Image reader

Abstract

To provide an image reading apparatus for appropriately correcting a streak image.SOLUTION: An image reading apparatus comprises: an ADF102 for conveying a document; a reading unit 129 for reading an image from the document conveyed by the ADF102; a white guide member 114 arranged at a reading position of the reading unit 129; a streak image detection unit for detecting a streak image based on a reading result of the white guide member 114 and the document by the reading unit 129; a content area detection unit for detecting a content area in which a content is printed from the reading result of the document by the reading unit 129; and a streak image correction unit for correcting the streak image in an area excluding the content area from the reading result of the document by the reading unit 129.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an image reading device that reads an image from a document and an image forming apparatus equipped with the image reading device. [Background technology]

[0002] In recent years, deregulation of the Electronic Bookkeeping Law has promoted electronic storage of expense documents. Accordingly, the demand for multifunction machines equipped with image reading devices for electronically storing paper documents is increasing. The image reading device may have a reading mode for reading a specific document such as a form. The image reading device set to the reading mode for the specific document can automatically detect contents (text information) from the read data obtained by reading the document using OCR (Optical Character Recognition). This allows the image reading device to electronically store the document including not only the image but also the text information printed on the document. In this way, the image reading device makes it easy to electronically store accounting documents. In addition, when electronically storing paper documents, it is necessary to consider not only automatic recognition by OCR but also the user's ability to recognize small point characters from the scanned image.

[0003] When using an automatic document feeder (hereinafter referred to as "ADF"), the readability of characters may be reduced due to the following phenomenon. For example, when an original is transported by the ADF, dirt, paper powder, and other adhesions may adhere to the reading glass provided at the reading position of the original. This causes vertical streaks (hereinafter referred to as "streaky images") to appear in the read image. Patent Document 1 discloses an image processing device equipped with a streaky image correction function that erases streaky images. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2001-285595 A Summary of the Invention [Problem to be solved by the invention]

[0005] In Patent Document 1, when a streak image and text overlap in a content area, the streak image is erased, and part of the content is also deleted. This reduces the accuracy of character detection by OCR and reduces the readability of small-point characters.

[0006] In the specific document reading mode, if the content is unclear, there is a possibility that the characters that were originally printed on the document cannot be read from the read image, and therefore the original purpose of electronic storage, which may not be achieved, may not be achieved. Content that is often included in accounting documents is highly important characters such as billing amounts, company names, and issue dates. Therefore, the loss of content or reduced readability due to streaky images can cause serious problems. On the other hand, if the streaky images are not erased, the streaky images will remain. In other words, the streaky images will remain outside the content area (for example, in blank areas), degrading the appearance of the read image.

[0007] SUMMARY OF THE PRESENT DISCLOSURE In view of the above problems, a primary object of the present invention is to provide an image reading device that appropriately corrects streaked images caused by a substance adhering to the reading position. [Means for solving the problem]

[0008] The image reading device of the present invention includes a conveying means for conveying an original, an image reading means for reading an image of the original conveyed by the conveying means using a reading sensor and generating image data representing the reading result, and a control means for detecting a streak image in an image represented by the image data based on the image data and performing a correction process to remove the streak image by correcting the image data, wherein the control means determines a content area in the image represented by the image data in which content is printed, performs the correction process on image data corresponding to the streak image in an area excluding the content area, and has a first reading mode in which the correction process is not performed on image data corresponding to the streak image in the content area. The first reading mode is a reading mode for reading a specific document. It is characterized by the above. Another image reading device of the present invention includes a conveying means for conveying an original document, an image reading means for reading an image of the original document conveyed by the conveying means using a reading sensor and generating image data representing the reading result, a control means for detecting a streak image in an image represented by the image data based on the image data and performing a correction process to remove the streak image by correcting the image data, and a white guide member arranged at a reading position by the image reading means on the opposite side of the image reading means with respect to a conveying path through which the original document passes, wherein the control means determines a content area in which content is printed in the image represented by the image data, performs the correction process on image data corresponding to the streak image in an area excluding the content area, and has a first reading mode in which the correction process is not performed on image data corresponding to the streak image in the content area, and the control means detects the streak image based on image data generated when the image reading means reads the white guide member and image data generated when the image reading means reads the original document. Another image reading device of the present invention includes a conveying means for conveying an original document, an image reading means for reading an image of the original document conveyed by the conveying means using a reading sensor and generating image data representing the reading result, and a control means for detecting a streak image in an image represented by the image data based on the image data and performing a correction process to remove the streak image by correcting the image data, wherein the control means has a first reading mode for determining a content area in which content is printed in the image represented by the image data, performing the correction process on image data corresponding to the streak image in an area excluding the content area, and not performing the correction process on image data corresponding to the streak image in the content area, and wherein the content area is an area specified by a user. Effect of the Invention

[0009] According to the present invention, it is possible to appropriately correct streak images, and to prevent loss of content and deterioration of readability. [Brief description of the drawings]

[0010] [Figure 1] FIG. 2 is a diagram illustrating the configuration of an image reading device. [Diagram 2] FIG. [Diagram 3] FIG. [Figure 4] FIG. 2 is a diagram illustrating the configuration of a read image processing unit. [Diagram 5] 4A and 4B are diagrams illustrating the relationship between an adhesion and a line sensor. [Figure 6] 6 is a diagram illustrating an example of digital values ​​at each position in a main operation direction. [Figure 7] 6A to 6C are explanatory diagrams illustrating a correction process for a streak image. [Figure 8] 4A to 4C are diagrams illustrating edge enhancement processing. [Figure 9] 5A to 5C are diagrams illustrating the processes performed by an isolation amount calculation unit and a halftone dot determination unit. [Figure 10] 11(a) to 11(c) are diagrams illustrating block selection processing. [Figure 11] 4A to 4C are diagrams illustrating streak image correction. [Figure 12] 13A to 13C are diagrams illustrating streak image correction for accounting documents. [Figure 13] 13A to 13C are diagrams illustrating examples of setting screens. [Figure 14] 11 is a flowchart showing an image reading process including streak image correction. [Figure 15] FIG. 2 is a diagram illustrating the configuration of a read image processing unit. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0012] (First embodiment) 1 is a diagram illustrating the configuration of an image reading device according to this embodiment. The image reading device 100 includes an image reading unit 101 and an ADF 102. The image reading device 100 may form part of an image forming device that forms an image on a sheet, such as a copier or multifunction device. The ADF 102 is disposed above the image reading unit 101. A reading glass 118 is provided on the surface of the image reading unit 101 facing the ADF 102. The ADF 102 is provided so as to be openable and closable relative to the reading glass 118.

