JP2013251779A - Image processing apparatus and image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve the problem in which: functions useful in outputting a read image include a function to determine only a document area, trim only the determined part, and perform inclination correction to output an image; and accurate determination of a document area and an inclination angle is required for the function.SOLUTION: A pixel group in a document area is extracted from a read image; a plurality of parallel lines in contact with the pixel group are created; and a rectangular area surrounded by parallel lines with the minimum distance therebetween and parallel lines orthogonal to said parallel lines is determined as a document area.

Description

  The present invention relates to an image processing apparatus that reads a document placed on a platen of the image reading apparatus, processes the read image, and outputs the processed image. In particular, the present invention trims only a document region placed on the platen. The present invention relates to an image processing apparatus capable of performing rotation correction and outputting.

  Devices used when digitizing paper media data and importing it into a PC or the like include imaging scanners and multifunction devices. These image input devices can read many types of originals such as monochrome images such as text originals, color images such as magazines, developed photographs, negatives and positive films.

  On the other hand, the reading range of the document table can be specified by the user, but may include an area other than the document to be read. Therefore, as a convenient function when outputting the read image, there is a function of determining a document area to be read, trimming only that portion, correcting the rotation, and outputting it. For these functions, it is necessary to accurately determine only the document area to be read from the read image. In general, the brightness and saturation of the read image are calculated, edge extraction and threshold processing are performed, the document area and the non-document area in the read image are discriminated, and the document area is extracted. Then, the shape and its inclination angle are obtained from the extracted document area. Therefore, when the color of the original and the color of the image reading device component that is the background of the read image data are very similar, the brightness and saturation of the original area and the non-original area in the read image are very similar. Therefore, the discrimination accuracy is lowered. Also, if you try to distinguish between a document area with very similar brightness and saturation and a non-document area by changing the threshold value, the noise in the scanned image and the noise in the document will be detected. As a result, the determination accuracy is lowered.

  As a method for obtaining the inclination angle of the document area, there is a technique for obtaining the edge inclination angle of the extracted document area. The number of samples is determined based on the length of the edge, and coordinates corresponding to the number of samples are extracted at regular intervals from the edge, whereby the inclination angle is obtained with high accuracy (Patent Document 1). In addition, as a method for obtaining the shape and tilt angle of the document area, the extracted document area is treated as a polygon polygon, the polygon polygon is rotated at a constant angular interval, and the rectangular area that encompasses the rotated polygon is minimized. In some cases, this angle is determined as the tilt angle (Patent Document 2).

JP-A-8-331328 JP 2004-129271 A

  However, the conventional technique is based on the premise that the extracted document area coincides with the document area to be read, and if the extracted document area is missing even a little, the tilt angle is incorrect. Also, rotating the extracted document area is computationally intensive and requires a large memory.

  Therefore, the present invention has an object of accurately performing shape determination, determining an inclination angle, trimming a range of the determined shape, and correcting rotation by the determined inclination angle even if the extracted document area is missing. And Further, it is an object to determine the shape and determine the tilt angle efficiently without requiring a large memory.

In order to solve the above-described problems, the image processing apparatus of the present invention has the following configuration.
A candidate pixel group extracting unit configured to extract a candidate pixel group estimated to be the original from the read original plate image in an image processing apparatus including a unit that reads an original placed on the original plate and obtains an original plate image; A parallel line creating unit that creates a plurality of parallel lines in contact with the candidate pixel group; and a parallel line created by the parallel line creating unit and surrounded by a parallel line having a minimum parallel line distance and a parallel line perpendicular thereto A document area determining means for determining the rectangular area as a document area, a clipping means for cutting out the area determined by the document area determining means from the document table image, and an image clipped by the clipping means as the document area. The determining unit includes a rotating unit that rotates by an inclination angle of the parallel line with the minimum parallel line distance with respect to the document table.

  Even if the candidate pixel group extraction unit cannot extract all of the true document area due to factors such as the color and thickness of the document placed on the document table, the parallel line creation unit and the document region determination unit save memory. In addition, an approximate rectangle of the true document area can be obtained accurately. Then, trimming and rotation correction can be performed accordingly.

It is a hardware block diagram of an image reading apparatus. 6 is a flowchart illustrating a reading operation of the scanner 101 by the host PC 17 in the first embodiment. 6 is a diagram illustrating that a document 302 is placed on a document table 301 in Embodiment 1. FIG. 6 is a flowchart illustrating an operation of extracting document area pixels in the first exemplary embodiment. FIG. 10 is a diagram illustrating a document area pixel extraction result 501 according to the first exemplary embodiment. 5 is a flowchart illustrating an operation of determining a rectangular area corresponding to a document area pixel 502 in Embodiment 1. 6 is a diagram illustrating parallel lines in contact with a document area pixel 502 and the ferret diameter L in Embodiment 1. FIG. 6 is a diagram illustrating a rectangular area 801 of document area pixels determined in step S312 in the first exemplary embodiment. FIG. 10 is a flowchart illustrating an operation of determining a rectangular area corresponding to a document area pixel 502 in Embodiment 2. In Example 2, it is a figure which shows the parallel line which contact | connected the original area pixel 502, and its ferret diameter sum L. FIG.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. The following embodiments do not limit the present invention according to the claims, and all combinations of features described in the present embodiments are essential to the solution means of the present invention. Not exclusively.

