JP2009181594A - Image reader - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image reader for generating high-quality image data. <P>SOLUTION: This invention provides the image reader having a reading means for generating the image data corresponding to a document by reading the document, a scan quality parameter extracting means for extracting a scan quality parameter for indicating quality of a scan image from the image data generated by the reading means based on scan miss detecting algorithm, a scan quality evaluating means for determining the existence of causing a scan miss based on the scan miss detecting algorithm and the scan quality parameter, and a notifying means for making a notification based on a determining result of the scan quality evaluating means. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an image reading apparatus. More specifically, the present invention relates to an image reading apparatus having a function for evaluating the image quality of a scanned image.

  A paper document (original) is read (scanned) by a scanner, digitized, used, or stored. At that time, it is desirable that the digitized document can faithfully reproduce the original document. For that purpose, for example, with regard to color tone and resolution, all paper documents may be scanned with multi-valued color and high resolution. However, color / monochrome is used as an expression mode of a document, and text / photographs are used as document contents. In practice, it is not always necessary to scan all documents with multi-valued color and high resolution. If such a document is scanned with color multi-value and high resolution, the work time and resources required for data storage will be consumed more than necessary.

  As described above, from the viewpoint of document reproducibility, a color multi-value and high-resolution scan is desirable, but from the viewpoint of work time and resource efficiency, a black-and-white binary and low-resolution scan is desirable. Therefore, during normal scanning operations, scanning is performed after an operator inputs processing parameters such as resolution and color / monochrome for each scanning operation. However, if the operator inputs inappropriate processing parameters, the quality of the scanned image (hereinafter referred to as “scan quality”) decreases, for example, the photograph or fine characters are crushed so that they cannot be identified. There is a risk that. For this reason, it is necessary to confirm the scan quality after the work (confirm whether the scan parameters entered by the operator are appropriate), but the scan quality is confirmed by the operator's visual inspection. Currently. However, it takes a lot of man-hours for visual confirmation by the worker, or because it depends on the worker, the judgment of good / bad for each worker is ambiguous. When scanning a large number of documents with multiple workers, However, there is a problem that the quality of the scanned image is not stable.

  As a technique for solving such a problem, Patent Document 1 discloses a technique for checking a file format of scanned image data and outputting the check result as a table. Patent Documents 2 and 3 disclose techniques for checking the operating state of an image output apparatus by scanning a document output by an image output apparatus such as a printer and checking the image. Patent Documents 4 and 5 disclose techniques for evaluating the image quality of image data.

JP 2002-77468 A Japanese Patent No. 3160898 Japanese Patent No. 3086520 Japanese Patent No. 3076692 Japanese Patent Application Laid-Open No. 11-203411

However, in the technique described in Patent Document 1, the check result is only presented as a table, and when performing a rescan, the operator must search for a document that is determined to be defective by itself and rescan it, There was a problem that work efficiency was bad. Further, the techniques described in Patent Documents 2 and 3 evaluate the characteristics of the image output apparatus by scanning statistical processing or a test pattern, and there is a problem that the scanned image cannot be evaluated. Furthermore, scan errors that can be detected by the image quality evaluation technique described in Patent Document 4 or 5 are very limited. In addition, there is a problem that even if the image data is checked with these conventional techniques, it does not contribute to the improvement of the skill of the worker.
The present invention has been made in view of the above circumstances, and an object of the present invention is to provide an image reading apparatus that is highly efficient, has a scan quality of a certain level or more, and contributes to improving the skill of the operator.

  According to a first aspect of the present invention, there is provided an image reading apparatus that reads a document and generates a scan image corresponding to the document, and a quality of the scan image generated by the reading unit based on a scan error detection algorithm. A scan quality parameter extracting means for extracting a scan quality parameter, a scan quality evaluation means for judging whether or not a scan error has occurred based on the scan error detection algorithm, and the scan quality parameter, and a scan quality evaluation means Notification means for making a notification based on the determination result, wherein the scan error detection algorithm divides the scanned image into a plurality of partial images, and calculates an estimated number of characters for each of the plurality of partial images; A character recognition process for each of the plurality of partial images. And a step of calculating a character recognition rate in each of the plurality of partial images based on the estimated number of characters and the number of characters recognized by the character recognition process, and at least one or more of the partial images When the character recognition rate is lower than the threshold value of the character recognition rate, it is determined that a scan error has occurred.

  According to a second aspect of the present invention, there is provided an image reading apparatus that reads a document and generates a scan image corresponding to the document, and a quality of the scan image generated by the reading unit based on a scan error detection algorithm. A scan quality parameter extracting means for extracting a scan quality parameter, a scan quality evaluation means for judging whether or not a scan error has occurred based on the scan error detection algorithm, and the scan quality parameter, and a scan quality evaluation means Notification means for making a notification based on the determination result, wherein the scan error detection algorithm divides the scanned image into a plurality of partial images, and calculates an estimated number of characters for each of the plurality of partial images; A character recognition process for each of the plurality of partial images. And a step of calculating a character recognition rate in each of the plurality of partial images based on the estimated number of characters and the number of characters recognized by the character recognition processing, and among the plurality of partial images A difference between a character recognition rate in the first partial image that is a partial region of the first partial image and a character recognition rate in the second partial image that is a region excluding the first partial image of the plurality of partial images; Alternatively, a configuration for determining that a scan error has occurred when a difference between a character recognition rate in the first partial image and an average of the character recognition rates in the first and second partial images is equal to or greater than a threshold value. Have.

  According to a third aspect of the present invention, there is provided an image reading apparatus that reads a document and generates a scan image corresponding to the document, and a quality of the scan image generated by the reading unit based on a scan error detection algorithm. A scan quality parameter extracting means for extracting a scan quality parameter, a scan quality evaluation means for judging whether or not a scan error has occurred based on the scan error detection algorithm, and the scan quality parameter, and a scan quality evaluation means A step of rewriting scan parameters so that the scan error detection algorithm performs a scan with a size larger than the document size input by the operator before the scan is performed. And the scanned image And extracting the image outside the document size area except for the document size image and determining whether a significant pixel exists in the image outside the area. When it is determined that a pixel exists, it is determined that a scan error has occurred.

  According to the present invention, even when a large amount of documents are scanned by a plurality of workers, the quality of the scanned image can be maintained at a certain level. Further, it is possible to improve the efficiency of scanning work by automatically detecting a scanning error. Furthermore, the operator's skill can be improved by presenting the scan error to the worker.