[0013] The ADF 102 includes an original tray 104 on which an original stack 103 consisting of one or more originals P is placed, a transport path 130, and a paper discharge tray 117. The originals P are transported one by one from the original tray 104 to the transport path 130 and discharged to the paper discharge tray 117. The originals P are read by the image reading unit 101 while being transported along the transport path 130. The original tray 104 includes a pair of width regulating plates 105 at both ends in a direction (width direction) perpendicular to the transport direction of the originals P. The pair of width regulating plates 105 align the ends of the original stack 103 in the width direction. A pickup roller 106 for feeding the originals P from the original stack 103 to the transport path 130 is disposed at the base end of the original tray 104.

[0014] A separation roller 108 and a separation pad 107 are provided on the transport path 130 downstream of the pickup roller 106 in the transport direction of the original P. The separation roller 108 and the separation pad 107 are disposed opposite each other with the transport path 130 in between. The separation roller 108 cooperates with the separation pad 107 to separate the originals P fed from the original tray 104 by the pickup roller 106 into the transport path 130, one by one from the top of the original stack 103.

[0015] In the conveying path 130 downstream of the separation roller 108 in the conveying direction of the document P, first and second registration rollers 109 and 110, first to fourth conveying rollers 111, 112, 113, and 115, and a paper discharge roller 116 are arranged in this order from the upstream side. Between the first conveying roller 111 and the second conveying roller 112, a document detection sensor 127 is arranged. Details of the document detection sensor 127 will be described later. The space between the second conveying roller 112 and the third conveying roller 113 is located on the upper surface of the reading glass 118 of the image reading unit 101, and is the reading position by the image reading unit 101. At the reading position, a white guide member 114 is arranged at a position facing the reading glass 118 across the conveying path 130.

[0016] The reading glass 118 of the image reading unit 101 is provided with a shading reference plate 128 at a position adjacent to the reading position via the third conveyor roller 113. A reading unit 129 is provided inside the housing of the image reading unit 101. The reading unit 129 includes lamps 119 and 120 as light sources, reflecting mirrors 121, 122, and 123, an imaging lens 124, and a line sensor 125. The reflecting mirrors 122 and 123 are disposed to face the reflecting mirror 121. The imaging lens 124 and the line sensor 125 are disposed to face the reflecting mirrors 122 and 123. A signal processing board 126 is electrically connected to the line sensor 125.

[0017] In the image reading device 100 configured as above, the document stack 103 placed on the document tray 104 of the ADF 102 is restricted in the width direction by the width restriction plate 105, thereby preventing the document P from being conveyed obliquely. The pickup roller 106 picks up the document P from the document stack 103 and feeds it to the conveying path 130. The separation pad 107 and the separation roller 108 work together to separate the document P one by one from the top of the document stack 103 and convey it.

[0018] The first registration roller 109 corrects the skew of the document P that is separated into one sheet and transported by the separation roller 108. The document P whose skew has been corrected is transported to the reading position by the second registration roller 110, the first transport roller 111, and the second transport roller 112. At the reading position, the document P is transported between the reading glass 118 and the white guide member 114. The white guide member 114 presses the document P toward the reading glass 118 so that the document P passing through the reading position does not separate from the reading glass 118 and float. In addition, the white guide member 114 is read by the reading unit 129 when the document P is not present at the reading position. Based on the reading result of reading the white guide member 114, it is determined whether or not there is any dirt or other attachment at the reading position.

[0019] The reading unit 129 performs a reading operation on the document P passing through the reading position as follows. The start timing of the reading operation of the document P by the reading unit 129 is controlled according to the detection timing of the document P on the transport path 130 by the document detection sensor 127.

[0020] Lamps 119 and 120 of the reading unit 129 irradiate light onto the original P passing through the reading position. Reflecting mirrors 121-123 reflect the light reflected by the original P and guide it to an imaging lens 124. The imaging lens 124 forms an image of the reflected light on the light receiving surface of a line sensor 125. The line sensor 125 has multiple imaging elements such as CCD sensors and CMOS sensors, and the reflected light forms an image on the light receiving surface of the imaging elements. The line sensor 125 converts the received reflected light into an electrical signal and transmits it to a signal processing board 126. The direction in which the multiple imaging elements are arranged is the main scanning direction, which is perpendicular to the transport direction of the original P. The transport direction of the original P is the sub-scanning direction.

[0021] The signal processing board 126 performs a predetermined process on the electrical signal obtained from the line sensor 125 to generate image data that is the result of reading the original P. The image data is a digital signal. When the image reading device 100 constitutes part of an image forming device, the image data is transmitted from the image reading device to an image forming section in the image forming device. In the case of copying processing, the image forming section forms an image on a sheet based on the image data. When recognizing the content (character information) of the original P, character recognition processing is performed by OCR based on the image data. When the original P is an accounting document, the result of character recognition by OCR and the image data are stored electronically.

[0022] The original P read at the reading position is transported to the third transport roller 113 by the second transport roller 112. The original P is transported to the third transport roller 113, the fourth transport roller 115, and the paper discharge roller 116 in this order, and is discharged to the paper discharge tray 117.

[0023] A shading reference plate 128 provided on the reading glass 118 is read by a reading unit 129 during shading correction. The line sensor 125 has manufacturing variations for each imaging element (each pixel). In addition, it is not easy to make the light irradiated from the lamps 119 and 120 uniform in the main scanning direction. For this reason, even if an image is read from an original P on which an image of uniform image density is formed, the digital values ​​of the image data, which are the reading results, may vary for each position in the main scanning direction.

[0024] In order to suppress such variations, shading correction is performed. Specifically, the reading unit 129 reads the shading reference plate 128. From the reading result of the shading reference plate 128, a correction value is calculated so that the reading result in the main scanning direction (e.g., luminance value) becomes uniform to a predetermined value. This correction value corrects the amount of irradiation of the lamps 119 and 120, the sensitivity variation of the imaging element, or the reading result of the image of the document P, thereby correcting the variation in image data for each position in the main scanning direction.

[0025] In order to read the shading reference plate 128, the lamps 119, 120 and the reflecting mirror 121 of the reading unit 129 can move in the left-right direction (sub-scanning direction) in the figure. When reading the shading reference plate 128, the lamps 119, 120 and the reflecting mirror 121 move from the reading position of the document to directly below the shading reference plate 128. When reading the image of the document P after shading correction, the lamps 119, 120 and the reflecting mirror 121 move to the reading position. Note that the reading unit 129 can also read the image of a document manually placed on the reading glass 118 by the user. In this case, the lamps 119, 120 and the reflecting mirror 121 of the reading unit 129 read the document line by line while moving in the sub-scanning direction.

[0026] 2 is a diagram illustrating an example of the document detection sensor 127. The document detection sensor 127 includes an actuator 127a and a transmission type sensor 301. The transmission type sensor 301 includes an irradiation unit 301a and a light receiving unit 301b.