<Embodiment 1>
The configuration of the image reading apparatus 101 will be described with reference to the block diagram of FIG. In this figure, 1 is a read original.

  The light source lamp 11 illuminates the document 1, and reflected light having an intensity corresponding to the density on the surface of the document is formed on a line image sensor such as a CCD sensor, which is a solid-state image pickup device 3, through a joint lens 2. To do.

  Reference numeral 10 denotes a light source lighting circuit for driving and lighting the light source lamp 11.

  Reference numeral 4 denotes an amplifier that amplifies the analog image signal output of the line image sensor 3.

  Reference numeral 12 denotes a motor drive circuit that drives an optical system drive motor 13 such as a stepper motor, and outputs an excitation signal of the drive motor 13 by a control signal from a CPU controller 9 that is a system control means of the image reading apparatus 101.

  An A / D converter 5 converts the analog image signal output from the amplifier 4 into a digital image signal.

  An image processing circuit 6 performs image processing such as offset correction, shading correction, digital gain adjustment, color balance adjustment, color masking conversion, and resolution conversion in the main and sub-scanning directions on a digital image signal. .

  Reference numeral 7 denotes a buffer memory composed of a RAM, which temporarily stores image data.

  An interface circuit 8 mediates command and image communication with the connected host PC 17. The interface 8 includes a USB interface, IEEE1394, LAN, or the like.

  A work memory 14 is used as a temporary work memory when the image processing circuit performs image processing. This working memory 14 is used for correcting an offset between RGB lines of image signals from RGB line sensors arranged in parallel on the line image sensor 1203 with a predetermined offset. The working memory 14 also temporarily stores various data such as shading correction.

  Reference numeral 15 denotes a gamma LUT for storing a density gamma conversion LUT and performing gamma correction.

  A CPU controller 9 controls the image reading apparatus 101 according to a command from the connected host PC, and controls the motor driving circuit 12, the light source lighting circuit 10, the image processing circuit 6, and the like.

  The state in which the switch provided on the operation panel 16 is pressed is detected by the CPU controller and notified to the host PC connected via the interface.

  In this embodiment, the three-line CCD sensor 3 for reading three colors of RGB and the white light source 11 are used. However, the same applies to the structure of the single-color one-line image sensor and the RGB three-color light sources that can be selectively turned on. Can be realized. Although not shown, the light source has a three-color LED configuration, and the CPU controller lights one color of the three light source LEDs by the light source lighting circuit, and reads the illuminating illumination light with the image sensor. By reading the LEDs that are turned on in order, the original image can be read by color separation based on the light emission color of the light source.

  Next, in Embodiment 1, the reading operation of the scanner 101 by the host PC 17 will be described.

  FIG. 2 is a flowchart illustrating the reading operation of the scanner 101 by the host PC 17 in the first embodiment. Assume that a document 302 to be read is placed on the document table 301 as shown in FIG.

  First, in S1, an image of the entire document table is read from the scanner. The resolution at the time of reading may be temporary or may be a resolution desired by the user. In S2, pixels in the area of the document 302 are extracted from the read image data.

  Details of the document area pixel extraction processing in S2 will be described.

  FIG. 4 is a flowchart showing the operation of extracting document area pixels. In the flowchart in FIG. 2, this corresponds to the process of S2.

  In S21, a threshold for binarization is determined from the image. This threshold value varies depending on the comparison method in S26 described later. In order to easily determine the threshold value, a fixed value may be determined in advance.

  In S22, the value of one pixel is acquired. In order to extract the document area pixels from the image, it is necessary to perform processing for all the pixels, but in S22, the processing can be performed for each pixel. Usually, the position of a certain pixel is specified using the X coordinate and the Y coordinate. At the start of processing, the X and Y coordinates are initialized with initial values (generally 0), and every time one pixel is processed, the X and Y coordinates are changed to scan all pixels.

  In S23, the color space of the pixel value acquired in S22 is converted. Generally, the color space for each scanner 101 differs depending on the characteristics of the CCD 3, the color filter, and the light source lamp 11. If a device-independent color space is used, there is a possibility that an object can be extracted without depending on the scanner 101. Therefore, the color space is converted in S23. When the parameter depending on the scanner 101 is adjusted to determine the threshold value in the process of S21, the process of S23 can be omitted.