1 is a block diagram showing a configuration of an image reading system 1 according to the present invention. 1 is a diagram illustrating a configuration of a scanner 100. FIG. It is a figure which shows the positional relationship of the book 61 and the platen glass 20. FIG. It is a figure which illustrates the scan mistake which concerns on a curve. It is a flowchart which shows the algorithm of bending determination. It is a flowchart which shows the algorithm of a top-and-bottom determination. 4 is a flowchart illustrating an operation of the scanner 100 according to the first embodiment. It is a figure which shows the example of the screen which notifies an operator of the detection result of a scanning mistake. 3 is a flowchart showing the operation of the scanner 100. 3 is a flowchart showing the operation of the scanner 100. 3 is a flowchart showing the operation of the scanner 100. It is a figure which illustrates the message which urges resetting of a manuscript to ADF. It is a figure which illustrates scan quality index conversion table TB3. It is a figure which shows the example of the screen which prompts the operator to judge rescanning.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
<1. Configuration>
<1.1 Image reading system configuration>
FIG. 1 is a block diagram showing a configuration of an image reading system 1 according to an embodiment of the present invention. The scanner 100 is an image reading apparatus having an ADF (Automatic Document Feeder) 10 and optically reads images such as characters and photographs written on the original from a sheet-like original 60 to be read. Electronic data corresponding to an image is generated. The PC 200 is connected to the scanner 100 via the cable 300, and receives, stores, or displays the electronic document from the scanner 100 on the display 210.

FIG. 2 is a diagram illustrating the configuration of the scanner 100.
In the present embodiment, the scanner 100 has an ADF 10. The operator automatically scans a plurality of documents one by one to the document reading position by using the ADF 10, or manually scans the documents one by one without using the ADF 10. You can also scan it on top.

  When scanning using the ADF 10, the operator sets a plurality of documents 60 on the tray 11. The tray 11 is provided with a sensor (not shown), detects the presence of a document set in the tray 11, and outputs a signal indicating that. The original 60 is conveyed to the conveying roller 13 one by one by the drawing roller 12. The transport roller 13 transports the document 60 toward the platen glass 20 while changing the document transport direction. The document 60 conveyed in this manner is pressed against the platen glass 21 by the back platen 17 and is finally discharged from the ADF 10 by the discharge roller 14. On the platen glass 20, four document reading positions are provided from the upstream to the downstream of the conveyance path. At each of these reading positions, the document 60 is conveyed at a constant speed. Reflected light (original image) emitted from the light source 31a and reflected by the original 60 at the reading position is changed in optical path by the mirror 31b, mirror 32a, and mirror 32b, condensed by the lens 33, and charged by a CCD (Charge Coupled Devices). ) An image is formed on the sensor 34. The CCD sensor 34 is, for example, a 4-line CCD sensor, and outputs analog image signals R, G, B, and BW to the subsequent circuit in accordance with the reflected light (original image) input at each reading position.

  When the reading is completed, the document 60 is conveyed by the conveyance roller 16 and is discharged to the discharge tray 18 via the discharge roller 14. The transport roller 16 has a function of changing the transport direction of the document 60 in accordance with a signal from the ADF drive control circuit 43, and the document 60 can be discharged to the discharge tray 19 via the discharge roller 15.

  When scanning without using the ADF 10, the operator sets the originals 60 one by one on the platen glass 20. When a scan instruction is input by a method such as pressing the start button of the operation unit 44, the first CRG 31 including the light source 31a reads the document image while moving in the direction A in FIG. The CCD sensor 34 outputs analog image signals R, G, B, and BW to the subsequent circuit as in the case of using the ADF 10.

  Analog image signals R, G, B, and BW output from the CCD sensor 34 are converted into digital image data R, G, B, and BW by an A / D conversion circuit 49. These digital image data R, G, B, and BW are corrected according to sensitivity variations of the CCD sensor 18 and light quantity distribution characteristics of the optical system by a shading correction circuit (not shown) or the like. The digital image data R, G, B, BW corrected in this way is output to the subsequent image processing circuit 50. Hereinafter, part or all of the digital image data R, G, B, and BW is simply referred to as “image data”, and an image indicated by the image data is referred to as “scanned image”.

  The image processing circuit 50 includes a scan quality parameter extraction unit 51 that extracts a parameter indicating the quality of a scanned image (hereinafter referred to as “scan quality parameter”) from image data according to a certain algorithm, and a scan error based on the extracted scan quality parameter. Is composed of a scan quality evaluation unit 52 that determines whether or not the image has occurred (evaluates scan image quality) and an image processing unit 53 that processes image data. Details of the image processing circuit 50 will be described later.

  The ADF drive control circuit 43 drives the ADF 10 according to a control signal from a CPU (Central Processing Unit) 41. The ADF drive control circuit 43 has a function of counting the number of documents (number of pages) conveyed by the ADF 10 and outputting a signal indicating the number of documents (number of pages). The display unit 46 is configured by a liquid crystal display, for example, and displays a message to the worker and a work status according to a control signal from the CPU 41. The operation unit 44 includes a numeric keypad, a start button, a stop button, a touch panel installed on the liquid crystal display, and the like, and outputs a signal corresponding to an operation input by the operator and a display screen at that time. The storage unit 45 stores various control programs, threshold values used in scan error determination described later, various tables, and the like. The I / F 47 is an interface for transmitting / receiving data to / from an external device such as the PC 200. These elements are connected to each other via a bus 48 and can exchange data.

<1.2 Image evaluation algorithm>
As an algorithm (hereinafter referred to as a “scan error detection algorithm”) for detecting a scan error (evaluation of the quality of a scan image) executed by the image processing circuit 50, for example, there is the following.

(1) Skew determination Scanning may be performed in a state where the document is inclined from the correct setting position because the document is set on the ADF 10 at an angle or the document is set on the platen glass 20 at an angle. . Scanning a document in such a tilted state is called skew (in this case, the tilt angle of the scanned image is called skew angle). Skew determination is an algorithm for detecting such a scan error.

  The scan quality parameter extraction unit 51 calculates a skew angle from scan image data, for example, by a technique described in Japanese Patent Application Laid-Open No. 2002-84420. The scan quality parameter extraction unit 51 adds an identifier indicating that the parameter relates to skew determination, and outputs the calculated skew angle data to the subsequent scan quality evaluation unit 52 as a scan quality parameter.

  The scan quality evaluation unit 52 determines that the data indicates the skew angle from the identifier of the input data. The scan quality evaluation unit 52 reads the threshold value of the skew angle from the storage unit 45 and compares it with the skew angle of the scan image. When the skew angle of the document image is larger than the threshold value, the scan quality evaluation unit 52 determines that the skew has occurred (that is, a scan error has occurred), and determines the occurrence of the scan error and the identifier of the scan error detection algorithm. Is output to the CPU 41. When it is not determined that the skew has occurred (that is, when it is determined that there is no scan error), the scan quality evaluation unit 52 outputs a signal indicating that to the CPU 41.

(2) Folding determination If a document is scanned while it is bent, the information described in the bent portion does not remain as data, and such a document needs to be rescanned. Break determination is an algorithm for detecting such a scan error.

  The scan quality parameter extraction unit 51 detects the outline (outline) of the document from the scan image data by a known method, and calculates the area C of the region surrounded by the outline. Subsequently, the scan quality parameter extraction unit 51 calculates a rectangular area A having a minimum area that envelops the detected contour. The scan quality parameter extraction unit 51 calculates an area ratio C / A, adds an identifier indicating that the parameter is related to breakage determination, and outputs it to the scan quality evaluation unit 52 as a scan quality parameter.

