graphs showing the relationship between the pixel position and the interpolation gradation value when it is assumed that the original image has the gradation edge Eg5 of FIG. 30 and the interpolation control unit 39D does not function. It is. As shown in FIGS. 31A to 31D, the gradation value sharply changes from 255 to zero at the gradation edge Eg5. FIG. 31A is a graph showing R component interpolation gradation values QR _n−1 , QR _n , QR _{n + 1} , and QR _{n + 2} , and FIG. 31B is a G component interpolation gradation value QG _{n−. 1} , QG _n , QG _{n + 1} , QG _{n + 2} , and FIG. 31C is a graph showing the B component interpolation gradation values QB _n−1 , QB _n , QB _{n + 1} , QB _{n + 2} . FIG. 31D is a graph showing the interpolated gradation values QR _n , QG _n , and QB _n of FIGS. 31A, 31B, and 31C.

As shown in FIGS. 31A to 31C, the interpolated gradation values QR _n , QG _n , and QB _n are all near zero. Therefore, almost no color misregistration occurs as shown in FIG. In this case, the interpolation position controller 44D may set all the correction amounts Δr, Δb, Δg to zero values.

Next, FIG. 32 is a diagram illustrating the transition of the relative positions of the line sensors 23R, 23G, and 23B in the scanning period, as in FIG. The original image formed on the read medium 2 in FIG. 32 has a gradation edge Eg6 having a large gradation change in the sub-scanning direction X. The gradation edge Eg6 is, G line sensor 23G at time _{T +1} is formed at the lower end near the B line sensor 23B when passing through the image reading position P (n + 1). At this time, the edge detector 36 does not detect an edge between the reference positions P (n) and P (n + 1). Therefore, the values V _EDR , V _EDG , and V _EDB of the determination signals EDR, EDG, and _EDB are all set to zero. In this case, the value V _SEP of the edge position information SEP is set to “0”.

33A to 33D are graphs showing the relationship between the pixel position and the interpolation gradation value when the original image has the gradation edge Eg6 of FIG. 32 and the interpolation control unit 39D does not function. As shown in FIGS. 33A to 33D, the gradation value sharply changes from 255 to zero at the gradation edge Eg6. FIG. 33A is a graph showing R component interpolation gradation values QR _n−1 , QR _n , QR _{n + 1} , and QR _{n + 2} , and FIG. 33B is a G component interpolation gradation value QG _{n−. 1} , QG _n , QG _{n + 1} , and QG _{n + 2} , and FIG. 33C is a graph showing B component interpolated gradation values QB _n−1 , QB _n , QB _{n + 1} , and QB _{n + 2} . FIG. 33 (D) is a graph showing the interpolated gradation values QR _n , QG _n , and QB _n of FIGS. 33 (A), (B), and (C).

As shown in FIGS. 33A to 33C, the interpolation gradation values QR _n , QG _n , and QB _n are all in the vicinity of 255. Therefore, almost no color misregistration occurs as shown in FIG. In this case, the interpolation position controller 44D may set all the correction amounts Δr, Δb, Δg to zero values.

From the above, the correspondence relationship between the position of the gradation edge and the combination of the values V _EDR , V _EDG , and V _EDB of the determination signals EDR, EDG, and EDB is arranged as shown in the following table.

  The edge position detection unit 49 can store the information in this table in advance as a lookup table, and can generate the edge position information SEP in accordance with the lookup table.

  In the present embodiment, the threshold values Thr, Thg, Thb may be set to values according to the characteristics of the R, G, B colors. For example, the G component includes a relatively large luminance component, and the gradation change is easily recognized as an edge. Therefore, the threshold value Thg may be made smaller than the threshold values Thr and Thb.

  In the present embodiment, the correction amount Δg for the G component is always set to a zero value. The reason why the correction amount Δg is set to zero is that the G component includes more luminance components than the R component and B component, and therefore the G component interpolation center position is set to a position shifted from the pixel interpolation position. This is because the change in the output image is conspicuous. However, by shifting the G component interpolation position from the pixel interpolation position, it is also possible to change the brightness of the edge portion to make the edge portion clearer.