  In S24, the value obtained in S23 is converted into a scalar value. When inputting a color image, it has RGB three-color values. In order to compare the RGB three-color value (vector value) with a threshold value (scalar value), the RGB three-color value is converted into a scalar value. When converting RGB three-color values to scalar values, a method of extracting only one color, a method of obtaining a luminance value by taking an appropriate weighted average of the RGB three-color values, a method of calculating saturation from the RGB three-color values, etc. There is. However, when the input image is a single color such as a gray scale, this processing is not necessary, so the processing of S24 can be omitted.

  In S25, an Nth order differential and a difference are calculated from the values obtained in S24. In the process of extracting the document area from the image, it may be easy to accurately determine the subsequent document area by extracting the document 302 placed on the document table 301 and other boundaries. In order to extract the boundary of the document 302 placed on the document table 301, an Nth order differential and a difference are calculated. Since this process depends on the characteristic of the value obtained in S24, the process of S25 can be omitted if not necessary.

  In S26, the value obtained in S25 is compared with the threshold value determined in S21. However, depending on the values obtained in S23 to S25, this relationship may be reversed, and if it is less than the threshold value, the document area may be set, and if it is equal to or greater than the threshold value, the document area may not be set. This relationship is determined in advance. For example, if the luminance value is less than the threshold value, the document area is set. If the saturation value is greater than the threshold value, the document area is set.

  In S27, the result of S26 is stored. Since there are only two types of results of S26, that is, the document area or the document area, 0 is encoded and stored such that 0 is the document area, 1 is not the document area, and the like.

  In S28, it is checked whether or not all the pixels have been processed in S27. If all the pixels have been processed, the process ends.

  In the first embodiment, the processing is performed in the order of the flowchart shown in FIG. 4, but the result of S25 may be required to determine the threshold value in S21, and the value of the conversion to the scalar value (S24) for the adjacent pixels is It may be necessary for the calculation of S25. For this reason, the processing order of the flowchart shown in FIG. 4 may be changed as necessary. In the first embodiment, the flowchart shown in FIG. 4 is executed only once. However, in some cases, it may be executed a plurality of times, and at this time, the internal processing method may be changed. For example, in the first processing, luminance is obtained without performing color space conversion, and processing is performed by second order differentiation. In the second process, the color space is converted, the saturation is obtained, and the process skips S25. Thereafter, a logical product or logical sum of the two results is obtained and combined. Whether to use a logical product or a logical sum depends on the encoding of S27, and is thus determined as appropriate.

  FIG. 5 is a diagram illustrating a document area pixel extraction result 501 in the first embodiment. A pixel group 502 filled with black is a pixel group extracted as a document area. This figure shows a case where a part of the original 302 cannot be extracted and is missing.

  In S <b> 3, a rectangular area corresponding to the original area pixel 502 is determined using a rectangle determining unit. A plurality of parallel lines that contact the document area pixel 502 are created, and are surrounded by a parallel line that minimizes the distance (Ferret diameter) of the parallel lines, and a parallel line that is orthogonal to the parallel line and contacts the document area pixel 502 The area is determined as a rectangular area of the document area pixel 502.

  Details of the rectangle determination process in S3 will be described.

  FIG. 6 is a flowchart showing an operation for determining a rectangular area corresponding to the document area pixel 502.

  An initialization process is performed in S301. The parallel line angle X is 0, the parallel line distance (Ferret diameter) is 0 when the minimum X is maintained, A is the maximum ferret diameter, FERET_MAX is 0, and the minimum ferret diameter is maintained. FERET_MIN to be initialized is initialized to the maximum settable value. Further, the search range R and the search interval D for checking the ferret diameter are set to predetermined values. Generally, the search range R is 180. More generally, the search interval D is determined from the calculation time and accuracy. If the calculation time is large but you want to check it accurately, decrease the value. If the accuracy is rough but you want to reduce the calculation time, increase the value.

  In step S <b> 302, a parallel line that is in contact with the document area pixel 502 and has an angle X with respect to the document table 301 is created. In S303, the distance (Ferret diameter L) of the parallel lines created in S302 is calculated. Here, not the distance of the parallel lines but the projected width of the parallel lines may be used.

  FIG. 7 is a diagram showing parallel lines in contact with the document area pixel 502 and the ferret diameter L thereof. FIG. 7A shows the parallel lines 701 and the ferret diameter L702 of the parallel lines 701 when X = 0. FIG. 7B shows the parallel lines 703 and the ferret diameter L704 of the parallel lines 703 when X = 45 degrees. FIG. 7C shows the parallel line 705 and the ferret diameter L706 of the parallel line 705 when X = 90 degrees.