[0027] When the original P conveyed on the conveying path 130 comes into contact with the actuator 127a, the actuator 127a tilts in the conveying direction of the original P. When the actuator 127a tilts in the conveying direction of the original P, the optical path between the irradiating unit 301a and the light receiving unit 301b of the transmission sensor 301 is blocked. When the optical path between the irradiating unit 301a and the light receiving unit 301b is blocked, the amount of light received by the light receiving unit 301b from the irradiating unit 301a, for example, the amount of infrared light, changes. The amount of light received by the light receiving unit 301b is converted into an electrical signal. The change in the amount of light received by the light receiving unit 301b results in a change in the level of the electrical signal. The change in the level of the electrical signal indicates that the original P has reached the detection position of the original detection sensor 127.

[0028] (controller) 3 is an explanatory diagram of a controller that controls the operation of the image reading device 100. The controller is built into the image reading device 100. In this embodiment, the controller is configured with a CPU (Central Processing Unit) 202, but this may also be a semiconductor device such as an MPU or ASIC. The CPU 202 controls the operation of the entire image reading device 100 by executing a predetermined computer program.

[0029] The CPU 202 is connected to the operation panel 201, the document detection sensor 127, the reading unit drive motor 204, the document transport motor 205, the line sensor 125, the AD converter 206, the read image processing unit 207, and the storage unit 208. The AD converter 206, the read image processing unit 207, and the storage unit 208 are mounted on the signal processing board 126.

[0030] As described above, the document detection sensor 127 is attached a predetermined distance upstream in the transport direction of the document P from the reading position of the document P. The CPU 202 determines that the leading edge of the document P has reached the detection position of the document detection sensor 127 based on the change in the level of the electrical signal output from the document detection sensor 127 as described above. After the leading edge of the document P has reached the detection position of the document detection sensor 127, the CPU 202 manages the timing at which the document P reaches the reading position by the number of pulses for driving the document transport motor 205. At the timing at which the document P reaches the reading position, the CPU 202 controls the line sensor 125 so that image reading is enabled. This makes it possible to read the image of the document P.

[0031] Similarly, CPU 202 determines that the rear end of document P has reached the detection position of document detection sensor 127 based on a change in the level of the electrical signal output from document detection sensor 127. After the rear end of document P has reached the detection position of document detection sensor 127, CPU 202 manages the timing at which document P passes the reading position by the number of pulses for driving document conveying motor 205. At the timing at which document P has passed the reading position, CPU 202 controls line sensor 125 to end reading of the image. This ends reading of the image of document P.

[0032] The operation panel 201 is a user interface equipped with an input interface and an output interface. The input interface includes key buttons, a touch panel, etc. The output interface includes a display, a speaker, etc. The CPU 202 controls the operation of the image reading device 100 in response to instructions, etc. input from the input interface of the operation panel 201, and outputs information such as the status of the image reading device 100 from the output interface of the operation panel 201. When setting reading conditions, etc., the CPU 202 also displays a setting screen on the display and accepts operation contents such as settings via the input interface.

[0033] The reading unit drive motor 204 is a drive source for moving the reading unit 129 in the sub-scanning direction under the control of the CPU 202. When reading an image from a document placed on the reading glass 118 or when reading the shading reference plate 128, the CPU 202 causes the reading unit drive motor 204 to move the reading unit 129 in the sub-scanning direction. The document transport motor 205 is a drive source for rotating various rollers arranged along the transport path 130. When the CPU 202 reads an image of the document P using the ADF 102, the CPU 202 controls the document transport motor 205 to feed the document P from the document tray 104. The document transport motor 205 is, for example, a pulse motor, and controls the transport of the document P in accordance with the number of pulses input from the CPU 202.

[0034] The CPU 202 controls the operation of the line sensor 125. The line sensor 125 converts the reflected light it receives into an electrical signal and outputs it. The line sensor 125 includes a plurality of line sensors in order to receive reflected light of a plurality of colors and read an image in a plurality of colors. In this embodiment, the line sensor 125 includes three line sensors corresponding to the three colors of R (red), G (green), and B (blue) in order to receive reflected light of the three colors. The electrical signal output from the line sensor 125 is an analog voltage of each color corresponding to R, G, and B.

[0035] The AD converter 206 acquires an analog voltage for each color from the line sensor 125. The AD converter 206 converts the acquired analog voltage for each color into read data, which is a digital value. The read data converted from the analog voltage for each color by the AD converter 206 is transmitted to the read image processing unit 207. The read image processing unit 207 performs a predetermined process on the read data for each color to generate image data representing the read image. The read image processing unit 207 stores the image data generated by reading the document P in the storage unit 208. The storage unit 208 can also store the result of character recognition by OCR. The image data generated by reading the document P may be directly transmitted to an image forming apparatus or another external device.

[0036] The image reading device 100 having such a configuration operates as follows, for example, when copying an original P. When copying, a user places an original P on the original tray 104 and inputs an instruction for copying processing via the operation panel 201.

[0037] When the CPU 202 receives a copy processing instruction from the operation panel 201, the reading unit drive motor 204 moves the lamps 119, 120 and the reflecting mirror 121 to a reading position of the shading reference plate 128. The CPU 202 causes the reading unit 129 to read the shading reference plate 128 and performs shading correction.

[0038] In shading correction, analog voltages of the R, G, and B colors, which are output from the line sensor 125 and indicate the results of reading the shading reference plate 128, are input to an AD converter 206. The AD converter 206 converts the analog voltages of the colors into read data and transmits them to a read image processing unit 207. The read image processing unit 207 performs shading correction so that the read data (digital values), which are the results of reading the shading reference plate 128, become approximately the same value.

[0039] When the shading correction is completed, the CPU 202 moves the lamps 119 and 120 and the reflecting mirror 121 to the reading position of the document P by the reading unit drive motor 204. The CPU 202 starts conveying the document P on the document tray 104 by the document conveying motor 205. The reading unit 129 reads the white guide member 114 until the document P is conveyed to the reading position. When the document detection sensor 127 detects the document P, the CPU 202 reads the document P by the reading unit 129 as described above. Based on the reading result of the white guide member 114 and the reading result of the document P, the position of the streak image caused by the attachment attached to the reading glass 118 is specified. Details of the streak image detection process for specifying the position of the streak image will be described later. The image data obtained by reading the document P is corrected for streak images as necessary and transmitted to the image forming device. The image forming device forms an image on a sheet based on the acquired image data. In this manner, the copying process is performed.

[0040] (Image reading processing unit) 4 is a diagram illustrating the configuration of the read image processing unit 207. The read image processing unit 207 includes a streak image detection unit 400, a content area detection unit 410, a streak image correction unit 420, and an image processing control unit 430. The read image processing unit 207 detects streak images and content areas (character areas) based on the read data acquired from the AD converter 206, and erases streak images while preventing loss of content. This allows the image reading device 100 to maintain the readability of small-point characters while maintaining the accuracy of character detection by OCR.