  As described above, in the second embodiment, the interpolation control unit 39D generates the interpolation control signals PHR, PHG, and PHB according to the edge position information SEP. Even in this case, the same effect as in the first embodiment can be obtained.

  In this embodiment, as shown in FIG. 3A, the linear image sensor 23 having a structure in which the line sensors 23R, 23G, and 23B are arranged at one pixel pitch is used. ), The same effect as when the linear image sensor 23 having a structure in which the line sensors 23R, 23G, and 23B are arranged at a 1/3 pixel pitch can be obtained.

Embodiment 3 FIG.
Next, a third embodiment according to the present invention will be described. FIG. 34 is a functional block diagram illustrating a schematic configuration of the image processing unit 14E according to the third embodiment. The configuration of the image reading apparatus of the third embodiment is the same as that of the image reading apparatus 1 (FIG. 1) of the first embodiment except for the image processing unit 14E of FIG.

  As shown in FIG. 34, the image processing unit 14E is similar to the image processing unit 14D (FIG. 16) of the second embodiment, and includes a shading correction unit 31, line delay units 32R, 32G, and 32B, and an interpolation processing unit 33R. , 33G, 33B and an edge position detector 49. The image processing unit 14E of the present embodiment further includes an edge detection unit 36E and an interpolation control unit 39E. The configuration of the image processing unit 14E is the same as the configuration of the image processing unit 14D of the second embodiment except for the edge detection unit 36E and the interpolation control unit 39E.

  The edge detection unit 36E includes an R edge detection unit 37ER, a G edge detection unit 37EG, a B edge detection unit 37EB, and an edge determination unit 38. The function of the edge determination unit 38 in FIG. 34 is the same as the function of the edge determination unit 38 in FIG.

The R edge detection unit 37ER, the G edge detection unit 37EG, and the B edge detection unit 37EB are similar to the R edge detection unit 37R, the G edge detection unit 37G, and the B edge detection unit 37B of the first embodiment. Each has a function of generating EDB, and further has a function of calculating the magnitude (absolute value) of the gradation change amount of each of the red edge, the green edge, and the blue edge. Specifically, the R edge detection unit 37ER, the G edge detection unit 37EG, and the B edge detection unit 37EB respectively calculate the red edge, the green edge, and the blue edge between two consecutive lines in the sub-scanning direction X according to the following formula. Values V _EVR , V _EVG , and V _EVB proportional to the magnitude (absolute value) of the tone change amount can be calculated.
V _EVR = KR × | R _n −R _{n + 1} |
V _EVG = KG × | G _n −G _{n + 1} |
V _EVB = KB × | B _n −B _{n + 1} |

  Here, KR, KG, and KB are proportional constants.

R edge detector 37ER, G edge detector 37EG and B edge detector 37EB supplies value _{V EVR,} V _{EVG, V EVB} edge detection information indicating each EVR, EVG, the EVB the interpolation control unit 39E. The interpolation control unit 39E has a function of calculating the correction amounts Δr, Δg, Δb according to the edge detection information EVR, EVG, EVB and controlling the interpolation position.

FIG. 35 is a functional block diagram schematically showing the configuration of the interpolation control unit 39E of the present embodiment. As illustrated in FIG. 35, the interpolation control unit 39E includes an interpolation position determination unit 43 and an interpolation position control unit 44E. The interpolation position control unit 44E receives the edge position information SEP and the edge detection information EVR, EVG, EVB as input, and similarly to the interpolation position control unit 44D of the second embodiment, the correction amount Δr, It has a function of calculating Δg, Δb and a function of changing the correction amounts Δr, Δg, Δb in accordance with the edge detection information EVR, EVG, EVB. For example, the interpolation position control unit 44E increases the amount of change of the correction amounts Δr, Δg, Δb as the values V _EVR , V _EVG , V _EVB are larger, and corrects as the values V _EVR , V _EVG , V _EVB are smaller. The amount of change of the amounts Δr, Δg, Δb can be reduced.

