JP5780064B2 - Image reading device - Google Patents

Description

  The technique disclosed in this specification relates to a technique used in an image reading apparatus that reads a document.

2. Description of the Related Art Conventionally, in an image reading apparatus that reads a document placed on a document table or a document to be conveyed, the document may be inclined and set on the document table, or the document may be inclined by conveying the document. In this case, since the reading unit such as the CIS reads in a state where the document is inclined, an image reading device (for example, Patent Document 1) that corrects the inclination of the document with respect to the read image read by the reading unit. Are known.
In the image reading apparatus of Patent Document 1, a point that is a candidate for the edge of a read image of a read original is detected, the point is approximated by a straight line using the least square method, and the inclination of the original with respect to the reading unit is obtained from the straight line. A technique for correcting the inclination of a document with respect to a read image read by a reading unit is disclosed.

JP 2009-164807 A

  However, the edge of the document is not always linear. For example, in a document that has been cut and shaped by the user, part of the edge protrudes outward from the other part, or conversely, part of the edge protrudes inward from the other part. The edges of the surface may be uneven. In such a document, even if a point that is a candidate for the edge of the scanned image of the document is detected and the point is linearly approximated using the least square method, if the edge of the document is not linear in the first place, The slope cannot be determined. However, the conventional image reading apparatus, regardless of whether or not the end of such a document is linear, approximates the point using a least square method, and the document from the straight line to the reading unit. I was looking for the slope.

  The present invention discloses a technique for correcting an inclination of a document with respect to a reading unit from a read image read by a reading unit included in the image reading device.

  The image reading apparatus disclosed in this specification includes a reading unit that reads an image in a reading range including a document image range obtained by reading a document in a main scanning direction, and the document image based on an image acquired by the reading unit. An edge detection unit that detects a plurality of edge points on one side of the range, an interval acquisition unit that acquires variation in the position of the edge point in a direction orthogonal or parallel to the main scanning direction, and the variation in the position as a reference An inclination correction unit that detects an inclination of the document with respect to the reading unit from a plurality of edge points detected by the edge detection unit and corrects an inclination with respect to an image in the document image range according to the inclination; When the position variation is larger than the reference interval, the edge detection unit detects a plurality of edge points on the side different from the previously detected side, and the interval detection is performed. And a control unit for obtaining a variation in the position of the edge points of the different sides in section, a.

  The image reading apparatus further includes a transport unit that transports a document along a transport path, and the reading unit is positioned on the transport path and reads a document transported by the transport unit. The edge detection unit detects a plurality of edge points with respect to one side of the leading edge of the document image range, and the control unit causes the edge detection unit to detect when the variation in the position of the edge point is larger than the reference interval. A plurality of edge points on one side of the rear end of the document image range may be detected.

  Further, in the above image reading apparatus, the control unit corrects the tilt when a variation in the positions of a plurality of edge points detected with respect to one side of the rear end of the document image range is larger than the reference interval. It is also possible to adopt a configuration in which the inclination is not corrected for the image in the original image range.

  In the image reading apparatus, the edge detection unit may be configured to detect the plurality of points from an image read at a central portion excluding both ends in the main scanning direction of the reading range.

  In the image reading apparatus, the control unit may correct the inclination of the image in the document image range to the inclination correction unit when the variation in position is larger than the second reference interval allowed by the inclination correction unit. It is good also as a structure which is not made to do.

  In the image reading apparatus disclosed in this specification, a plurality of edge points are detected with respect to one side of the document image range, and the variation in the position of the edge points is more than the reference interval in a direction perpendicular or parallel to the main scanning direction. If there is a large variation, an edge point is detected for another side of the image in the document image range. According to this image reading apparatus, when one side of the document image range varies more than the reference interval, the edge point is detected with respect to the other side, and the inclination can be detected. It is possible to correct the inclination of the document without using the inclination detected from the side that greatly varies.

Schematic sectional view of an image reading device Block diagram of image reader 6 is a flowchart illustrating a reading process according to the first embodiment. The flowchart which shows the inclination correction process of Embodiment 1. The flowchart which shows the edge detection process of Embodiment 1. The flowchart which shows the edge detection process of Embodiment 1. Flow chart showing tilt detection processing The figure which shows the edge image EG of Embodiment 1. The flowchart which shows the inclination correction process of Embodiment 2. The flowchart which shows the inclination correction process of Embodiment 2. The flowchart which shows the edge detection process of Embodiment 2. The flowchart which shows the edge detection process of Embodiment 2. The figure which shows the edge image EG of Embodiment 2.