  When the scan quality evaluation unit 52 determines that the input data relates to the break determination, the scan quality evaluation unit 52 reads the C / A threshold value from the storage unit 45. The scan quality evaluation unit 52 compares the scan quality parameter C / A with the threshold value, and determines that a break has occurred (a scan error has occurred) if the C / A is smaller than the threshold value. The scan quality evaluation unit 52 outputs a signal indicating the occurrence of the scan error and the identifier of the scan error detection algorithm to the CPU 41. When it is not determined that a fold has occurred (when it is determined that there is no scan error), the scan quality evaluation unit 52 outputs a signal indicating that to the CPU 41.

  Alternatively, the break determination may be performed as follows. The scan quality parameter extraction unit 51 detects the contour (outer shape) of the document from the scan image data by a known method. Subsequently, the scan quality parameter extraction unit 51 calculates the number of corners of the detected contour and the angle of each corner. The scan quality parameter extraction unit 51 adds an identifier indicating that the parameter relates to breakage determination, and outputs the number of corners and angle data to the scan quality evaluation unit 52 as scan quality parameters.

  When the scan quality evaluation unit 52 determines that the input data is related to breakage determination, it first determines whether the number of corners is four. If the number of corners is not 4, it is determined that a break has occurred (a scan error has occurred), and the scan quality evaluation unit 52 outputs a signal indicating that to the CPU 41. When the number of corners is 4, the scan quality evaluation unit 52 further determines whether the angles of the four corners are all 90 °. If there is even one angle other than 90 °, it is determined that a fold has occurred. The subsequent operation is as described above.

(3) Curvature Determination FIG. 3 is a diagram illustrating a positional relationship between the book 61 and the platen glass 20 when the facing page of the book 61 is read as a document. As shown in FIG. 3, when a document such as a book that is bound on one side is opened and placed on the platen glass 20 for reading, the binding portion of the document may float from the platen glass 20. When reading is performed in this state, characters and figures are distorted, or blurs and black shadows are generated near the binding portion of the scanned image (FIG. 4). Bending determination is an algorithm for detecting such a scan error.

  FIG. 5 is a flowchart showing an algorithm for bending determination. The scan quality parameter extraction unit 51 first calculates an average gradation value of all pixels of the scan image. The scan quality parameter extraction unit 51 reads the character number estimation table TB1 stored in the storage unit 45 and recorded in association with the average gradation value and the estimated number of characters. The scan quality parameter extraction unit 51 calculates an estimated value of the number of characters of the entire document based on the character number estimation table TB1 and the calculated average gradation value (step S101). Subsequently, the scan quality parameter extraction unit 51 divides the read document image into a plurality of regions (step S102). The dotted line in FIG. 4 shows an example in which the document image is equally divided into 16 areas. The scan quality parameter extraction unit 51 performs OCR (Optical Character Reader) processing on each of these areas (step S103). The scan quality parameter extraction unit 51 calculates the character recognition rate of each region using the number of characters recognized by the OCR process and the estimated number of characters per region (1/16 of the estimated number of characters of the entire document). (Step S104). The scan quality parameter extraction unit 51 adds an identifier indicating that the parameter relates to the bending determination, and outputs the character recognition rate data of each region to the scan quality evaluation unit 52 as a scan quality parameter.

When the scan quality evaluation unit 52 determines that the input data relates to the bending determination, the scan quality evaluation unit 52 reads the threshold value data of the character recognition rate from the storage unit 45. The scan quality evaluation unit 52 compares the character recognition rate with a threshold value for each region (step S105). If there is a region where the character recognition rate is lower than the threshold, it is determined that a curve has occurred (step S106), and the scan quality evaluation unit 52 indicates the occurrence of the scan error and the identifier of the scan error detection algorithm. The signal is output to the CPU 41. If it is not determined that a curve has occurred (step S107), the scan quality evaluation unit 52 outputs a signal indicating that to the CPU 41.
It should be noted that the step of determining the occurrence of bending may be configured to determine that there is a bending when the difference between the character recognition rate of a certain area and the average character recognition rate of all the areas is equal to or greater than a threshold value. . Alternatively, when the difference between the character recognition rate in the central portion of the image and the character recognition rate in the peripheral portion of the image is greater than or equal to a threshold value, it may be determined that there is a bend.

(4) Top-to-bottom determination If the original is not set in the orientation determined by the structure of the scanner 100, it will be read upside down or rotated 90 ° clockwise or counterclockwise. The top / bottom determination is an algorithm for detecting such a scan error.

  FIG. 6 is a flowchart showing the top / bottom determination algorithm. The scan quality parameter extraction unit 51 first performs OCR processing on the scanned image as it is (that is, a 0 ° rotated image), and calculates the number of recognized characters (steps S201 and S202). The scan quality parameter extraction unit 51 further performs OCR processing on the image obtained by rotating the scanned image by 90 °, 180 °, and 270 °, and calculates the number of recognized characters (steps S202, S203, and S206). The scan quality parameter extraction unit 51 adds an identifier indicating that this parameter is related to the top / bottom determination, and uses the four rotation angles and the data of the number of recognized characters corresponding to each as the scan quality parameter. To 52.

When the scan quality evaluation unit 52 determines that the input data is related to the upside-down inversion, it extracts the maximum number of recognized characters. The scan quality evaluation unit 52 determines whether or not the extracted maximum number of recognized characters corresponds to the 0 ° rotated image (step S204). If the data does not correspond to the rotation angle of 0 °, it is determined that the document is not set correctly (step S205), and the scan quality evaluation unit 52 determines the occurrence of the scan error and the identifier of the scan error detection algorithm. The signal shown is output to the CPU 41. Otherwise (step S207), the scan quality evaluation unit 52 outputs a signal indicating that to the CPU 41.
The scan quality parameter extraction unit 51 outputs the rotation angle corresponding to the maximum number of recognized characters to the scan quality evaluation unit 52 as the scan quality parameter, and the scan quality evaluation unit 52 receives the input rotation angle. If the angle is not 0 °, it may be determined that a scan error has occurred.

(5) Overflow determination In some cases, a sticky note is pasted on an original, and it is required to scan each character written down on the sticky note. Here, for example, if a sticky note is attached to the edge of an A4 size document, even if this sticky note original is scanned in A4 size, the sticky note portion is located outside the reading range and cannot be read. There is. The protrusion determination is an algorithm for detecting such a scan error.
When performing the protrusion determination, the scanner 100 scans with a size larger than the document size specified by the scan instruction. That is, when the execution of the protrusion determination is instructed before scanning, the CPU 41 reads the scan size conversion table TB2 from the storage unit 45 at the time of scanning. In the scan size conversion table TB2, a document size and a document size one size larger than that are recorded in association with each other. The CPU 41 extracts the document size from the scan parameters at the time of scanning, and extracts a document size one size larger than the document size from the scan size conversion table. The CPU 41 performs a scan operation after rewriting the scan parameters so that the extracted document size one size larger becomes the document size at the time of scanning.