  As described above, in the third embodiment, the correction amounts Δr, Δg, and Δb can be dynamically controlled according to the magnitude of the detected gradation change amount of the edge. Compared to the case, the R, G, B gradation balance can be further improved, and the color shift can be further suppressed.

Embodiment 4 FIG.
Next, a fourth embodiment according to the present invention will be described. All or part of the functions of the image processing units 14, 14D, and 14E according to the first to third embodiments can be realized by hardware resources, but may be realized by cooperation of hardware resources and software. This function can also be realized by a computer program executed by a microprocessor including a CPU (central processing unit). When a part of the function is realized by a computer program, the microprocessor can realize a part of the function by loading and executing the computer program from a computer-readable recording medium.

  Such a computer program may be provided by being recorded on a computer-readable recording medium such as an optical disk, or may be provided via a communication line such as the Internet.

  FIG. 36 is a functional block diagram of an arithmetic unit 14F showing a configuration when the functions of the image processing units 14, 14D, and 14E according to the first to third embodiments are realized by a computer program. As shown in FIG. 36, the arithmetic device 14F includes a processor 51 including a CPU, a RAM (Random Access Memory) 52, a nonvolatile memory 53, a large-capacity recording medium 54, an input / output interface 55, and a bus 56. . As the non-volatile memory 53, for example, a flash memory can be used. Further, as the large-capacity recording medium 54, for example, a hard disk (magnetic disk) or an optical disk can be used.

  The input / output interface 55 transfers the digital signals R2, G2, and B2 transferred from the A / D converter 13 of FIG. 1 to the processor 51 and the image signals Rout, Gout, and Bout transferred from the processor 51 to the outside. And a function of giving a control signal transferred from the controller 15 to the processor 51.

  The processor 51 can realize any of the functions of the image processing units 14, 14 </ b> D, and 14 </ b> E by loading and executing a computer program from the nonvolatile memory 53 or the large-capacity recording medium 54.

  FIG. 37 is a flowchart illustrating an example of a processing procedure according to the fourth embodiment in the case where the processor 51 executes a computer program that implements the function of the image processing unit 14 according to the first embodiment.

  First, the processor 51 performs shading correction on the input digital signals R2, G2, and B2 to generate corrected pixel signals R3, G3, and B3, similarly to the shading correction unit 31 in FIG. 4 (step S11). Next, similarly to the line delay units 32R, 32G, and 32B, the processor 51 delays the correction pixel signals R3, G3, and B3 to correct the phase shift between the correction pixel signals R3, G3, and B3 in units of lines. (Step S12). Next, the processor 51 performs color edge detection processing in the same manner as the R edge detection unit 37R, G edge detection unit 37G, and B edge detection unit 37B, and generates determination signals EDR, EDG, and EDB that indicate the presence or absence of color edges. (Step S13). Next, similarly to the edge determination unit 38, the processor 51, based on the determination signals EDR, EDG, and EDB, determines the sub-image in the original image of the read medium 2 according to the combination of the presence / absence of red edge, green edge, and blue edge. The presence / absence of an edge in the scanning direction X is determined, and an edge determination signal ED indicating the determination result is generated (step S14).

  Thereafter, the processor 51 calculates the parameters r, g, and b indicating the pixel interpolation position in the same manner as the interpolation position determination unit 43 (step S15). Next, similarly to the interpolation position control unit 44 of the interpolation control unit 39, the processor 51 calculates correction amounts Δr, Δg, Δb corresponding to the R, G, B interpolation positions (step S16). Next, when it is determined that there is an edge (YES in step S17), the processor 51 corrects one of the parameters r, g, and b indicating the pixel interpolation position to interpolate the control signals PHR, PHG, and PHB. Is generated (step S18). On the other hand, if it is determined that there is no edge, the process proceeds to the next step S18.

  Then, similarly to the interpolation processing units 33R, 33G, and 33B, the processor 51 calculates interpolation coefficients based on the interpolation control signals PHR, PHG, and PHB (step S19), and executes an interpolation operation using these interpolation coefficients. The image signals Rout, Gout, and Bout are generated (step S20).