<Embodiment 1>
The first embodiment will be described with reference to FIGS. 1 to 8.

1. FIG. 1 is a schematic cross-sectional view of an image reading apparatus 1 of the present embodiment. The image reading apparatus 1 includes a paper feed tray 2, a main body unit 3, and a paper discharge tray 4. The image reading apparatus 1 conveys a plurality of originals G placed on the paper feed tray 2 by a user to the paper discharge tray 4 and reads the original G being conveyed using the CIS 24 included in the main body 3. It is a scanner.

  The main body 3 is provided with a conveyance path 22 that connects the paper feed tray 2 and the paper discharge tray 4. Around the conveyance path 22, a paper feed roller 20, a separation pad 21, a conveyance roller 23, and the like are provided. , CIS 24, front sensor (hereinafter referred to as F sensor) 25, and rear sensor (hereinafter referred to as R sensor) 26. The CIS 24 is an example of a reading unit.

  The paper feed roller 20 is in contact with the original G placed on the paper feed tray 2, and draws a plurality of originals G placed on the paper feed tray 2 into the main body 3 by frictional force. . The separation pad 21 separates a plurality of documents G into a single document G by a frictional force. The plurality of documents G placed on the sheet feed tray 2 are separated into a single document G and drawn into the main body 3.

  The conveyance roller 23 is driven by a motor M (see FIG. 2), and conveys the document G drawn into the main body 3 onto the conveyance path 22. The CIS 24 is located on the transport path 22 in a posture in which the main scanning direction D1 as the reading direction is orthogonal to the transport direction D2 along the transport path 22, and reads the document G transported in the transport path 22.

  Further, the transport roller 23 sends the document G transported on the transport path 22 to the paper discharge tray 4. The original G discharged to the paper discharge tray 4 by the transport roller 23 is stacked on the paper discharge tray 4. That is, the paper feed roller 20 and the transport roller 23 form a transport unit 27 that transports a plurality of documents G placed on the paper feed tray 2 along the transport path 22.

  The F sensor 25 is disposed in the paper feed tray 2 and is set to be turned on when the original G is placed on the paper feed tray 2 and turned off when the original G is not placed on the paper feed tray 2. Has been. The R sensor 25 is disposed upstream of the CIS 24 on the transport path 22 and is turned on when the document G passes through the transport path 22 and turned off when the document G does not pass through the transport path 22. Is set to In addition, the main body 3 includes a power switch and various setting buttons. The input unit 5 (see FIG. 2) that receives an operation command from the user and the status of the image reading apparatus 1 includes an LED. The display unit 6 (see FIG. 2) and the output device 9 are included. The output device 9 includes, for example, a USB memory connection unit and the like, and is used to take out an image read from the original G using the CIS 24 to the outside of the image reading device 1.

2. FIG. 2 is a block diagram schematically showing the electrical configuration of the image reading apparatus 1. As shown in FIG. 2, the image reading apparatus 1 includes an ASIC (application-specific integrated circuit) 10 that controls each unit of the image reading apparatus 1. The ASIC 10 includes a central processing unit (hereinafter referred to as CPU) 11 and a ROM 12.
RAM 13, device control unit 14, and analog front end (hereinafter referred to as AFE) 1
5, a drive circuit 16, and an image processing circuit 17, to which an F sensor 25, an R sensor 26, and the like are connected via a bus 18 .

  Various programs for controlling the operation of the image reading apparatus 1 are stored in the ROM 12. The CPU 11 functions as an interval acquisition unit 31, an inclination correction unit 32, a control unit 33, and the like according to the program read from the ROM 26. Then, each part is controlled. The device control unit 14 is connected to the CIS 24 and transmits a signal for controlling reading based on a command from the CPU 11 to the CIS 24. The CIS 24 reads the document G over the reading range H (see FIG. 8) based on a signal from the device control unit 14, and outputs the read data to the AFE 15.

  The AFE 15 is connected to the CIS 24, and converts read data, which is an analog signal output from the CIS 24, into gradation data, which is a digital signal, based on a command from the CPU 11. The read data and the gradation data are examples of images. The AFE 15 converts the read data output from the CIS 24 into a scan image SG intended to be stored in the image reading apparatus 1 or taken out of the image reading apparatus 1, and reads the original G from the read data. The original image range GH (see FIG. 8) is converted into an edge image EG for specifying. The edge image EG uses a predetermined reference luminance, a first gradation value indicating that the edge image EG has a luminance higher than the reference luminance, and a second gradation indicating that the edge image EG has a luminance lower than the reference luminance. The value is binarized. The scan image SG and the edge image EG are stored in the RAM 13 via the bus 18. The image processing circuit 17 specifies the document image range GH based on the edge image EG stored in the RAM 13 and generates a scan image SG of the document image range GH.