[0041] (Streak image detection) The streak image detection unit 400 includes an average brightness calculation unit 401, a brightness difference calculation unit 402, and a streak image determination unit 403. The streak image detection unit 400 detects a streak image while reading a predetermined number of lines of the document P after the document detection sensor 127 detects the document P. If the document P contains printed information such as thin lines, it is difficult to distinguish between a streak image caused by an attachment and the printed thin lines. Therefore, the streak image detection unit 400 identifies candidates for abnormal pixels having a streak image in the main scanning direction (streak pixel candidates) from read data obtained by reading the white guide member 114 at a timing when the document P is not present at the reading position. The streak image detection unit 400 also identifies candidates for abnormal pixels having a streak image in the main scanning direction (streak pixel candidates) from read data obtained by reading the document P at a timing when the document P is present at the reading position. The streak image detection unit 400 compares the positions in the main scanning direction of the streak pixel candidates obtained from the reading results of the white guide member 114 and the streak pixel candidates obtained from the reading results of the document P. If there are streak image candidates at the same position (pixel) in the main scanning direction, the pixel becomes the streak pixel.

[0042] If a substance adheres to white guide member 114, document P passes closer to reading unit 129 than white guide member 114, and therefore no streak image resulting from white guide member 114 will appear in the reading results of document P. However, if a substance adheres to reading glass 118, a streak image resulting from the substance will appear in the reading results of document P, because reading glass 118 is closer to reading unit 129 than the document. For this reason, streak image detection unit 400 identifies streak pixel candidates from the respective read data of white guide member 114 and document P, and if there are streak pixel candidates at the same position (pixel) in the main scanning direction, it detects the pixel as a streak pixel.

[0043] The streak image detection unit 400 detects streak pixel candidates for each pixel in the main scanning direction based on the luminance value of each pixel included in the read data acquired from the AD converter 206. The streak image detection unit 400 judges pixels in which the difference between the average luminance value of the pixel of interest and the luminance value of the pixel of interest is equal to or greater than a predetermined threshold value to be in the range of streak image candidates. The streak image detection unit 400 notifies the streak image correction unit 420 of the detected range of streak image candidates. Each unit of the streak image detection unit 400 will be described.

[0044] The average luminance calculation unit 401 calculates the average luminance value within a predetermined area (for example, an area of ​​7 pixels in the X direction and 7 pixels in the Y direction) based on the luminance value included in the read data. When the coordinate of the pixel of interest in the main scanning direction is X and the coordinate in the sub-scanning direction is Y, and when referring to three pixels before and after the pixel of interest in both the main scanning direction and the sub-scanning direction, the average luminance value AVE1 is calculated by the following formula 1.

[0045]

number

[0046] The average luminance value AVE1 is sent to a luminance difference calculation unit 402. The luminance difference calculation unit 402 calculates a luminance difference DELTA between the average luminance value AVE1 and the luminance value of the pixel of interest. The luminance difference DELTA is calculated by the following Equation 2.

[0047]

number

[0048] The streak image determination unit 403 determines that the pixel of interest has a substance attached thereto when the luminance difference DELTA is greater than a predetermined threshold value THdetect. The streak image determination unit 403 performs a substance determination for the pixel of interest for each line in the main scanning direction. The streak image determination unit 403 counts the number of pixels of interest determined to have a substance attached thereto continuously in the sub-scanning direction. The count result is stored in DUST[X] and is calculated using the following formula 3.

[0049]

number

[0050] The streak image determination unit 403 performs such calculations for each line at each position in the main scanning direction. If the count value DUST[X] of a position determined to have a substance after calculations for a predetermined number of lines is equal to or greater than a predetermined value, the streak image determination unit 403 determines that the position in the main scanning direction is a streak pixel candidate. The streak pixel candidate determination result DUST is calculated using the following mathematical formula 4.

[0051]

number

[0052] The streak image detection unit 400 determines that a streak pixel candidate exists at a position (pixel) in the main scanning direction where it is determined that there are pixels with attached matter attached for a predetermined number of consecutive lines or more, based on the respective read data of the white guide member 114 and the document P, as described above. The streak image detection unit 400 detects a pixel as a streak pixel when the streak pixel candidate exists at the same position (pixel) in the main scanning direction for the white guide member 114 and the document P. In this way, it is determined that a streak image has occurred at a position in the main scanning direction where it is determined that there are streak pixels attached for a predetermined number of consecutive lines or more. The reason for detecting the position of the attached matter for a predetermined number of consecutive lines is that the attached matter is particles such as paper powder, and may move during reading, and there is a possibility of erroneous detection when trying to identify the position of the attached part in one line. The pixel corresponding to the position of the streak image in the main scanning direction is called an "abnormal pixel."

[0053] FIG. 5 is an explanatory diagram of the relationship between dust (adherent matter) attached to the reading position and line sensor 125. Line sensor 125 is a line sensor in which a plurality of image pickup elements are arranged in the main scanning direction. Light reflected by white guide member 114 or document P is imaged on the light receiving surface of the line sensor by imaging lens 124 via reflecting mirrors 121, 122, and 123. In FIG. 5, dust 501 is attached to the reading position. In FIG. 5(a), dust 501 is attached to the reading position of white guide member 114. In FIG. 5(b), dust 501 is attached to the reading position of reading glass 118. In FIG. 5(a), the reflected light received by line sensor 125 from white guide member 114 is partially blocked in the main scanning direction by dust 501 attached to white guide member 114. In FIG. 5B, the reflected light received by line sensor 125 from white guide member 114 is partially blocked in the main scanning direction by dust 501 adhering to reading glass 118.

[0054] A method for detecting the position of an abnormal pixel will be described below. First, the output of the line sensor 125 when the line sensor 125 reads the white guide member 114 will be described.

[0055] 6 is an example diagram of read data (digital values) at each position in the main scanning direction based on the reading result output from line sensor 125 when dust 501 is attached to reading position 502. In Fig. 6, the horizontal axis indicates each position in the main scanning direction, and the vertical axis indicates the digital value of the read data. When AD converter 206 outputs an 8-bit digital value, the digital value of the read data when the analog voltage output from line sensor 125 is AD converted will be a value in the range of 0 to 255.

[0056] The digital values ​​obtained when the white guide member 114 is read after shading correction are approximately uniform. However, as shown in FIG. 5(a), for example, if dust 501 is attached to the white guide member 114 at the reading position 502, the dust 501 blocks the light irradiated onto the white guide member 114 at the position where the dust 501 is present. As a result, the digital value at this position is smaller than that at other positions. The streak image detection unit 400 compares the digital values ​​for each position in the main scanning direction when the white guide member 114 is read with a predetermined threshold value. The streak image detection unit 400 determines, as a result of the comparison, pixels corresponding to positions where the digital value is smaller than the predetermined threshold value as candidates for abnormal pixels (candidate abnormal pixels).