Embodiment 5 FIG.
Next, Embodiment 5 according to the present invention will be described with reference to FIG. FIG. 38 shows an example of a processing procedure according to the fifth embodiment when the processor 51 executes a computer program for realizing the function of the image processing unit 14D of the second embodiment using the arithmetic device 14F of FIG. It is a flowchart.

  Referring to FIG. 39, the processor 51 executes steps S11 to S14 as in the processing procedure according to the fourth embodiment. Next, similarly to the edge position detector 49 (FIG. 16), the processor 51 determines the position of the edge when it is determined that there is an edge according to the determination signals EDR, EDG, EDB and the edge determination signal ED. , And edge position information SEP indicating the detection result is generated (step S14D). Next, the processor 51 calculates the parameters r, g, and b indicating the pixel interpolation position, similarly to the interpolation position determination unit 43 (step S15). Further, similarly to the interpolation position control unit 44D (FIG. 17), the processor 51 calculates correction amounts Δr, Δg, Δb corresponding to the R, G, B interpolation positions based on the edge position information SEP (step). S16D).

  Thereafter, steps S17 to S20 are executed in the same manner as the processing procedure according to the fourth embodiment.

Embodiment 6 FIG.
Next, Embodiment 6 according to the present invention will be described with reference to FIG. FIG. 39 shows an example of a processing procedure according to the sixth embodiment when the processor 51 executes a computer program that realizes the function of the image processing unit 14E of the third embodiment using the arithmetic device 14F of FIG. It is a flowchart.

  Referring to FIG. 39, the processor 51 executes steps S11 and S12 in the same manner as the processing procedure according to the fourth embodiment. Next, the processor 51 generates determination signals EDR, EDG, and EDB, respectively, similarly to the R edge detection unit 37ER, G edge detection unit 37EG, and B edge detection unit 37EB, and further, a red edge, a green edge, and a blue edge The magnitude (absolute value) of each gradation change amount is calculated (step S13E).

  Thereafter, similarly to the edge determination unit 38, the processor 51, based on the determination signals EDR, EDG, and EDB, determines the sub-image in the original image of the read medium 2 in accordance with the combination of the presence of the red edge, the green edge, and the blue edge. The presence / absence of an edge in the scanning direction X is determined, and an edge determination signal ED indicating the determination result is generated (step S14). Next, the processor 51 calculates correction amounts Δr, Δg, Δb according to the edge position information SEP, and further corrects them according to the edge detection information EVR, EVG, EVB, similarly to the interpolation position control unit 44E. The amounts Δr, Δg, Δb are changed (step S16E).

  Thereafter, steps S17 to S20 are executed in the same manner as the processing procedure according to the fourth embodiment.

Modified examples of the first to sixth embodiments.
Although various embodiments according to the present invention have been described above with reference to the drawings, these are examples of the present invention, and various forms other than the above can be adopted. For example, in the first to sixth embodiments, the linear image sensor 23 has the three rows of line sensors 23R, 23G, and 23B that receive the three primary colors, respectively. However, the present invention is not limited to this. Alternatively, a line sensor that receives infrared light may be further included.

  The linear image sensor 23 is a contact image sensor, but is not limited thereto. It is also possible to appropriately change the configuration of the image reading apparatus according to any of the first to sixth embodiments so as to have an image sensor that uses a reduction imaging optical system for reducing and imaging the light reflected from the document on a plurality of line sensors. is there.

  Further, the edge detection unit 36 according to the first embodiment can be replaced with the edge detection unit 36D according to the second embodiment to configure an image processing unit of another form.

  A part of the configuration of the image processing units 14, 14D, and 14E can be realized by an LSI (Large Scale Integrated circuit). Also, part of the configuration of the image processing units 14, 14 </ b> D, and 14 </ b> E can be realized by an FPGA (Field-Programmable Gate Array) or an ASIC (Application Specific Integrated Circuit).

  In addition, this invention is not limited to the said embodiment, In the range which does not deviate from the summary of this invention, it can implement with a various aspect.