  The scan quality parameter extraction unit 51 removes the original document size image from the scan image, and extracts an image outside the paper size area specified by the scan instruction. The scan quality parameter extraction unit 51 inspects whether there is a significant pixel value in an image outside this region. That is, the scan quality evaluation unit 52 reads the threshold value of the pixel gradation value and the threshold value of the magnitude from the storage unit 45. The scan quality evaluation unit 52 inspects an image outside this region, and if pixels having a gradation value equal to or greater than the threshold value are continuously detected in the region greater than or equal to the threshold value, the scan quality evaluation unit 52 It is determined that there is a significant pixel. If the scan quality parameter extraction unit 51 determines that a significant pixel exists, the scan quality parameter value is set to, for example, “1”, and if it is determined that there is no significant pixel outside the region, the scan quality parameter value is set. For example, “0” is set. The scan quality parameter extraction unit 51 adds an identifier indicating that this scan quality parameter relates to the protrusion determination, and outputs the identifier to the scan quality evaluation unit 52.

When the scan quality evaluation unit 52 determines that the input data relates to the protrusion determination, the scan quality evaluation unit 52 determines whether the value of the scan quality parameter is 1. If the value of the scan quality parameter is 1, it is determined that there is an overflow, and the scan quality evaluation unit 52 outputs a signal indicating the occurrence of the scan error and the identifier of the scan error detection algorithm to the CPU 41. If it is not determined that there is a protrusion, the scan quality evaluation unit 52 outputs a signal indicating that to the CPU 41.
Note that, at the time of scanning, it may be configured to scan with the maximum scan size of the scanner 100 instead of scanning with an image size one size larger than the scan instruction.

(6) Resolution determination If a document on which a photo or fine text is written has a reading resolution that is too low, the text or photo will be crushed and information cannot be used. Resolution determination is an algorithm for detecting such a scanning error.

  The scan quality parameter extraction unit 51 first performs a Fourier transform on the scanned image data, and calculates the power density as a function of frequency. The scan quality parameter extraction unit 51 calculates an average power density in a high frequency region that is a predetermined frequency region. The scan quality parameter extraction unit 51 adds an identifier indicating that this parameter relates to resolution determination, and outputs average power density data to the scan quality evaluation unit 52 as a scan quality parameter.

  When the scan quality evaluation unit 52 determines that the input data is related to resolution determination, the scan quality evaluation unit 52 reads the power density threshold value from the storage unit 45. The scan quality evaluation unit 52 compares the input average power density with a threshold value. When the average power density is smaller than the threshold value, it is determined that the set resolution is too low, and the scan quality evaluation unit 52 outputs a signal indicating the occurrence of the scan error and the identifier of the scan error detection algorithm to the CPU 41. Otherwise, the scan quality evaluation unit 52 outputs a signal indicating that to the CPU 41.

(7) Color / monochrome / binary / multi-value determination If a full-color photo is scanned with binary binary, the photo image cannot be identified. When scanned and stored, the data size becomes larger than necessary, which is inconvenient. Color / monochrome / binary / multi-value determination is an algorithm for detecting a scan error due to a mismatch between such a scan instruction and a document.
When performing color / monochrome / binary / multi-value determination, the scanner 100 always performs scanning with color multi-value regardless of the scan parameter specified by the scan instruction. That is, if execution of color / monochrome / binary / multi-value determination is instructed before scanning, the CPU 41 extracts the color / monochrome / binary / multi-value parameters from the scan parameters at the time of scanning. The scan operation is performed after the parameter is rewritten to the value.

  The scan quality parameter extraction unit 51 determines whether the scan image is color or black and white, binary image or multi-value image by a known technique. The scan quality parameter extraction unit 51 rewrites the scan quality parameter with a value indicating the determination result, adds an identifier indicating that the parameter relates to color / monochrome / binary / multi-value determination, and evaluates the scan quality To the unit 52.

  The scan quality evaluation unit 52 compares the input parameter and the parameter specified by the scan instruction, and if they are different from each other, determines that a scan error has occurred. A signal indicating the identifier is output to the CPU 41. Otherwise, the scan quality evaluation unit 52 outputs a signal indicating that to the CPU 41.

(8) Tone / Gamma / Contrast Determination When scanning a photograph or the like, parameters such as tone, gamma, and contrast may be set manually in order to make the gradation of the scanned image appropriate. In such a case, if the setting values of parameters such as tone, gamma, and contrast are not appropriate, the original image is accurately displayed such that the dark part is crushed in black or the bright part is skipped in white. It becomes impossible to reproduce. Tone / gamma / contrast determination is an algorithm for detecting such a scan error.

  The scan quality parameter extraction unit 51 creates a histogram of the scan image. That is, the scan quality parameter extraction unit 51 equally divides the gradation value from shadow (0) to highlight (for example, 255 in the case of 8 bits) into a predetermined number of areas, and sets variables corresponding to each area. . Each variable represents the frequency of pixels in the corresponding gradation area. The scan quality parameter extraction unit 51 examines the gradation value of a certain pixel and increases the value of the variable corresponding to the region to which the gradation value belongs by one. The scan quality parameter extraction unit 51 creates a gradation value histogram by examining gradation values for all pixels. Subsequently, the scan quality parameter extraction unit 51 reads the frequency threshold value from the storage unit 45 and compares it with the value of each variable. The scan quality parameter extraction unit 51 updates the value of a variable whose value is lower than the threshold value to 0. Subsequently, the scan quality parameter extraction unit 51 calculates a gradation range whose frequency is a positive value. That is, the scan quality parameter extraction unit 51 extracts the variable having the positive tone value and the tone value having the maximum tone value and calculates the difference, that is, the tone value width. The scan quality parameter extraction unit 51 adds an identifier indicating that this parameter relates to tone / gamma / contrast determination, and outputs the data of the gradation value width to the scan quality evaluation unit 52 as a scan quality parameter.

  When the scan quality evaluation unit 52 determines that the input data relates to tone / gamma / contrast determination, the scan quality evaluation unit 52 reads the threshold value of the gradation value width from the storage unit 45. The scan quality evaluation unit 52 compares the input gradation range with a threshold value. If the input gradation value width is narrower than the threshold value, it is determined that the tone / gamma / contrast setting is not appropriate, and the scan quality evaluation unit 52 determines the occurrence of the scan error and the identifier of the scan error detection algorithm. The signal shown is output to the CPU 41. If the input gradation value width is greater than or equal to the threshold value, the scan quality evaluation unit 52 outputs a signal indicating that to the CPU 41.

  The scan quality evaluation unit 52 calculates a recommended contrast value so that the operator can set the contrast to an appropriate value when re-scanning when the gradation value width is equal to or smaller than the threshold value. However, the recommended value may be displayed on the display unit 46. Specifically, the storage unit 45 stores a table in which the gradation value width and the recommended contrast value are recorded in association with each other, and the scan quality evaluation unit 52 refers to this table to determine the recommended contrast value. calculate.

(9) Character image quality determination In order to reduce the amount of data at the time of storage, it is desirable to scan with a reduced resolution. However, if the resolution is lowered too much, characters are crushed and information cannot be used. The character image quality determination is an algorithm for detecting such a scan error.