  The drive circuit 16 is connected to the motor M, and transmits a pulse signal to the motor M based on a command from the CPU 11. When the motor M rotates in accordance with the pulse signal input from the drive circuit 16, the conveyance unit 27 is driven in accordance with this, and the original G is step-conveyed along the conveyance path 22. That is, when conveying the document G, the CPU 11 transmits a pulse signal to the motor M via the drive circuit 16, and the conveying unit 27 accordingly sets the specified distance L to the number of pulses transmitted from the pulse signal. The original G is conveyed by the multiplied distance. Hereinafter, the number of pulses of the pulse signal that the drive circuit 16 transmits to the transport unit 27 is referred to as a step number.

3. Reading Processing Next, processing in the CPU 11 when reading the original G using the CIS 24 will be described with reference to FIGS.

  FIG. 3 shows a flowchart of processing executed by the CPU 11 according to a predetermined program. The CPU 11 starts processing when it is confirmed that the document G is placed on the paper feed tray 2 using the F sensor 25 and an instruction to read the document G is input via the input unit 5.

(Basic processing)
When starting the processing, the CPU 11 drives the transport unit 27 using the drive circuit 16 and starts transporting the document G (S2). The CPU 11 detects the position of the document G being conveyed using the R sensor 26 (S4: NO). When the R sensor 26 is turned on (S4: YES), reading by the CIS 24 is started after the R sensor 26 is turned on. It waits until the first step number ST1 until it is counted (S6: NO). The first step number ST1 is determined based on the distance between the R sensor 26 and the CIS 24 in the transport direction D2 and the reading range H.

  When the first step number ST1 is counted after being turned on by the R sensor 26 (S6: YES), the CPU 11 starts reading the document G using the CIS 24 (S8). The CPU 11 waits for the R sensor 26 to turn off (S10: NO). When the R sensor 26 is turned off (S10: YES), the CPU 11 turns off the R sensor 26 and ends reading by the CIS 24. The second step number ST2 is counted (S12) and the R sensor 26 is turned on again by the next original G (S22). Similarly to the first step number ST1, the second step number ST2 is determined based on the distance between the R sensor 26 and the CIS 24 in the transport direction D2 and the reading range H.

  When the second step number ST2 is counted before the R sensor 26 is turned on again (S12: YES, S14: NO), the CPU 11 detects that the reading of the document G over the reading range H has been completed. The CPU 11 ends the reading of the original G (S22) and executes an inclination correction process described later (S24). In other words, in the present embodiment, the CPU 11 executes the tilt correction process using the edge image EG corresponding to the entire reading range H after completing the reading of the original G over the entire reading range H. After executing the inclination correction process, the CPU 11 checks whether or not the F sensor 25 is turned on (S26). If the F sensor 25 is turned off (S26: NO), the CPU 11 remains unread on the paper feed tray 2. It is determined that there is no original G, the conveyance of the original G is stopped, and the reading process is terminated.

  Further, when the F sensor 25 is on (S26: YES), the CPU 11 executes processing for the document G to be conveyed next. The CPU 11 detects the position of the next original G to be conveyed using the R sensor 26 (S28: NO), and repeats the processing from S6 when the R sensor 26 is turned on (S28: YES).

  On the other hand, if the R sensor 26 is turned on again before the second step number ST2 is counted (S12: NO, S14: YES), the CPU 11 continues to count the second step number ST2 (S16: NO). When the second step number ST2 is counted (S16: YES), the CPU 11 ends the reading of the original G (S18) and executes an inclination correction process described later (S20). After executing the inclination correction process, the CPU 11 returns to S6 and executes the process on the next original G to be conveyed.

(Tilt detection processing)
The tilt correction process will be described with reference to FIG. In the inclination correction process, the CPU 11 first executes an edge point detection process (S30). When the CPU 11 reads the document G while transporting it on the transport path 22, even if the transported document G is inclined with respect to the transport direction D <b> 2, the CPU 11 reads the document G without causing image loss. A reading range H wider than G is set. Therefore, as shown in FIG. 8, the image of the original image range GH obtained by reading the original G is included in the edge image EG obtained by reading the entire reading range H. In the edge point detection process, the CPU 11 uses the image processing circuit 17 to detect, from the edge image EG stored in the RAM 13, an edge point P that is a point on the boundary line of the document image range GH from which the document G is read. Execute.