DESCRIPTION OF SYMBOLS 1 Image reader, 2 Medium to be read, 10 Translucent board, 11 Scanning drive part, 12 Imaging sensor module, 13 A / D converter (ADC), 14, 14D, 14E Image processing part, 14F arithmetic unit, 15 Controller, 16 operation panel, 21A, 21B light source, 22 magnification optical system (lens array), 23 linear image sensor, 23R R line sensor, 23G G line sensor, 23B B line sensor, 24R, 24G, 14B 31 Shading correction unit, 32R, 32G, 32B line delay unit, 33R, 33G, 33B interpolation processing unit, 36, 36E, edge detection unit, 37R, 37ERR R edge detection unit, 37G, 37EG G edge detection unit, 37B, 37EB B edge detection unit, 38 edge determination unit, 39, 39D, 39E Interpolation control unit, 40R _{1 to} 40R _N , 40G _{1 to} 40G _N , 40B _{1 to} 40B _N line memory, 41R, 41G, 41B selector, 43 Interpolation position determination unit, 44, 44D, 44E Interpolation position control unit, 45R, 45G , 45B adder, 47R, 47G, 47B interpolation calculation unit, 48R, 48G, 48B interpolation coefficient calculation unit, 49 edge position detection unit, 51 processor, 52 RAM, 53 nonvolatile memory, 54 large capacity recording medium, 55 input / output Interface, 56 buses.

Claims (20)