  The scan quality parameter extraction unit 51 first performs a well-known layout extraction process on the scanned image, and decomposes the scanned image into partial images. The scan quality parameter extraction unit 51 performs OCR processing on each partial image, and measures the number of characters that can be recognized for each partial image. The scan quality parameter extraction unit 51 performs a Fourier transform on the partial image having the largest number of recognized characters, and calculates the average power density of the high-frequency component specific to the character image. The scan quality parameter extraction unit 51 adds an identifier indicating that this parameter relates to character image quality determination, and outputs average power density data to the scan quality evaluation unit 52 as a scan quality parameter.

When the scan quality evaluation unit 52 determines that the input data relates to character image quality determination, the scan quality evaluation unit 52 reads a threshold value of power density related to character image quality determination from the storage unit 45. The scan quality evaluation unit 52 compares the input power density with a threshold value. If the input power density is lower than the threshold value, it is determined that the resolution is insufficient, and the scan quality evaluation unit 52 sends a signal indicating the occurrence of the scan error and the identifier of the scan error detection algorithm to the CPU 41. Output to. Otherwise, the scan quality evaluation unit 52 outputs a signal indicating that to the CPU 41.
Note that the scan quality parameter extraction unit 51 may calculate the area of each partial image and perform frequency analysis on the partial image having the largest number of characters per unit area.

<2. Operation>
<2.1 Outline of operation>
The operation of the scanner 100 according to the present embodiment is generally as follows. That is, when scanning the document 60, the scanner 100 checks the quality of the scanned image according to at least one of the above-described scan error detection algorithms or a known scan error detection algorithm. If it is determined that a scan error has occurred, the operator is notified of this. The scanner 100 stores a program for executing at least one of the plurality of scan error detection algorithms described above and other known scan error detection algorithms in the storage unit 45. In the present embodiment, the operator can select the scan error detection algorithm to be applied prior to the scanning operation. Further, the scanner 100 can perform scanning without executing a scan error detection algorithm. Hereinafter, the scan mode in which scanning is performed while applying the scan error detection algorithm is referred to as “normal scan mode”, and the scan mode in which scanning is performed without applying the scan error detection algorithm is referred to as “forced scan mode”. Note that the scan error detection algorithm executed at the time of scanning may be predetermined by the scanner 100 and unchanged.

  Hereinafter, in the first embodiment, an aspect in which scanning is performed by manually setting the original 60 on the platen glass 20 one by one without using the ADF 10 will be described, and in the second embodiment, an aspect in which scanning will be performed using the ADF 10 will be described. To do. In the first and second embodiments, the evaluation of the scanned image is either a scan error or an alternative, but in the third embodiment, an aspect in which the evaluation of the scanned image is expressed by a numerical value will be described. In the fourth embodiment, a mode in which a scanned image is automatically corrected will be described.

<2.2 First Embodiment>
FIG. 7 is a flowchart showing the operation of the scanner 100 according to the first embodiment. FIG. 8 is a diagram showing an example of a screen for notifying the operator of the detection result of the scan error. First, the operator places one of the plurality of originals 60 on the platen glass 20, performs an operation such as selecting a scan mode, and then issues a scan instruction to the scanner 100 by pressing a start button. Then, image data corresponding to the document 60 is generated in the scanner 100 as described above (step S301).

  The scanner 100 executes at least one of the scan error detection algorithms described above on the scanned image (step S302). As a result of applying the scan error detection algorithm, if a scan error is detected (step S303: YES), as described above, the scan quality evaluation unit 52 sends the CPU 41 the occurrence of the scan error, and the identifier of the scan error detection algorithm. Is output. When receiving a signal indicating that a scan miss has been detected, the CPU 41 reads a scan miss message table from the storage unit 45. The scan miss message table is a table in which an identifier of a scan miss detection algorithm and a message for notifying an operator of a scan miss are associated with each other and recorded. The CPU 41 extracts a message corresponding to the scan error from the scan miss message table (step S304), and displays the extracted message on the display unit 46 (step S305). For example, when a document break is detected by the above-described folding determination algorithm, a message “Scan failed. Please stretch the paper and rescan” is displayed (FIG. 8B). If it is determined by the protrusion determination algorithm that there is an protrusion, a message “Scanning failed. Change the image size and rescan” is displayed (FIG. 8D). Alternatively, when it is determined that the contrast is insufficient by the tone / gamma / contrast determination, a message “Scanning failed. Re-scan with the contrast as xx” is displayed (FIG. 8C). ). As a result, the operator can learn that the scan parameter should have been set in this way for the document, and the skill of the operator can be improved.

  When a scan error is not detected (step S303: NO), the CPU 41 controls the display unit 46 and displays a message indicating that (step S306). For example, “Scan succeeded. Next scan is possible” is displayed (FIG. 8A), and the operator is prompted to proceed to the next scan operation.

  A plurality of algorithms may be applied during scanning. When a plurality of scan mistakes are detected by a plurality of algorithms, a plurality of messages corresponding to the plurality of scan mistakes may be displayed on the display unit 46. Alternatively, if the storage unit 45 stores a threshold value for the number of scan errors and the number of scan errors exceeding the threshold value is detected, the cause of the scan error cannot be specified, and the “scan” You may display a message such as “Failed, check originals or scan parameters and rescan, or scan in forced scan mode” (FIG. 8E).

  According to this embodiment, when a document is manually scanned, the presence or absence of a scanning error is immediately detected, so that the process of visually confirming the scanned image by an operator can be omitted, and the efficiency of the scanning operation can be reduced. Can be planned.

<2.3 Second Embodiment>
In the present embodiment, a mode in which scanning is performed using the ADF 10 will be described. When the ADF 10 is used, a plurality of documents 60 are conveyed to the reading position one by one and scanned. As in the first embodiment, the scanner 100 executes at least one scan miss detection algorithm. When a scan miss is detected, the timing for instructing a rescan is (1) when a scan miss is detected, and (2 ) Two methods after the end of all page scan are possible. Hereinafter, each case will be described separately.

(1) When Rescan is Instructed when Scan Miss is Detected FIG. 9 is a flowchart showing the operation of the scanner 100 according to the present embodiment. The operator sets a plurality of documents 60 on the ADF 10 and presses the start button. Then, the plurality of documents 60 are conveyed one by one to the reading position, and image data corresponding to the documents 60 is generated inside the scanner 100 as described above (step S401). The image processing circuit 50 detects a scan error in accordance with the above-described scan error detection algorithm every time one page of the document is read (steps S402 and S403). When receiving a signal indicating the occurrence of a scan error, the CPU 41 outputs a signal instructing to stop the ADF 10 to the ADF drive control circuit 43. In accordance with this signal, the ADF drive control circuit 43 controls the ADF 10 to stop the conveyance of the document (step S404). Then, the scanner 100 presents the document in such a manner that the operator can identify the document in which the scanning error has occurred (step S405). That is, the CPU 41 controls the ADF drive control circuit 43 so that only the document in which the scan error has occurred is discharged to the discharge tray 19. As a result, only the document that needs to be rescanned is discharged to a discharge tray 19 that is different from the discharge tray 18 from which the document that has been successfully scanned is discharged. At this time, a message prompting rescanning as in the first embodiment (for example, “Scanning failed. Remove the document from the discharge tray, stretch the paper, place the document on the platen, and press the start button”) Is displayed on the display unit 46 (steps S406 and S407). The operator takes out a document that needs to be rescanned from the discharge tray 19, stretches the paper, and places it on the platen glass 20. Subsequently, the worker presses the start button as instructed by the message. When the CPU 41 receives a signal corresponding to pressing of the start button after displaying a message prompting re-scanning, the CPU 41 uses an optical system such as the first CRG 31 and the second CRG 32 so as to scan one document placed on the platen glass 20. Control. When the original is read from the platen glass 20, the CPU 41 controls the ADF drive control circuit 43 to perform scanning using the ADF 10 from the page next to the page where the scan error has occurred.