  The edge point detection process will be described with reference to FIGS. When starting the edge point detection process, the CPU 11 selects an edge image EG of the central portion IB excluding both end portions OB in the main scanning direction D1 from the edge image EG corresponding to the entire reading range H (S42). Since the main scanning direction D1 of the edge image EG is equal to the main scanning direction D1 of the CIS 24, the central portion IB of the edge image EG is an edge image read at the central portion excluding both ends in the main scanning direction D1 of the CIS 24. It can be said that it is EG. In the following description of the edge point detection process, a process of detecting the edge point P is performed on the selected edge image EG of the central portion IB.

  Next, the CPU 11 sets various parameters for detecting the edge point P. The CPU 11 has parameters of Xstart, Ystart, dX, dY, Ypeakmax, Ypeakmin, and n as parameters. Here, as shown in FIG. 8, (X, Y) represents a coordinate system in the edge image EG, X represents a coordinate axis parallel to the main scanning direction D1, and Y represents a coordinate axis parallel to the transport direction D2. Yes. In the present embodiment, as the origin of the coordinate system (X, Y), the point where the X coordinate and the Y coordinate are the smallest among the selected edge images EG of the central portion IB is used as the origin. In this case, the X coordinate is The X coordinate of the maximum point is Xmax, and the Y coordinate of the point where the Y coordinate is maximum is Ymax.

  In the present embodiment, when the CPU 11 detects the edge point P, the CPU 11 advances the detection of the edge point P in the transport direction D2 with respect to a plurality of set X coordinates, and each Y at a position that satisfies the condition of the edge point P described later. Detect coordinates. Then, the CPU 11 functioning as the interval acquisition unit 31 acquires variations in the positions of the plurality of edge points P corresponding to the plurality of X coordinates as variations in the Y coordinate. Xstart indicates the smallest X coordinate among the plurality of X coordinates, and dX indicates an interval between the plurality of X coordinates. Ystart indicates a Y coordinate at which detection starts in the transport direction D2, and dY indicates a detection resolution when the detection proceeds in the transport direction D2. Ypeakmax indicates the Y coordinate of the edge point P having the largest Y coordinate among the detected edge points P, and Ypeakmin indicates the edge point P having the smallest Y coordinate among the detected edge points P. The Y coordinate is shown.

  When detecting the edge point P from the edge image EG, the CPU 11 first detects a plurality of edge points P (hereinafter referred to as the leading edge point P) from one side of the leading edge of the edge image EG. The CPU 11 sets parameters for detecting a plurality of leading edge points P. For example, the CPU 11 sets Xstart = dX = 8, Ystart = 0, and dY = 1 (S44). The CPU 11 sets Ypeakmax = 0 and Ypeakmin = Ymax (S46), and resets the peak value stored in the previous edge point detection process.

  Next, the CPU 11 starts detecting a plurality of edge points P using the set parameters. The CPU 11 has a detection point E (Xe, Ye) used for detection of the edge point P, and sets Xe = Xstart (S48). Further, n indicating the number of detected leading edge points P is set to n = 0 (S50).

  Next, the CPU 11 increases n by “1” (S52), sets Ye = Ystart (= 0) (S54), and checks whether the detection point E is the edge point P (S56). Specifically, when the gradation value of the detection point (Xe, Ye) is the first gradation value, it is determined that the detection point (Xe, Ye) is the edge point P. On the other hand, when the gradation value of the detection point E is the second gradation value, it is determined that the detection point E is not the edge point P.

  When the CPU 11 detects that the detection point E is not the edge point P (S56: NO), it increases Ye by dY (S58), and the increased Ye is greater than Ymax, or the increased Ye is less than “0”. It is confirmed whether or not any of the following conditions is satisfied (S60). When the increased Ye does not satisfy any of the above conditions (S60: NO), the CPU 11 repeats the processing from S56 and advances the detection of the edge image EG in the transport direction D2.

  If the increased Ye satisfies any of the above conditions (S60: YES), the CPU 11 increases Xe by dX (S62), and confirms whether the increased Xe is greater than Xmax (S64). ). When the increased Xe is equal to or less than Xmax (S64: NO), the CPU 11 repeats the processing from S54. Further, when the increased Xe is larger than Xmax (S64: YES), the CPU 11 determines that the detection of the edge point P has failed.