[0057] As shown in Fig. 5(b), if dust 501 is attached to the reading glass 118 at the reading position 502, the dust 501 blocks the light irradiated to the white guide member 114 at the position where the dust 501 is present. Therefore, the digital value at this position will be smaller than other positions. The streak image detection unit 400 compares the digital value for each position in the main scanning direction when the white guide member 114 is read with a predetermined threshold value. As a result of the comparison, the streak image detection unit 400 determines pixels corresponding to positions where the digital value is smaller than the predetermined threshold value as candidates for abnormal pixels (candidate abnormal pixels).

[0058] In the example of Fig. 6, the average luminance value AVE1 when the white guide member 114 is read is "about 200", and the luminance difference DELTA is "20". Therefore, if the threshold value is, for example, "180", it can be determined that dust 501 has adhered to a reading position where the digital value is lower than "180", causing a streaked image. In the example of Fig. 6, the pixels of dust 501 at positions 1015 to 1020 in the main scanning direction are determined to be candidate abnormal pixels.

[0059] Next, the output of line sensor 125 when line sensor 125 reads the image of document P passing reading position 502 will be described.

[0060] 5(a), when dust 501 is attached on white guide member 114 at reading position 502, dust 501 is hidden by document P when viewed from line sensor 125 as document P passes over reading glass 118. As a result, no streaky image caused by dust 501 appears in the image representing the results of reading document P. That is, the values ​​at positions 1015 to 1020 in the main scanning direction in FIG. 6 do not become smaller than a predetermined threshold value caused by dust 501.

[0061] As shown in FIG. 5(b), if dust 501 is attached on the reading glass 118 at the reading position 502, when the original P passes over the reading glass 118, the original P will be hidden by the dust 501 when viewed from the line sensor 125. As a result, a streak image due to the dust 501 appears in the image representing the reading result of the original P. That is, the values ​​at positions 1015 to 1020 in the main scanning direction in FIG. 6 become smaller than a predetermined threshold value due to the dust 501. The streak image detection unit 400 determines pixels corresponding to positions where the digital value is smaller than the predetermined threshold value as candidates for abnormal pixels (candidate abnormal pixels).

[0062] The streak image detection unit 400 determines an abnormal pixel based on image data obtained during the period from when reading of the white guide member 114 begins until a predetermined length of image is read from the leading edge of the document P. If candidate abnormal pixels are consecutive in the sub-scanning direction at the same position in the main scanning direction in the image data for that period, the streak image detection unit 400 determines the pixel at the position of the candidate abnormal pixel in the main scanning direction as an abnormal pixel. In other words, if a streak image appears consecutively before and after the document P reaches the reading position 502, the streak image detection unit 400 determines the pixel at the position of the streak image in the main scanning direction as an abnormal pixel.

[0063] (Streak image correction) 7 is an explanatory diagram of the streak image correction process. The streak image determination unit 403 outputs the streak image determination result as streak image information (DETECT). In the streak image information, "1" is set to pixels in the main scanning direction that are determined to have streak images, and "0" is set to pixels that are determined to have no streak images. The streak image information makes it possible to know which pixel in the main scanning direction has a streak image. Streak image correction is performed based on the streak image information.

[0064] In the streak image information in Figure 7, the pixel of interest is "1", indicating that a streak image has occurred at the pixel of interest. The streak image information for pixels N-3, N-2, N-1, N+1, N+2, and N+3 surrounding the pixel of interest in the main scanning direction is "0", indicating that no streak image has occurred. The digital value of the read data for the pixel of interest is "80", which is a smaller value than the digital values ​​at the surrounding positions.

[0065] First, the streak image correcting unit 420 identifies the pixel (position) of the streak image based on the streak image information obtained from the streak image determining unit 403. Next, the streak image correcting unit 420 refers to the digital values ​​of the pixels (positions) adjacent in the main scanning direction to the identified pixel (position) of the streak image, and linearly interpolates the digital values ​​of the pixels of the streak image based on this. The streak image is removed by linear interpolation.

[0066] 7, the digital values ​​of pixels N-1 and N+1 adjacent to the pixel of interest where the streak image occurs are "190" and "192," respectively. If the digital value of the pixel of interest is D[N], the digital values ​​of the adjacent pixels are D[N-1] and D[N+1], and the digital value of the pixel of interest after correction (after linear interpolation) is D[N]', then the digital value D[N]' is expressed by the following formula. D[N]'=D[N-1]+(D[N+1]-D[N-1]) / 2 D[N]'=190+(192-190) / 2 =191

[0067] Therefore, the digital value D[N]' of the pixel of interest after linear interpolation becomes "191." In this way, the streak image corrector 420 performs streak image correction by converting the digital value "80" of the pixel of interest before correction into "191."

[0068] (Content Area Detection) The content area detection unit 410 detects a content area (character area) based on the luminance value of each pixel included in the read data acquired from the AD converter 206. The content area detection unit 410 calculates the difference between the average luminance value in the vicinity of a pixel of interest and the luminance value of the edge-emphasized read data, and binarizes the data by setting "1" to pixels whose difference is equal to or less than a predetermined threshold and "0" to pixels whose difference exceeds the threshold. The content area detection unit 410 detects isolated pixel groups by detecting pixels whose binarization results (binarized data) match a predetermined pattern by pattern matching.

[0069] The content area detection unit 410 converts the detected isolated pixel groups from "1" to "0" to exclude dot areas from the binarized data. The content area detection unit 410 determines that areas from which dot areas have been excluded are pixels containing content. The content area detection unit 410 determines whether or not each predetermined area contains pixels containing content. By determining that areas containing pixels containing content are "1" and areas that do not contain content are "0", areas near the image containing content are detected.

[0070] As described above, the streak image correcting unit 420 performs streak image correction near the position in the main scanning direction where the streak image is located, based on the streak image information obtained from the streak image detecting unit 400. However, the streak image correcting unit 420 does not perform streak image correction in an area determined by the content area detecting unit 410 to contain content. This causes the content to be processed without streak image correction. The read data in which the streak image has been corrected is transmitted to the image processing control unit 430 without losing information near the content. This prevents the loss of content due to streak image correction.

[0071] The following describes each unit of the content region detection unit 410. The content region detection unit 410 includes an average luminance calculation unit 411, an edge enhancement calculation unit 412, a binarization determination unit 413, an isolation amount calculation unit 414, a halftone dot determination unit 415, a content determination unit 416, and a block selection 417.