  If a scan error is not detected by the scan error detection algorithm (step S403: NO), the CPU 41 determines whether or not scanning has been completed for all documents (all pages) (step S408). This determination is made, for example, by determining whether or not a document remains in the tray 11 based on a signal from a sensor provided on the tray 11. When all the originals have been scanned, the operation of the scanner 100 ends. If the document still remains, the next document is conveyed (step S409), and the above operation is repeated until the scanning of all the documents is completed.

  Note that the timing for instructing rescanning is not limited to when a scan error on one page of a document is detected. For example, a configuration may be adopted in which rescanning is instructed when scanning mistakes are continuously detected for a predetermined number of documents. That is, a constant indicating the allowable number of consecutive scan errors is stored in the storage unit 45. Each time the CPU 41 receives a signal indicating the occurrence of a scan error, the CPU 41 increments the value of a variable indicating the number of occurrences of the scan error by one. The CPU 41 compares this variable with a constant, and when they match (or when the variable becomes larger than the constant), stops the ADF 10 and instructs rescan. In addition, a constant indicating the allowable number of consecutive scan errors may be used as a variable, and the value may be rewritten by an operator's instruction input or the like. In addition, the execution order of the above-described presentation of a scan error-occurring document (step S405) and the display of a message indicating a scan error (steps S406 and S407) may be switched.

  According to the present embodiment, during a scanning operation using the ADF, the presence or absence of a scanning error is checked every time one page of the document is scanned. In addition, when a scan error is detected, the document (page) in which the scan error has occurred is presented in a form that can be specified, and information indicating the type of scan error is displayed, so rescanning can be performed easily. Can do.

(2) When Re-Scanning is Instructed After All-Page Scans FIGS. 10A and 10B are flowcharts showing the operation of the scanner 100 according to this embodiment. The operator sets a plurality of documents 60 on the ADF 10 and presses the start button. Then, the plurality of documents 60 are conveyed one by one to the reading position, and image data corresponding to the documents 60 is generated inside the scanner 100 as described above (step S501). The image processing circuit 50 evaluates the scanned image in accordance with the scan error detection algorithm every time one page of the document is read (steps S502 and S503). When the CPU 41 receives a signal indicating the occurrence of a scan error, the CPU 41 stores in the storage unit 45 information identifying the page where the scan error has occurred, such as the page number of the document counted up by the ADF drive control circuit 43 (step S504). ). When a scan error is detected in a plurality of pages, the CPU 41 stores a plurality of pieces of information for specifying the page in which the scan error has occurred in the storage unit 45. The CPU 41 determines whether scanning of all documents (all pages) has been completed (step S505). If all the originals have not been scanned, the next original is conveyed (step S512), and the above operation is repeated.

  When reading of all the documents set on the tray 11 is completed, the CPU 41 determines whether or not a scan error has occurred from the information stored in the storage unit 45. When a scan miss occurs, the CPU 41 extracts information corresponding to the youngest page number from the information specifying the page where the scan miss has occurred. The CPU 41 sets a variable indicating a page to be rescanned as the initial value using the extracted page number. Subsequently, the CPU 41 notifies the operator of the occurrence of a scanning error and re-sets the document to the ADF, such as "Scanning failed. Set all documents in the ADF and press the start button." A prompt message is displayed (step S506). FIG. 11A shows an example of this message.

  In the present embodiment, both a document that has been successfully scanned and a document that has caused a scanning error are both discharged to the discharge tray 18. Since the scanned originals are discharged one by one to the discharge tray, the originals are arranged on the discharge tray 18 in the order in which they were set on the tray 11 when the scanning of all the originals is completed. The operator removes all originals from the discharge tray 18 and sets them on the ADF. When the operator presses the start button, the CPU 41 only carries the document without scanning up to the page where the scan error has occurred (step S507). For example, it is assumed that a document consisting of 50 pages in total is scanned using the ADF 10 and scan errors are detected on the third, fifteenth and forty-eighth pages. In this case, the storage unit 45 stores information indicating that a scan error has occurred on the 3rd, 15th, and 48th pages, and the initial value of the variable is set to 3. The CPU 41 outputs a signal to the ADF drive control circuit 43 and the CCD drive circuit 42 instructing that the document is transported without scanning until one page before the page designated by the variable (in this case, the second page). The scanner 100 only carries the document without scanning up to the second page. When the conveyance of the original to the second page is completed, the third-page original, which is a page where a scan error has occurred, appears on the resurface of the tray 11.

  In this state, the CPU 41 displays a message corresponding to the scan error on the display unit 46 as in the first embodiment (step S508). For example, when a scan error is detected by the fold determination, a message “Please extend the paper and rescan” is displayed on the display unit 46. The operator takes out the original on the resurface of the tray 11, corrects the folding of the original, and returns it to the tray 11. The operator presses the start button and inputs a rescan instruction. When the start button is pressed, the CPU 41 conveys one document and outputs an instruction to the ADF drive control circuit 43 and the CCD drive circuit 42 to scan the document. As a result, only the page where the scan error is detected is rescanned (step S509). The scanned document is checked for scan errors according to a scan error detection algorithm as usual. When no scan error is detected, the CPU 41 deletes information indicating that a scan error has occurred on the third page from the storage unit 45. Based on the information stored in the storage unit 45, the CPU 41 determines whether or not all the rescans of the document (page) that needs to be rescanned have been completed (step S510). Specifically, for example, the CPU 41 determines whether or not rescanning is completed by determining whether or not information indicating a page in which a scanning error has occurred is stored in the storage unit 45. If the rescan has not been completed, the CPU 41 sets the value of the variable to 15, which is the page where the next scan error is detected. Subsequently, similarly, the manuscript is preliminarily fed up to the 14th page and the 15th page is rescanned.

  If a scan error is detected again for the rescanned image, the CPU 41 does not delete the information indicating that a scan error has occurred on the third page, and the next scan error is detected using the variable value. The page is set to 15, and page skip and rescan continue. Or it is good also as a structure which discharges | emits to another discharge | emission tray similarly to 1st Embodiment, and promotes the rescan of the 3rd page again.