An image reading device for reading a two-dimensional image from a read medium extending in a main scanning direction and a sub-scanning direction different from the main scanning direction,
An image sensor including a plurality of line sensors arranged at predetermined intervals along the sub-scanning direction;
A scanning drive unit that causes the image sensor to receive light reflected by the surface of the read medium by moving the image sensor relative to the read medium in the sub-scanning direction;
Based on a plurality of series of read pixel signals output in parallel from the plurality of line sensor at regular time intervals within a scanning period in which the image sensor moves relative to the read medium in the sub-scanning direction. An image processing unit for generating image data,
The plurality of lines of line sensors photoelectrically convert light of a plurality of different colors, respectively, and output the plurality of series of read pixel signals in parallel.
The image processing unit
An interpolation position determining unit for determining a pixel interpolation position for correcting a phase shift between a read pixel signal of a reference series among the read pixel signals of the plurality of series and a read pixel signal of another series other than the reference series; ,
An interpolation processing unit that performs an interpolation operation for each of the plurality of read pixel signals and calculates a gradation value of each of the plurality of colors at the pixel interpolation position;
An edge detector that determines the presence or absence of an edge having a gradation change in the sub-scanning direction based on at least one series of read pixel signals among the plurality of series of read pixel signals;
When it is determined that the edge is present, the interpolation processing unit, with respect to the read pixel signal of the reference series, the first one of the pixel interpolation position and the position shifted from the pixel interpolation position. The interpolation calculation is performed using a weighting coefficient set with reference to the interpolation position of the second interpolation position, and a second interpolation position different from the first interpolation position is used as a reference for the read pixel signals of the other series. An image reading apparatus that performs the interpolation operation using a weighting coefficient set as.   The image reading apparatus according to claim 1, wherein when the interpolation processing unit determines that there is no edge, the interpolation processing unit uses the weighting coefficient set with reference to the pixel interpolation position for each series. An image reading apparatus that performs an interpolation operation.   The image reading apparatus according to claim 1, further comprising an interpolation position control unit that individually controls the first interpolation position and the second interpolation position in accordance with a detection result by the edge detection unit. An image reading apparatus comprising:   4. The image reading apparatus according to claim 3, wherein when the interpolation position control unit determines that the edge is present, the interpolation position control unit determines from the read pixel signals of the other series among the gradation values of the plurality of colors. 5. The first interpolation position and the second interpolation position are controlled such that the calculated color gradation value approaches the color gradation value calculated from the reference series of read pixel signals. Image reading device. The image reading apparatus according to claim 3, wherein
An edge position detector that detects the position of the edge when it is determined that the edge is present;
The image reading apparatus, wherein the interpolation position control unit individually controls the first interpolation position and the second interpolation position according to a detection position of the edge. The image reading apparatus according to claim 5,
The edge detection unit determines the presence / absence of a color edge corresponding to each of the plurality of colors and having a gradation change in the sub-scanning direction from the plurality of read pixel signals, and a determination result of the presence / absence of the color edge According to the combination of, the presence or absence of an edge in the sub-scanning direction is determined,
The image reading apparatus, wherein the interpolation position control unit individually controls the first interpolation position and the second interpolation position based on a determination result of the presence or absence of an edge by the edge detection unit .   The image reading apparatus according to claim 6, wherein the edge position detection unit includes a lookup table in which correspondence between the presence / absence of the color edge and the positions of a plurality of edges is determined in advance. An image reading apparatus that generates information indicating the position of the edge according to a table. The image reading apparatus according to any one of claims 3 to 7,
The edge detection unit detects a magnitude of a gradation change amount of the edge;
The image reading apparatus, wherein the interpolation position control unit individually controls the first interpolation position and the second interpolation position in accordance with the magnitude of the gradation change amount.   9. The image reading apparatus according to claim 1, wherein the reference so as to reduce a phase shift between the read pixel signal of the reference series and the read pixel signal of the other series. An image reading apparatus, further comprising: a delay memory that delays the read pixel signals of the other series relative to the read pixel signals of the series.   10. The image reading apparatus according to claim 9, wherein the delay memory relatively delays the read pixel signals of the other series by a delay time corresponding to a variable reading magnification with respect to the read medium. An image reading apparatus. For a read medium extending in a main scanning direction and a sub-scanning direction different from the main scanning direction, photoelectric conversion is performed for light of a plurality of different colors arranged at predetermined intervals along the sub-scanning direction. An image sensor including a plurality of lines of line sensors that output a plurality of series of read pixel signals in parallel, and a surface of the read medium by moving the image sensor relative to the read medium in the sub-scanning direction An image processing apparatus incorporated in an image reading apparatus including a scanning drive unit that causes the image sensor to receive light reflected by the image sensor,
Among the plurality of series of read pixel signals output in parallel from the plurality of line sensor at regular time intervals within a scanning period in which the image sensor moves relative to the read medium in the sub-scanning direction. An interpolation position determination unit for determining a pixel interpolation position for correcting a phase shift between the read pixel signal of the reference series and the read pixel signal of other series other than the reference series;
An interpolation processing unit that performs an interpolation operation for each of the plurality of read pixel signals and calculates a gradation value of each of the plurality of colors at the pixel interpolation position;
An edge detection unit that determines the presence or absence of an edge having a gradation change in the sub-scanning direction based on at least one series of read pixel signals among the plurality of series of read pixel signals;