  On the other hand, when the CPU 11 detects that the detection point E is the edge point P (S56: NO), the CPU 11 stores the detection point E in the RAM 13 as the edge point Pn which is the nth edge point P (S66). The CPU 11 further compares Ye of the detection point E stored as the edge point Pn with Ypeakmax and Ypeakmin (S68, S72). When Ye is larger than Ypeakmax (S68: YES), the CPU 11 resets the value of Ypeakmax to the value of Ye (S70), and when Ye is equal to or less than Ypeakmax (S68: NO), the value of Ypeakmax is set to the current value. Keep the value. Further, when Ye is smaller than Ypeakmin (S72: YES), the CPU 11 resets the value of Ypeakmin to the value of Ye (S74), and when Ye is equal to or greater than Ypeakmin (S72: NO), the value of Ypeakmin is set. Keep the current value.

  Next, the CPU 11 subtracts Ypeakmin from Ypeakmin to calculate a difference value ΔY indicating the variation in the Y coordinate of the edge point P. As described above, the difference value ΔY can be said to indicate variations in the positions of the plurality of edge points P. Then, the CPU 11 compares the difference value ΔY with a reference interval K stored in advance (S76). When the difference value ΔY is equal to or smaller than the reference interval K (S76: NO), the CPU 11 checks whether n is equal to or larger than a necessary number nK, which is the number of edge points necessary for an inclination detection process described later (S78). The CPU 11 determines that the edge point P has been successfully detected when n is equal to or greater than the necessary number nK (S78: YES).

  When n is smaller than the required number nK (S78: NO), the CPU 11 increases Xe by dX (S80), and checks whether the increased Xe is larger than Xmax (S82). When the increased Xe is equal to or less than Xmax (S82: NO), the CPU 11 repeats the processing from S52. Further, when the increased Xe is larger than Xmax (S82: YES), the CPU 11 determines that the detection of the edge point P has failed.

  On the other hand, when the difference value ΔY is larger than the reference interval K (S76: YES), the CPU 11 functioning as the control unit 33 confirms whether the edge point P currently detected is the leading edge point P (S84). . The CPU 11 confirms whether the edge point P currently detected from the currently set parameters such as Ystart is the tip edge point P. When the currently detected edge point P is the leading edge point P (S84: YES), the CPU 11 ends the detection of the leading edge point P, and next, with respect to one side of the trailing edge of the edge image EG A plurality of edge points P (hereinafter, rear end edge points P) are detected.

  The CPU 11 sets parameters for detecting a plurality of trailing edge points P. For example, the CPU 11 sets Xstart = dX = 8, Ystart = Ymax, dY = −1 (S86), returns to S46, calculates the difference value ΔY for the trailing edge point P, and compares it with the reference interval K. (S76) and the like are repeated. Then, when the difference value ΔY calculated also at the trailing edge point P is larger than the reference interval K (S76: YES, S84: NO), the CPU 11 determines that the detection of the edge point P has failed.

  As illustrated in FIG. 4, when the CPU 11 functioning as the control unit 33 fails to detect the edge point P in the edge point detection process (S32: NO), an inclination detection process (S34) described later and a scan image rotation process are performed. Without executing (S38), the process returns to S6 or S26 of FIG. On the other hand, when the edge point P is successfully detected in the edge point detection process (S32: YES), the CPU 11 executes the inclination detection process (S34).

  The tilt detection process will be described with reference to FIG. When starting the tilt detection process, the CPU 11 linearly approximates the obtained edge point P, and calculates a tilt θ that is an angle of the straight line with respect to the main scanning direction D1 (S92). The CPU 11 confirms whether or not the inclination θ is equal to or smaller than a correctable limit θL that is an upper limit value that allows scanning image rotation processing described later (S94). Since the inclination θ is determined based on the variation in the position of the edge point P, the correctable limit θL is used to classify whether or not the variation in the position of the edge point P is allowed in the scan image rotation process. It can be said to be an example of two reference intervals.

  When the inclination θ is larger than the correctable limit θL (S94: NO), the CPU 11 determines that the detection of the inclination θ has failed. On the other hand, when the inclination θ is equal to or less than the correctable limit θL (S94: YES), the inclination correction amount θH is calculated from the inclination θ (S96), and it is determined that the inclination θ is successfully detected. When the leading edge point P and the trailing edge point P are detected as the edge points P as in the present embodiment, the inclination correction amount θH matches the inclination θ. Further, when the edge point P is detected in the main scanning direction D1 and the edge point P is detected, the inclination correction amount θH is a value obtained by subtracting 90 ° from the inclination θ, or subtracting the inclination θ from 90 °. .