  In the above-described embodiment, a description has been given of a mode in which all pages of the document are set in the ADF 10 at the time of re-scanning, and a page in which no scan error is detected is skipped. It may be configured to be set and rescanned. In this case, after all the pages of the document have been scanned, the CPU 41 displays a message indicating the page on which the scanning error has occurred on the display unit 46. In the above example, the message “Scan failed. Please rescan pages 3, 15, and 48.” is displayed (FIG. 11B). The operator extracts the third, fifteenth and forty-eighth pages from the document and sets them on the ADF 10. When the operator presses the start button, these pages are rescanned.

  The electronic data of the scanned image is electronically sorted so that the page order is correct. That is, the CPU 41 temporarily stores the image data of each page in the storage unit 45 together with information specifying the page number. When scanning of all pages including rescanning is completed, the pages are rearranged in a regular order based on the information specifying the page numbers, and these images are combined into one file.

  According to the present embodiment, when a scanning error is detected in the scanning operation using the ADF, the operator simply puts all the documents on the tray, and the ADF is fed to the document (page) where the scanning error occurs, and Since information indicating the type of scan error is displayed, rescanning can be performed easily.

<2.4 Third Embodiment>
In the present embodiment, a description will be given of a mode in which the quality of a scanned image is not a good / bad choice, but is presented in numerical form. The scan quality evaluation unit 52 further has a function of calculating a numerical value representing the quality of the scanned image. That is, the storage unit 45 stores a scan quality index conversion table TB3 for converting the scan quality parameter output from the scan quality parameter extraction unit 51 into a numerical value representing the scan quality (hereinafter referred to as “scan quality index”). is doing.

  12A to 12C are diagrams showing examples of the scan quality index conversion table TB3. The storage unit 45 stores one or more such tables. The scan quality index conversion table TB3 includes an identifier for identifying the type of scan quality parameter ("skew angle" in FIG. 9A, "C / A" in (b), "tone value width" in (c). "), The value of the scan quality parameter, and the corresponding scan quality index are recorded.

  In the present embodiment, the paper transport method may use either the ADF 10 or the manual setting by the operator. When an operator inputs a scan instruction by a method such as pressing a start button, the scanner 100 reads an original and generates image data in the same manner as in the first or second embodiment. The scan quality parameter extraction unit 51 outputs a scan quality parameter according to any of the algorithms described above. Upon receiving the scan quality parameter from the scan quality parameter extraction unit 51, the scan quality evaluation unit 52 reads the scan quality index conversion table TB3 from the storage unit 45. The scan quality evaluation unit 52 converts the scan quality parameter into a scan quality index based on the scan quality index conversion table TB3. The scan quality evaluation unit 52 adds an identifier for specifying the scan quality evaluation algorithm to the obtained numerical value, and outputs it to the CPU 41.

  Examples of the method for converting the scan quality parameter into the scan quality index based on the scan quality index conversion table TB3 include the following. The first method is to obtain a scan quality index using the values described in the table as criteria. In the table illustrated in FIG. 12A, scan quality indexes are assigned to skew angles 0, 5, 10, and 45 °. In this example, when the skew angle θ is 0 ° ≦ θ <5 °, the scan quality index is 100, and when 5 ° ≦ θ <10 °, the scan quality index is 90. The quality index is converted. In the second method, values between values described in the table are obtained by interpolation. In the example of FIG. 12A, for example, when a skew angle of 4 ° is output from the scan quality parameter extraction unit 51 as a scan quality parameter, the scan quality evaluation unit 52 determines that the scan quality evaluation unit 52 is 4 ° from the table of FIG. The values (0 °, 5 °) listed in the table before and after and the scan quality index (100, 90) corresponding to each skew angle are extracted, and the scan quality corresponding to 4 ° is extracted from these values by linear interpolation. The index is calculated as 92.

  In place of the scan quality index conversion table TB3, a function for converting the scan quality parameter into a numerical value is stored in the storage unit 45, and the scan quality evaluation unit 52 substitutes the scan quality parameter for this function to scan quality. It is good also as composition which asks for an index.

  The scan quality evaluation unit 52 adds, to the scan quality index data obtained as described above, an identifier indicating that the data is the scan quality index and an identifier that identifies the algorithm that calculated the scan quality parameter. To the CPU 41. When the CPU 41 determines that the received data is a scan quality index, the CPU 41 reads out the non-defective product threshold value and the defective product threshold value from the storage unit 45. In the present embodiment, 90 and 50 are stored as the threshold values for the non-defective product and the defective product, respectively.

  First, the CPU 41 compares the scan quality index with a threshold value for non-defective products. If the scan quality index is greater than the non-defective threshold value, it is determined that no scan error has occurred (there is no problem with image quality), and the CPU 41 does not issue a rescan instruction. If the scan quality index is smaller than the non-defective product threshold, the CPU 41 compares the scan quality index with the defective product threshold. If the scan quality index is smaller than the threshold value for defective products, it is determined that a scan error has occurred, and the CPU 41 performs processing related to rescanning in the same manner as in the first or second embodiment. If the scan quality index is equal to or higher than the defective product threshold value, that is, if (defective product threshold value) ≦ (scan quality index) <(good product threshold value), the CPU 41 scans the scan quality index. A message is generated based on the data and other necessary data, and is output to the display unit 46. The display unit 46 displays a scan quality index and a screen for prompting rescan determination.

  FIG. 13 is a diagram illustrating an example of a screen that prompts the operator to determine rescanning. For example, if the scan quality index is 70, the scan quality index satisfies the condition of (defective product threshold value) ≦ (scan quality index) <(good product threshold value). Data, an identifier for specifying a scan error detection algorithm, and information for specifying a page in which a scan error is detected when scanning is performed using the ADF 10 as described in the second embodiment. Based on the (number of pages), a message as shown in FIG. 13 is generated.

  The operator can know from the displayed message the page where the suspected scan error was detected and the type of scan error. For example, the operator can check the image data of the page by operating the PC 200, etc. It can be determined whether a scan is necessary. The operator operates the operation unit 44 based on his / her own judgment and inputs an instruction on whether or not rescanning is necessary. The CPU 41 performs a rescan process in the same manner as in the first embodiment or the second embodiment described above according to the support input of the worker.

Note that the non-defective product / defective product threshold value stored in the storage unit 45 may be changed by an operator's operation input. Alternatively, every time the operator inputs the necessity of rescanning, the CPU 41 stores the necessity of rescanning and the scan quality index at that time in association with each other in the storage unit 45 and accumulates a certain amount or more of these data. Then, the CPU 41 may automatically determine whether or not rescanning is necessary based on statistical processing of these data. Moreover, it is good also as a structure which an operator can select automatic judgment / manual judgment.
Furthermore, in the above description, threshold values are provided for both non-defective products / defective products. However, only one of them may be provided with a threshold value, and the other judgment may be left to the operator.

  According to the present embodiment, the operator can compare an image suspected of having a scan error with a numerical value that objectively represents the degree of the scan error (image quality). On the other hand, it is possible to learn how much the scan quality index is necessary or not to be rescanned through the scanning operation, so that the skill of the operator can be improved.