[0072] Similar to the average brightness calculation unit 401 of the streak image detection unit 400, the average brightness calculation unit 411 calculates the average brightness value within a predetermined area (for example, an area of ​​7 pixels in the X direction and 7 pixels in the Y direction) based on the brightness value included in the read data. When the coordinate of the pixel of interest in the main scanning direction is X and the coordinate in the sub-scanning direction is Y, and when referring to three pixels before and after the pixel of interest in both the main scanning direction and the sub-scanning direction, the average brightness value AVE2 is calculated by the following formula 5.

[0073]

number

[0074] The edge emphasis calculation unit 412 performs edge emphasis processing on the read data to detect content areas and dot areas from the read data obtained from the AD converter 206. When a pixel of interest is multiplied by a 7x7 filter coefficient K for each of three pixels before and after the pixel of interest in both the main scanning direction and the sub-scanning direction, the luminance value EDGE of the read data after edge emphasis is calculated by the following formula 6.

[0075]

number

[0076] FIG. 8 is an explanatory diagram of edge enhancement processing. FIG. 8(a) shows an image based on read data in which the background has a brightness value of "200", the diagonal line has a brightness value of "150", and the isolated point has a brightness value of "190". The image size is 10x10 pixels. FIG. 8(b) shows an edge enhancement filter with a size of 5x5 pixels. FIG. 8(c) shows the brightness value of the read data after the edge enhancement filter performs a convolution operation based on the above formula 4. Values ​​below 0 as a result of the convolution operation based on formula 4 are clipped to "0", and values ​​above 255 are clipped to "255".

[0077] In the read data after the convolution operation (Fig. 8(c)), the brightness value of the diagonal line is "255" and the edges are emphasized. The brightness value of the background is "0" due to the edge emphasis. In this way, the edges are extracted. The isolated points, which have a small brightness difference from the background brightness value, are not significantly emphasized after edge emphasis and have a brightness value of "80".

[0078] The binarization determination unit 413 compares the difference between the average brightness value AVE2 output from the average brightness calculation unit 411 and the brightness value EDGE output from the edge emphasis calculation unit 412 with a threshold value THedge. The binarization determination unit 413 outputs binarized data BIN that is "1" when the difference is equal to or smaller than the threshold value THedge, and is "0" when the difference is greater than the threshold value THedge. This binarization process emphasizes and extracts information that is clearly printed on the document, such as characters, and dots in halftone printing. The binarization determination unit 413 performs binarization determination using the following equation 7.

[0079]

number

[0080] In the example of Fig. 8(c), the threshold value THedge is set to "100", pixels with a difference of 100 or more are set to "0", and pixels with a difference of 99 or less are set to "1" for binarization processing. As a result, only pixels with diagonal lines are assigned BIN=1, making it possible to extract diagonal lines.

[0081] The threshold value THedge is set to a value that detects the content but does not detect the streak image. For the streak image and the content printed on the accounting document, the luminance value of the content tends to be lower than the luminance value of the streak image. The adhering matter is mainly paper dust and the like that peeled off when the document P was transported by the ADF 102. Because the particle diameter of the paper dust is small, the luminance value when the image is read tends to be higher than the luminance value of the content. As a result, the luminance value of the streak image becomes higher than the luminance value of the content.

[0082] The isolated amount calculation unit 414 detects isolated pixel groups with no surrounding pixels with BIN=1, such as halftone dots printed on the document P. The isolated amount calculation unit 414 prepares in advance an image for pattern matching that is an image that is assumed when halftone dots are read, and detects areas where an image matching this image is found as isolated pixel groups. The halftone dot determination unit 415 converts the binary data BIN value of the detected pixel groups from "1" to "0." This process makes it possible to remove the halftone dot portion from the binary data.

[0083] Fig. 9 is an explanatory diagram of the processing of the isolation amount calculation unit 414 and the halftone dot determination unit 415. Fig. 9(a) is an example diagram of binarized data BIN. Fig. 9(a) illustrates binarized data BIN including a halftone dot area in which the character "A" and 2x2 dots are printed. Fig. 9(b) illustrates an image for pattern matching to detect 2x2 dots. Fig. 9(c) illustrates binarized data after processing by the isolation amount calculation unit 414 and the halftone dot determination unit 415.

[0084] In the binary data BIN in Fig. 9(a), there are two areas that match the image for pattern matching (areas surrounded by a thick frame in Fig. 9(a)). By converting the binary data BIN in this area from "1" to "0", the binary data in Fig. 9(c) is obtained. In this example, the 2x2 dot portion has been removed, leaving the letter "A," so the halftone dot area has been removed and only the content has been extracted. The binarized data DOT from which the halftone dot area has been removed is transmitted from the halftone dot determination unit 415 to the content determination unit 416. The content determination unit 416 determines that the pixels set to "1" from the binarized data DOT from which the halftone dot area has been removed are the content area, as shown in Fig. 9(c).

[0085] The block selection 417 performs area mapping by dividing the image read by the reading unit 129 into a predetermined size (for example, a rectangular size of 32x32 pixels) to detect a content area. FIG. 10 is an explanatory diagram of the process of the block selection 417.

[0086] For each mapped area, the block selection 417 determines whether the area contains a pixel whose binary data DOT is "1". The block selection 417 determines that an area containing a pixel whose binary data DOT is "1" contains content. In the example of FIG. 10(b), the letter "A" is printed within Area 11, Area 12, Area 21, and Area 22. Since this area contains a pixel whose binary data DOT is "1", it is determined that the area surrounded by the thick frame contains content.

[0087] 10(c), the block selection 417 outputs areas that are determined to have content as "1" and areas that do not have content as "0." Through this process, areas near pixels that have content are extracted.

[0088] In the present embodiment, as an example, if there are pixels in the area of ​​interest whose binary data DOT is "1", the area is determined to contain content. Alternatively, if there are a predetermined number or more pixels in the area of ​​interest whose binary data DOT is "1", the area may be determined to contain content. When the number of pixels in each area whose binary data DOT is "1" is calculated, it is presumed that an area with a small number of pixels does not contain content. For this reason, the average number of pixels in which the binary data DOT is "1" is calculated for a plurality of areas with a small number of pixels whose binary data DOT is "1". If the difference between this average number and the number of pixels in the area of ​​interest whose binary data DOT is "1" is a predetermined number or more, the area of ​​interest may be determined to contain content.

[0089] The output data BLK of the block selection 417 is sent to the streak image correction unit 420. The streak image correction unit 420 performs linear correction to remove streak images as described above. In this embodiment, the streak image correction unit 420 prevents loss of content by not performing streak image correction in areas where the output data BLK obtained from the block selection 417 is "1". The area where the output data BLK is "1" becomes the content area.