  As shown in FIG. 4, when the CPU 11 functioning as the control unit 33 fails to detect the inclination θ in the inclination detection process (S36: NO), the CPU 11 does not execute the scan image rotation process (S38) described later. The process returns to S6 or S26. On the other hand, when the CPU 11 functioning as the inclination correction unit 32 succeeds in detecting the inclination θ in the inclination detection process (S36: YES), the CPU 11 executes a scan image rotation process (S38). The CPU 11 rotates the scan image SG in the document image range GH stored in the RAM 13 by the tilt correction amount θH calculated in the tilt detection process. After completing the scan image rotation process, the CPU 11 returns to the process of S6 or S26 of FIG.

4). Advantages of the present embodiment (1) In the image reading apparatus 1, a plurality of edge points P are detected with respect to one side of the leading edge of the document image range GH of the edge image EG, and a difference value indicating variation in the position of the edge point P is detected. When ΔY is larger than the reference interval K, the edge point P is detected with respect to one side of the rear end of the document image range GH. Therefore, the inclination θ is calculated from one side of the leading edge of the document image range GH of the edge image EG in which the difference value ΔY varies more than the reference interval K, and the document image range GH is corrected without correcting the inclination of the document G. It is possible to suppress the deterioration of the specified specific accuracy.

(2) In the image reading apparatus 1, when the difference value ΔY with respect to one side of the leading edge and the trailing edge of the document image range GH of the edge image EG varies more than the reference interval K, the image reading apparatus 1 is inclined with respect to the document image range GH. The detection process and the scan image rotation process are not executed. Therefore, it is possible to prevent the original image range GH of the scan image SG from being rotated inaccurately, and it is possible to suppress the deterioration of the specific accuracy for specifying the original image range GH.

(3) In the image reading apparatus 1, as shown in FIG. 8, the edge image EG of the central portion IB excluding both end portions OB in the main scanning direction D1 is selected from the edge images EG corresponding to the entire reading range H. Then, an edge point P is detected for the edge image EG of the central portion IB. Normally, when the original G is read by the CIS 24, the original G is often arranged at the approximate center of the reading range H of the CIS 24. Therefore, the edge image EG of the central portion IB, which is the edge image EG read at the central portion of the CIS 24, includes one side of the document image range GH, but includes other sides. Often not. According to the image reading apparatus 1, when a plurality of edge points P are detected for one side of the leading edge and the trailing edge of the document image range GH, the edge point P for the other side is erroneously detected. And the occurrence of problems such as deterioration in tilt detection accuracy due to erroneous detection is suppressed.

(4) In the image reading apparatus 1, the difference value ΔY with respect to one side of the leading edge and the trailing edge of the document image range GH of the edge image EG varies more than the reference interval K, and the CPU 11 functioning as the inclination correcting unit 32 accurately tilts. When the correction cannot be performed, the scan image rotation process is not performed on the document image range GH of the scan image SG. Therefore, it is possible to prevent the original image range GH of the scan image SG from being rotated inaccurately, and it is possible to suppress the deterioration of the specific accuracy for specifying the original image range GH.

<Embodiment 2>
The second embodiment will be described with reference to FIGS. 9 to 13. In the present embodiment, the CPU 11 does not store the edge image EG corresponding to the entire reading range H in the RAM 13 when reading over the entire reading range H of the document G. Different from device 1. In the present embodiment, when the CPU 11 performs reading over the entire reading range H of the document G, as shown in FIG. 13, the CPU 11 uses the AFE 15 to select a range that includes one side of the leading edge of the document image range GH. The leading edge image EG1 is stored in the RAM 13, the trailing edge image EG2 in the range including one side of the trailing edge of the document image range GH is stored in the RAM 13, and the inclination correction process is executed using these edge images EG1 and EG2. To do. In the following description, the same description as that of the first embodiment will not be repeated.

1. Reading process (basic process)
As shown in FIGS. 9 and 10, the CPU 11 starts the process, and when the R sensor 26 is turned on (S4: YES), the CPU 11 clears the inclination correction amount θH (S102). Then, when the first step number is counted after the R sensor 26 is turned on (S6: YES), the CPU 11 starts reading the original G using the CIS 24 (S104). At this time, as described above, the CPU 11 uses the AFE 15 to store the scan image SG in the RAM 13 over the entire reading range H of the document G, while the edge image EG includes the leading edge image EG1 and the trailing edge. Only the edge image EG2 is stored in the RAM 13.