When it is determined that the edge is present, the interpolation processing unit, with respect to the read pixel signal of the reference series, the first one of the pixel interpolation position and the position shifted from the pixel interpolation position. The interpolation calculation is performed using a weighting coefficient set with reference to the interpolation position of the second interpolation position, and a second interpolation position different from the first interpolation position is used as a reference for the read pixel signals of the other series. An image processing apparatus that performs the interpolation operation using a weighting coefficient set as.   The image processing apparatus according to claim 11, further comprising an interpolation position control unit that individually controls the first interpolation position and the second interpolation position in accordance with a detection result by the edge detection unit. An image processing apparatus. The image processing apparatus according to claim 12,
An edge position detector that detects the position of the edge when it is determined that the edge is present;
The said interpolation position control part controls the said 1st interpolation position and the said 2nd interpolation position separately according to the detection position of the said edge, The image processing apparatus characterized by the above-mentioned. The image processing apparatus according to claim 12 or 13,
The edge detection unit detects a gradation change amount of the edge,
The image processing apparatus, wherein the interpolation position control unit individually controls the first interpolation position and the second interpolation position according to the magnitude of the gradation change amount. For a read medium extending in a main scanning direction and a sub-scanning direction different from the main scanning direction, light of a plurality of different colors arranged in the sub-scanning direction at different intervals is photoelectrically converted. An image sensor including a plurality of lines of line sensors that output a plurality of series of read pixel signals in parallel, and the image sensor is moved relative to the read medium in the sub-scanning direction and reflected by the surface of the read medium. An image processing method in an image reading apparatus comprising a scanning drive unit that causes the image sensor to receive the light that has been obtained,
Among the plurality of series of read pixel signals output in parallel from the plurality of line sensor at regular time intervals within a scanning period in which the image sensor moves relative to the read medium in the sub-scanning direction. Determining a pixel interpolation position for correcting a phase shift between a read pixel signal of a reference series and a read pixel signal of another series other than the reference series;
Determining the presence or absence of an edge having a gradation change in the sub-scanning direction based on at least one series of read pixel signals among the plurality of series of read pixel signals;
When it is determined that the edge is present, the reference series read pixel signal is set with the first interpolation position as one of the pixel interpolation position and the position shifted from the pixel interpolation position as a reference. Calculating a tone value of a color corresponding to the reference series at the pixel interpolation position by performing an interpolation operation using the weighting coefficient,
When it is determined that the edge is present, an interpolation operation using a weighting coefficient set with reference to a second interpolation position different from the first interpolation position is performed on the read pixel signals of the other series And calculating a color gradation value corresponding to the other series at the pixel interpolation position. The image processing method according to claim 15, comprising:
Detecting the position of the edge when it is determined that the edge is present;
And a step of individually controlling the first interpolation position and the second interpolation position in accordance with the detected position of the edge. The image processing method according to claim 15 or 16,
Detecting a gradation change amount of the edge;
And a step of individually controlling the first interpolation position and the second interpolation position in accordance with the magnitude of the gradation change amount. For a read medium extending in a main scanning direction and a sub-scanning direction different from the main scanning direction, light of a plurality of different colors arranged in the sub-scanning direction at different intervals is photoelectrically converted. An image sensor including a plurality of lines of line sensors that output a plurality of series of read pixel signals in parallel, and the image sensor is moved relative to the read medium in the sub-scanning direction and reflected by the surface of the read medium. A scanning drive unit that causes the image sensor to receive the received light, and the line sensors of the plurality of rows at regular time intervals within a scanning period in which the image sensor moves relative to the read medium in the sub-scanning direction. An image reading apparatus comprising: an image processing unit configured to generate image data based on the plurality of series of read pixel signals output in parallel; and performing image processing on the image processing unit A computer program to,
The computer program is read from a computer-readable recording medium by the image processing unit and executed.
The image processing is
An interpolation position determination process for determining a pixel interpolation position for correcting a phase shift between a read pixel signal of a reference series among the read pixel signals of the plurality of series and a read pixel signal of a series other than the reference series; ,
Edge detection processing for determining the presence or absence of an edge having a gradation change in the sub-scanning direction based on at least one series of read pixel signals among the plurality of series of read pixel signals;
When it is determined that the edge is present, the reference series read pixel signal is set with the first interpolation position as one of the pixel interpolation position and the position shifted from the pixel interpolation position as a reference. A first interpolation process for calculating a tone value of a color corresponding to the reference series at the pixel interpolation position by executing an interpolation operation using the weighting coefficient;
When it is determined that the edge is present, an interpolation operation using a weighting coefficient set with reference to a second interpolation position different from the first interpolation position is performed on the read pixel signals of the other series And a second interpolation process for calculating a gradation value of a color corresponding to the other series at the pixel interpolation position. A computer program according to claim 18, comprising:
The image processing is
Edge position detection processing for detecting the position of the edge when it is determined that the edge is present;
A computer program further comprising: a first interpolation position control process for individually controlling the first interpolation position and the second interpolation position in accordance with a detection position of the edge. A computer program according to claim 18 or 19, comprising:
The image processing is
A process of detecting a gradation change amount of the edge;
A computer program comprising: a second interpolation position control process for individually controlling the first interpolation position and the second interpolation position in accordance with the magnitude of the gradation change amount.

Images