[0090] Fig. 11 is an explanatory diagram of streak image correction. Fig. 11(a) illustrates an image 1101 near the character "1" printed on a document, and a streak image 1111. Fig. 11(b) illustrates an example of a case where streak image correction is performed on the entire image of Fig. 11(a). Fig. 11(c) illustrates an example of a case where streak image correction is not performed on the content area detected by block selection 417 for the image of Fig. 11(a).

[0091] In Fig. 11(a), the character "1" is printed, and a streak image 1111 appears where it overlaps with the character. In Fig. 11(b), the printed "1" has almost completely disappeared due to streak image correction being performed on the entire image, reducing the readability of the content. Fig. 11(c) shows an image in which streak image correction is not performed on a content area 1113 detected by block selection 417. Because streak image correction is not performed on the content area 1113, a streak image 1114 remains in the image within the content area 1113, and the streak image outside the content area 1113 has been erased.

[0092] FIG. 12 is an explanatory diagram of streak image correction for accounting documents. FIG. 12(a) is an example diagram of a read image of an invoice, which is an example of an accounting document, read by the reading unit 129. No streak image occurs in the read image of this invoice. FIG. 12(b) illustrates an example of a read image of an invoice when streak images 1201 and 1202 occur in two places in the main scanning direction. FIG. 12(c) illustrates an example of a read image of an invoice when streak image correction is not performed in the content area for the image of FIG. 12(b). The frame shown near the content area in FIG. 12(c) indicates an area determined to be a content area by block selection 417. Streak images 1211, 1212, 1221, 1222, 1223, 1224, and 1225 within this frame are not subjected to streak image correction.

[0093] The image reading device 100 of this embodiment reads accounting documents as specific documents in a special reading mode. This is because the contents of accounting documents such as invoices are treated as extremely important information, and loss of the contents may lead to conversion to erroneous information. Since loss of numbers (characters) on an invoice may lead to falsification of the invoice amount, invoice issue date, and expiration date, loss of the contents is extremely problematic. Therefore, priority is given to not performing image processing in areas including contents, and streak image correction is not performed.

[0094] The streak image correction performed outside the content area has a low risk of losing content because the area does not have content. Also, as shown in Fig. 12(b), if streak image correction is not performed outside the content area, streak images will remain over the entire area of ​​the scanned image in the sub-scanning direction, resulting in an image that is unpleasant for the user. For this reason, streak image correction is performed outside the content area.

[0095] Fig. 13 is an example diagram of a setting screen displayed on the display of operation panel 201 when setting a reading mode for a specific document. Note that Fig. 13 is an example for explaining this embodiment, and any setting screen may be used as long as it has a reading mode for a specific document such as an accounting document, and is not limited to these setting screens.

[0096] Fig. 13(a) is the processing mode setting screen. In Fig. 13(a), a "Document Scan" button is provided separately, in addition to the "Copy" button and the "Scan and Save" button for saving scanned images that are standard on copiers and multifunction devices. The "Document Scan" button is used to select the reading mode for a specific document. Pressing the "Document Scan" button on the processing mode setting screen transitions to the setting screen in Fig. 13(b).

[0097] FIG. 13(b) is an example of the top screen of the setting screen for document scan mode. This setting screen displays options for the type of document to be scanned, such as "invoice," "quote," "order form," "minutes," "contract," and "proposal." By registering a folder in advance to store scanned images for each type of document, it is possible to store folders in which scanned images are registered for each type of document, such as invoices and quotes. In addition, it is possible to set optional functions according to the type of document, such as displaying the OCR function for invoices and displaying an email send button for minutes. For example, pressing the "invoice" button on the screen in FIG. 13(b) transitions to the setting screen for invoice scan mode in FIG. 13(c).

[0098] 13(c) is the top screen of the invoice scan mode, which has standard functions such as color mode and resolution settings, as well as buttons for scan density, high compression, barcode, rotation correction, OCR, mark sheet, tilt correction, encryption, handwritten characters, etc. Displaying these frequently used functions on the top screen improves convenience. For example, when the OCR button is pressed and then the start button is pressed, OCR processing is performed on the scanned image, and the scanned image can be saved in the storage unit 208 as a PDF file including the content. OCR is performed, for example, by the CPU 202.

[0099] (Image reading processing) 14 is a flowchart showing the image reading process with streak image correction by the image reading device 100 configured as described above. This process is started when the user presses the image reading start button provided on the operation panel 201. Note that the reading mode for the specific document is set before the image reading start button is pressed.

[0100] The CPU 202 detects that a user has pressed an image reading start button provided on the operation panel 201 (S101). The CPU 202 detects the pressing of the image reading start button by acquiring a signal transmitted from the operation panel 201 when the image reading start button is pressed. The CPU 202 determines whether or not an original P is placed on the original tray 104 of the ADF 102 (S102). This determination is made, for example, based on the detection result of a sensor disposed in the original tray 104 for detecting the presence or absence of an original on the original tray 104.

[0101] When no document P is placed on the document tray 104 (S102: N), the CPU 202 reads the image of the document P placed on the reading glass 118 (pressure plate reading). When the document P is placed on the document tray 104 (S102: Y), the CPU 202 moves the lamps 119 and 120 and the reflecting mirror 121 to the reading position by the reading unit drive motor 204, and starts conveying the document P by the document conveying motor 205.

[0102] The CPU 202 determines whether the document detection sensor 127 has detected the conveyed document P (S104). If the document detection sensor 127 has not detected the document P (S104: N), the CPU 202 determines whether a predetermined time has elapsed (S105). That is, the CPU 202 determines whether the document P has reached the detection position of the document detection sensor 127 before the predetermined time has elapsed. If the document P has not reached the detection position of the document detection sensor 127 before the predetermined time has elapsed (S105: Y), the CPU 202 determines that a jam has occurred and ends the process (S124).

[0103] If the document P reaches the detection position of the document detection sensor 127 before the predetermined time has elapsed (S104: Y), the CPU 202 starts detection of streak images by the streak image detection unit 400 (S106). The CPU 202 reads a predetermined number of lines of the white guide member 114 by the reading unit 129 within the predetermined time until the document P reaches the reading position (S107, S108). If the predetermined number of lines cannot be read within the predetermined time (S108: Y), the CPU 202 determines that a jam has occurred and ends the process (S124).

[0104] If the predetermined number of lines are read within the predetermined time (S107: Y), the CPU 202 controls the transport amount of the document P by the number of pulses of the document transport motor 205, and judges whether the leading edge of the document P reaches the reading position within the predetermined time (S109, S110). The CPU 202 judges whether the document P reaches the reading position by judging whether the document P has been transported a predetermined amount by the number of pulses. If the document P has not been transported a predetermined amount within the predetermined time (S110: Y), the CPU 202 judges that a jam has occurred and ends the process (S124). If the document P has been transported a predetermined amount within the predetermined time (S109: Y), the CPU 202 judges that the leading edge of the document P has reached the reading position.