  The CPU 11 checks whether or not the third step number ST3 corresponding to the leading edge image EG1 has been counted since the reading of the CIS 24 was started (S106). Here, the third step number ST3 is determined based on the reading range H and the document size. When the third step number ST3 is counted from the start of reading of the CIS 24 (S106: YES), the CPU 11 performs an inclination correction process (S108) shown in FIG. 4 on the read leading edge image EG1. In the inclination correction processing for the leading edge image EG1, the CPU 11 functioning as the inclination correction unit 32 rotates the scan image SG of the document image range GH read so far by the calculated inclination correction amount θH, and then reads it. A process of sequentially rotating the scanned image SG in the original image range GH is executed.

  Further, when the third step number ST3 has not been counted since the reading of the CIS 24 is started (S106: NO), the CPU 11 executes the process of S10 without performing the inclination correction process. In the process of S10, the CPU 11 checks whether or not the R sensor 26 is turned off (S10), and when the R sensor 26 is turned off (S10: YES), the second step number ST2 is counted ( While waiting for S12), it waits for the R sensor 26 to turn on again (S22). On the other hand, when the R sensor 26 is not turned off (S10: NO), the CPU 11 repeats the processing from S106. When the CPU 11 repeats the process from S106, the third step number ST3 is counted in the process so far (S106: YES), and the inclination correction process (S108) is performed on the leading edge image EG1. In this case, the inclination correction process (S108) is not executed again.

  Next, when the second step number ST2 is counted before the R sensor 26 is turned on again (S12: YES, S14: NO), the CPU 11 reads the scanned image SG over the entire reading range H of the document G. And that the reading of the trailing edge image EG2 is completed. The CPU 11 finishes reading the document G (S22), and checks whether or not the inclination correction amount θH has been calculated (S116).

  When the inclination correction amount θH is calculated in the inclination correction processing for the leading edge image EG1, and the inclination correction amount θH has already been calculated (S116: YES), the CPU 11 advances the processing to S26. On the other hand, when the inclination correction amount θH has not been calculated in the inclination correction processing for the leading edge image EG1, and the inclination correction amount θH has not yet been calculated (S116: NO), the shift processing necessary for the trailing edge image EG2 is performed. This is executed (S118), and the inclination correction process (S120) shown in FIG. 4 is performed on the read rear edge image EG2. Here, the shift processing is, for example, when the rear end edge image EG2 is stored as a ring buffer so as to share the storage area in the RAM 13 with the front end edge image EG1, and the stored rear end edge image EG2 is read. The process of rearranging in the order shown.

  On the other hand, if the R sensor 26 is turned on again before the second step number ST2 is counted (S12: NO, S14: YES), the CPU 11 continues to count the second step number ST2 (S16: NO). When the second step number ST2 is counted (S16: YES), the CPU 11 ends the reading of the original G (S18), and confirms whether or not the inclination correction amount θH has been calculated as in S116 ( S110).

(Tilt detection processing)
The tilt detection process of this embodiment is the same as the tilt detection process of the first embodiment, but the edge point detection process (S30) to be executed is different. The edge point detection processing of this embodiment will be described with reference to FIGS. When starting the edge point detection process, the CPU 11 selects the leading edge image EG1 or the trailing edge image EG2 of the central portion IB as shown in FIG. 13 (S42).

  Next, the CPU 11 checks whether the edge point P currently detected is the leading edge point P or the trailing edge point P (S122). The CPU 11 identifies whether the currently processed edge image EG is the leading edge image EG1 or the trailing edge image EG2 depending on whether or not the shift processing has been performed prior to the processing, and currently detects it. It is confirmed whether the edge point P is the leading edge point P or the trailing edge point P. When the edge point P currently detected is the leading edge point P (S122: leading edge), the CPU 11 sets a plurality of detection parameters for detecting a plurality of trailing edge points P (S44). Further, when the edge point P currently detected is the rear end edge point P (S122: rear end), the CPU 11 sets a plurality of detection parameters for detecting a plurality of rear end edge points P. (S86).

  Then, the CPU 11 calculates a difference value ΔY with respect to the leading edge point P or the trailing edge point P and compares it with the reference interval K (S76). When the difference value ΔY is larger than the reference interval K (S76: YES), the CPU 11 functioning as the control unit 33 does not check whether the edge point P currently detected is the tip edge point P, and the edge point It is determined that detection of P has failed.