<2.5 Fourth Embodiment>
In the present embodiment, the scanner 100 has a function of automatically correcting a scanned image when a scanning error occurs. The scanner 100 scans a document and generates image data corresponding to the scanned document. The scan quality parameter extraction unit 51 applies the above-described scan error detection algorithm to the generated image data, and outputs the scan quality parameter to the scan quality evaluation unit 52. The operation up to this point is the same as in the first to third embodiments described above.

  In the present embodiment, when the scan quality evaluation unit 52 detects a scan error, the scan quality evaluation unit 52 determines whether the scan error can be corrected. For this purpose, the scan quality evaluation unit 52 reads the correction possibility table TB4 from the storage unit 45. In the correctability table TB4, an identifier for specifying a scan error detection algorithm and a flag indicating whether or not a scan error can be corrected are recorded in association with each other. For example, the skew can be corrected by performing image processing of rotation processing, but the fold cannot be corrected by image processing and rescanning is essential. Accordingly, in the correctability table TB4, a flag “1” indicating that correction is possible is associated with skew, and a flag “0” indicating that correction is not possible is associated with breakage. If the scan quality evaluation unit 52 determines that the scan error cannot be corrected based on the correction possibility table TB4, the scan quality evaluation unit 52 indicates that a scan error has occurred, as in the first to third embodiments. A signal indicating information for specifying the detection algorithm is output to the CPU 41.

  When it is determined that the scan error can be corrected, the scan quality evaluation unit 52 outputs the image data and the scan quality parameter to the image processing unit 53. The image processing unit 53 corrects the image data based on the input image data and the scan quality parameter. For example, when a scan error is detected by skew determination, the scan quality parameter output from the scan quality evaluation unit 52 includes an identifier indicating that this parameter relates to skew determination and a skew angle. . The image processing unit 53 determines from the scan quality parameter that it is necessary to perform skew correction. The image processing unit 53 corrects the image data by performing known rotation processing, affine transformation processing, and the like based on the skew angle. The image processing unit 53 outputs the corrected data to the CPU 41 in the same manner as normal image data (no scan error was detected).

  In addition, for example, in the case of the protrusion determination, the image processing unit 53 corrects the image size data including the significant pixel value portion. In the case of the top / bottom determination, the image processing unit 53 performs a rotation process of an angle that maximizes the number of recognized characters. Alternatively, in the case of tone / gamma / contrast determination, the image processing unit 53 expands the gradation value width to an appropriate value and corrects the gradation value with reference to the lookup table stored in the storage unit 45. Do.

  For example, in the tone / gamma / contrast determination, when the gradation value width is equal to or smaller than the threshold value stored in the storage unit 45, it may be determined that automatic correction is impossible and a re-scan is promoted. Good. That is, the image processing unit 53 outputs to the CPU 41 a signal indicating that a scan error has occurred and information specifying the scan error detection algorithm.

  According to the present embodiment, a scan error that can be automatically corrected by the scanner is automatically corrected, and only the scan error is notified to the operator that automatic correction is impossible. Therefore, it is possible to further improve the efficiency of the scan operation.

<2.6 Modification>
In each of the above-described embodiments, the means for notifying the operator of the occurrence of a scan error is not limited to a mode in which a message is displayed on the display unit 46, and may be configured to issue a message by voice from a speaker or the like.
In each of the above-described embodiments, the scanner 100 has been described as an example of the image reading apparatus. However, the image reading apparatus is not limited to the scanner, and may be a copier, a multifunction machine, or the like.

DESCRIPTION OF SYMBOLS 1 ... Image reading system, 10 ... ADF, 11 ... Tray, 12 ... Pull-in roller, 13 ... Conveyance roller, 14 ... Ejection roller, 15 ... Ejection roller, 16 ... Conveyance roller, 17 ... Back platen, 18 ... Ejection tray, 19 DESCRIPTION OF SYMBOLS ... Discharge tray, 20 ... Platen glass, 31 ... 1st CRG, 32 ... 2nd CRG, 33 ... Lens, 34 ... CCD sensor, 41 ... CPU, 42 ... CCD drive circuit, 43 ... ADF drive control circuit, 44 ... Operation part, 45: Storage unit, 46: Display unit, 47 ... I / F, 48 ... Bus, 49 ... A / D conversion circuit, 50 ... Image processing circuit, 51 ... Scan quality parameter extraction unit, 52 ... Scan quality evaluation unit, 53 ... Image processing unit, 60 ... Original, 61 ... Book, 100 ... Scanner, 200 ... PC, 210 ... Display, 300 ... Cable

Claims (3)

Reading means for reading a document and generating a scanned image corresponding to the document;
Scan quality parameter extraction means for extracting a scan quality parameter indicating the quality of the scan image generated by the reading means based on a scan error detection algorithm;
A scan quality evaluation means for determining the presence or absence of a scan error based on the scan error detection algorithm and the scan quality parameter;
Notification means for making a notification based on the determination result of the scan quality evaluation means,
The scan error detection algorithm is
Dividing the scanned image into a plurality of partial images;
Calculating an estimated number of characters for each of the plurality of partial images;
Performing character recognition processing on each of the plurality of partial images;
Calculating a character recognition rate in each of the plurality of partial images based on the estimated number of characters and the number of characters recognized by the character recognition process,
It is determined that a scan error has occurred when a character recognition rate in at least one of the partial images is lower than a threshold value of the character recognition rate. Reading means for reading a document and generating a scanned image corresponding to the document;
Scan quality parameter extraction means for extracting a scan quality parameter indicating the quality of the scan image generated by the reading means based on a scan error detection algorithm;
A scan quality evaluation means for determining the presence or absence of a scan error based on the scan error detection algorithm and the scan quality parameter;
Notification means for making a notification based on the determination result of the scan quality evaluation means,
The scan error detection algorithm is
Dividing the scanned image into a plurality of partial images;
Calculating an estimated number of characters for each of the plurality of partial images;
Performing character recognition processing on each of the plurality of partial images;
Calculating a character recognition rate in each of the plurality of partial images based on the estimated number of characters and the number of characters recognized by the character recognition process,
The character recognition rate in the first partial image that is a partial region of the plurality of partial images, and the second partial image that is a region excluding the first partial image of the plurality of partial images. A scan error occurs when the difference between the character recognition rate or the difference between the character recognition rate in the first partial image and the average of the character recognition rates in the first and second partial images is equal to or greater than a threshold value. An image reading apparatus characterized by determining that it has occurred. Reading means for reading a document and generating a scanned image corresponding to the document;
Scan quality parameter extraction means for extracting a scan quality parameter indicating the quality of the scan image generated by the reading means based on a scan error detection algorithm;
A scan quality evaluation means for determining the presence or absence of a scan error based on the scan error detection algorithm and the scan quality parameter;
Notification means for making a notification based on the determination result of the scan quality evaluation means,
The scan error detection algorithm is
Rewriting scan parameters so that scanning is performed at a size larger than the document size input by the operator before scanning;
Excluding the document size image from the scanned image, extracting an image outside the document size area, and determining whether a significant pixel exists in the image outside the area, and
An image reading apparatus, wherein when it is determined that a significant pixel exists in an image outside the region, it is determined that a scan error has occurred.

Images