[0105] When the leading edge of the document P reaches the reading position, the CPU 202 starts reading the document P by the reading unit 129 (S112). The CPU 202 detects a streak image based on streak pixel candidates obtained from the read data of the document P and the read data of the white guide member 114 (S113). This identifies the position of the streak image in the main scanning direction. When the streak image detection is completed, the CPU 202 judges whether or not a read mode for a specific document is set (S113). If the read mode for a specific document is set (S113: Y), the CPU 202 detects a content area by the content area detection unit 410 (S114). If the read mode for a specific document is not set (S113: N), the CPU 202 does not detect a content area.

[0106] The CPU 202 judges whether a streak image is detected at the pixel of interest by the process of S109 (S115). If a streak image is detected (S115: Y), the CPU 202 judges whether the pixel of interest is outside the content area (S117). If the pixel of interest is outside the content area (S117: Y), the CPU 202 performs streak image correction by the streak image corrector 420 (S118). If a streak image is not detected (S115: N), or if a streak image is detected but the pixel of interest is within the content area (S117: N), the CPU 202 does not perform streak image correction by the streak image corrector 420 (S116). Note that if the reading mode for a specific document is not set, streak image correction is performed regardless of the content area.

[0107] The CPU 202 repeats the processes of S115 to S118 for each line until the rear end of the document P passes the reading position (S119: N, S120). When the CPU 202 detects that the document P has been conveyed a predetermined amount based on the number of pulses of the document conveying motor 205, it determines that the rear end of the document P has passed the reading position (S119: Y).

[0108] After the rear end of the document P passes the reading position, the CPU 202 judges whether the read document is the last sheet or not (S121). If the next document is in the document tray 104 and the read document is not the last sheet, the CPU 202 starts feeding the next document and repeats the processes of S110 to S118 (S121: N, S122). If the read document is the last sheet (S121: Y), the CPU 202 ejects the last sheet onto the paper ejection tray 117, stops the document transport motor 205, stops feeding of the documents, and ends the job normally (S123).

[0109] In the specific document reading mode, the image reading device 100 of the present embodiment as described above does not perform streak image correction in the vicinity of an area where content is printed on the document P. Therefore, the image reading device 100 can erase unnecessary streak images as much as possible while suppressing the risk of content loss.

[0110] Second embodiment The configuration of the image reading device 100 of the second embodiment is the same as that of the first embodiment, so the description will be omitted. The image reading device 100 of the second embodiment differs from the first embodiment in the configuration of the read image processing unit 207. The differences from the first embodiment will be described.

[0111] 15 is a diagram illustrating the configuration of the read image processing unit 207. The read image processing unit 207 of the second embodiment is different in configuration from the read image processing unit 207 of the first embodiment in that the streak image correction unit 1520 does not acquire the output from the content area detection unit 1510. The other configurations are the same.

[0112] In FIG. 15, the streak image correction unit 1520 does not obtain the output data BLK from the content area detection unit 1510, and outputs image data CORRECT after the streak image correction to the image processing control unit 1530. Since the content area is not specified in the image data CORRECT, the content is erased together with the streak image. The image processing control unit 1530 receives the image data CORRECT after the streak image correction, the read data before the streak image correction, and the output data BLK indicating the content area output from the content area detection unit 1510. The image processing control unit 1530 generates image data by restoring the content erased together with the streak image by the streak image correction by adding an image of the content area of ​​the read data before the correction to the read data after the correction. As a result, the streak image is erased from the read image while leaving the content.

[0113] In the second embodiment having the above configuration, image reading processing involving streak image correction is performed in the same manner as in the first embodiment. In the image reading device 100 of the second embodiment, like the first embodiment, it is possible to erase unnecessary streak images as much as possible while suppressing the risk of content loss.

[0114] The method of detecting the content area is not limited to the process by the content area detection unit 410 described in the first and second embodiments. For example, the content area may be detected for each predetermined pixel block by a machine learning model such as deep learning. That is, a machine learning model using a large number of image data representing content as training data may be used to detect the content area.

[0115] In the specific document reading mode, for example, the read image may be displayed on the operation panel 201, and the user may specify an area (i.e., a content area) where the streak image correction is not performed. The streak image corrector 420 performs the streak image correction on the area not specified by the user.

Claims

1. A conveying means for conveying the document; an image reading means for reading an image of the document conveyed by the conveying means using a reading sensor and generating image data representing the reading result; a control unit that detects a streak image in an image represented by the image data based on the image data and performs a correction process to remove the streak image by correcting the image data, the control means has a first reading mode in which a content area in an image represented by the image data in which content is printed is determined, the correction process is performed on image data corresponding to the streak image in an area excluding the content area, and the correction process is not performed on image data corresponding to the streak image in the content area; The first reading mode is a reading mode for reading a specific document. Image reading device.

2. A conveying means for conveying the document; an image reading means for reading an image of the document conveyed by the conveying means using a reading sensor and generating image data representing the reading result; a control unit that detects a streak image in an image represented by the image data based on the image data and performs a correction process to remove the streak image by correcting the image data; a white guide member disposed at a reading position of the image reading means on an opposite side to the image reading means with respect to a transport path through which the document passes; the control means has a first reading mode in which a content area in an image represented by the image data in which content is printed is determined, the correction process is performed on image data corresponding to the streak image in an area excluding the content area, and the correction process is not performed on image data corresponding to the streak image in the content area; The control means detects the streak image based on image data generated when the image reading means reads the white guide member and image data generated when the image reading means reads the document. Image reading device.

3. A conveying means for conveying the document; an image reading means for reading an image of the document conveyed by the conveying means using a reading sensor and generating image data representing the reading result; a control unit that detects a streak image in an image represented by the image data based on the image data and performs a correction process to remove the streak image by correcting the image data, the control means has a first reading mode in which a content area in an image represented by the image data in which content is printed is determined, the correction process is performed on image data corresponding to the streak image in an area excluding the content area, and the correction process is not performed on image data corresponding to the streak image in the content area; The content area is an area designated by a user. Image reading device.

4. The control unit is provided with a second reading mode in which the correction process is performed on image data corresponding to the streak image regardless of the presence or absence of the content area.

4. The image reading device according to claim 1.

5. The first reading mode is a reading mode for reading an accounting document as the original.

2. The image reading device according to claim 1.

6. The control means determines an area of ​​a predetermined size including the content as the content area.

3. The image reading device according to claim 1 or 2.

7. The content area is an area in which an image is printed.

3. The image reading device according to claim 1 or 2.

8. The content area is an area in which characters are printed.

3. The image reading device according to claim 1 or 2.

Images