2. Advantages of the present embodiment (1) In the image reading apparatus 1, when executing the tilt correction process, only the leading edge image EG1 or the trailing edge image EG2 is used, and the edge image EG corresponding to the entire reading range H is required. And not. Therefore, the storage capacity of the RAM 13 for storing the edge image EG can be reduced, and the manufacturing cost of the apparatus can be reduced.

(2) Further, in the image reading device 1, when the inclination correction amount θH is calculated in the leading edge image EG1, it is not necessary to read the trailing edge image EG2. Therefore, the processing load of the image reading apparatus 1 in the reading process can be reduced.

<Other embodiments>
The present invention is not limited to the embodiments described with reference to the above description and the drawings, and for example, the following various aspects are also included in the technical scope of the present invention.
(1) Although the above embodiment has been described using the image reading apparatus 1, the present invention is not limited to this. For example, the image forming apparatus may be a multifunction machine having a printer function for forming an image and at least one function such as a copy function and a facsimile function together with a scanner function.

(2) In the above embodiment, the image reading apparatus 1 has one ASIC 10 and will be described using an example in which reading processing is executed by one CPU 11 that functions as the interval acquisition unit 31, the inclination correction unit 32, and the control unit 33. However, the present invention is not limited to this. For example, the reading process may be executed by a plurality of ASICs, CPUs, and the like.

(3) In the above embodiment, when the edge point P is detected, the example in which the detection of the edge point P is advanced in the transport direction D2 in the plurality of set X coordinates has been described. Not limited to. For example, the detection of the edge point P may be advanced in the main scanning direction D1 at a plurality of set X coordinates. Further, the detection of the edge point P in the main scanning direction D1 and the detection of the edge point P in the transport direction D2 may be performed in combination. That is, when the difference value ΔY is larger than the reference interval K with respect to one side of the document image range GH, a plurality of edge points P are detected for the adjacent sides, and the difference value ΔY is calculated. Also good.

(4) In the above embodiment, the edge image EG is generated from the gradation data read from the entire area in the main scanning direction D1 of the CIS 24, and the edge image EG of the central portion IB is selected in the edge point detection process. However, the present invention is not limited to this. For example, the edge image EG may be generated from the gradation data read in the center portion of the CIS 24 in the main scanning direction D1. Thereby, the storage capacity of the RAM 13 for storing the edge image EG can be reduced, and the manufacturing cost of the apparatus can be reduced.

1: Image reading apparatus, 10: ASIC, 11: CPU, 14: device control unit, 15: AFE, 16: drive circuit, 17: image processing circuit, 22: transport path, 24: CIS, 25: F sensor, 26 : R sensor, 31: interval acquisition unit, 32: tilt correction unit, 33: control unit, D1: main scanning direction, D2: transport direction, EG: edge image, SG: scanned image, H: reading range, GH: document Image range, P, Pn: Edge point, ΔY: Difference value, θH: Correction amount, K: Reference interval, nK: Required number, θL: Limit for correction

Claims (2)

A transport unit that transports a document along a transport path;
A reading unit that is located on the conveyance path, reads the document conveyed by the conveyance unit in a main scanning direction, and outputs an image of a reading range including a document image range obtained by reading the document;
An edge image conversion unit that performs binarization on the image output by the reading unit and converts the image into an edge image;
CPU,
With
The CPU
Edge point detection processing for detecting a plurality of edge points from the edge image converted by the edge image conversion unit to each of one side of the leading edge of the document image range and one side of the trailing edge of the document image range When,
An interval acquisition process for acquiring a difference between a maximum value and a minimum value of the position of the edge point in a direction orthogonal to the main scanning direction;
When the difference between the positions on one side of the leading edge of the document image range acquired by the interval acquiring process is equal to or smaller than a reference interval, a plurality of one side of the leading edge of the document image range detected by the edge point detecting process Detecting the inclination of the document with respect to the reading unit from the edge point, and correcting the inclination of the image in the document image range according to the inclination ,
When the difference in the position on one side of the leading edge of the document image range acquired by the interval acquisition process is larger than the reference interval, one side of the trailing edge of the document image range detected by the edge point detection process Detecting the inclination of the document with respect to the reading unit from a plurality of edge points on the side, and correcting the inclination of the image in the document image range according to the inclination,
When the difference between the positions of a plurality of edge points detected with respect to one side of the rear end of the document image range is larger than the reference interval, tilt correction is not performed on the image in the document image range.
Tilt correction processing,
An image reading apparatus that executes The image reading apparatus according to claim 1 ,
The edge point detection processing detects a plurality of edge points from the image read in the central portion excluding both end portions in the main scanning direction of the reading range, the image reading apparatus.

Images