  In S304 and S305, when the ferret diameter L obtained in S302 is smaller than the minimum ferret diameter FERET_MIN held, L is set in FERET_MIN, and X is set in the angle A of the parallel line at the minimum ferret diameter. In S306 and S307, when the ferret diameter L obtained in S302 is larger than the maximum ferret diameter FERET_MAX held, L is set in FERET_MAX.

  If the parallel line angle X is within the search angle range R in S308, the search interval D is added to the angle X in S309, and the process returns to S302. If it is out of range, go to S310.

  A circle is determined in S310, and if it is determined to be a circle, rotation is not performed. When the document area pixel 502 is a perfect circle, the minimum ferret diameter and the maximum ferret diameter are equal. Therefore, the obtained maximum ferret diameter FERET_MAX and the minimum ferret diameter FERET_MIN are compared. If the difference is equal to or smaller than the threshold value, it is determined that the document area pixel 502 is circular, and the parallel line angle A at the minimum ferret diameter is 0. Is set (S311). In S312, a parallel line having an angle of A and in contact with the document area pixel 502 and a parallel line perpendicular to the parallel line having an angle of A and in contact with the document area pixel 502 are created, and their four intersections are obtained. A rectangle having the obtained four points as vertices is determined as a rectangular region 801 corresponding to the document region pixel 502.

  FIG. 8 is a diagram showing a rectangular area 801 of document area pixels determined in S312. A parallel line 802 at an angle A and a parallel line 803 orthogonal to the parallel line 802 are created, and their four intersections 804, 805, 806, and 807 are obtained. A rectangle 801 having vertices at intersections 804, 805, 806, and 807 is determined.

  In S4, the rectangular area 801 determined in S3 is cut out from the read image data.

  In S5, the image data of the rectangular area 801 cut out in S4 is rotated by the angle A obtained in S4. The rotated image data is the final output data.

<Embodiment 2>
Since the configuration of the image reading apparatus 101 used in this embodiment can be the same as that described with reference to FIG. 1, description thereof will be omitted.

  Since the reading operation of the scanner 101 by the host PC 17 is the same as the flowchart shown in FIG. 2, the description thereof is omitted.

  Assume that a document 302 to be read is placed on the document table 301 as shown in FIG. Assume that the document area pixel 502 shown in FIG. 5 is obtained as a result of the image area pixel extraction S2.

  FIG. 9 is a flowchart showing an operation for determining a rectangular area for the document area pixel 502.

  Initialization is performed in S401. Holds the angle X when the angle X of the parallel line is 0, and the sum of the distance of the parallel line of the angle X (Ferret diameter) and the distance of the parallel line orthogonal to the parallel line of the angle X (Ferret diameter sum) is minimum. A to be initialized is initialized to 0. FERET_SUM_MAX that holds the maximum ferret diameter sum is initialized to 0, and FERET_SUM_MIN that holds the minimum ferret diameter sum is initialized to a settable maximum value. Further, the search range R and the search interval D for checking the ferret diameter sum are set to predetermined values. In general, the search range R is 90 degrees. More generally, the search angle D is determined from the calculation time and accuracy. If the calculation time is large but you want to check it accurately, decrease the value. If the accuracy is rough but you want to reduce the calculation time, increase the value. In step S <b> 402, parallel lines that are in contact with the document area pixel 502 and have an angle X with respect to the document table 301 and parallel lines that are in contact with the document area pixel 502 and have an angle of X are created. In S403, the sum of the distances of the parallel lines created in S402 (Ferret diameter sum L) is calculated. Here, not the distance of the parallel lines but the projected width of the parallel lines may be used.

  FIG. 10 is a diagram showing the parallel lines in contact with the document area pixel 502 and the ferret diameter thereof. FIG. 10A shows a parallel line 1001 and a ferret diameter L1 1002 of the parallel line 1001 when X = 0, and a parallel line 1003 orthogonal to the parallel line 1001 and a ferret diameter L2 1004 of the parallel line 1003. FIG. 7B shows the ferret diameter L1 1006 of the parallel line 1005 and the parallel line 1005 when X = 45 degrees, and the ferret diameter L2 1008 of the parallel line 1007 and the parallel line 1007 when X = 45 degrees + 90 degrees. Yes. The ferret diameter sum L at this time is L1 + L2.

  In S404 and S405, when the ferret diameter sum L obtained in S402 is smaller than the minimum ferret diameter sum FERET_SUM_MIN held, L is set in FERET_SUM_MIN, and the parallel line angle A at the minimum ferret diameter sum is set to X Set. In S406 and S407, when the ferret diameter sum L calculated in S402 is larger than the maximum ferret diameter sum FERET_SUM_MAX held, L is set in FERET_SUM_MAX.

  If the angle X of the parallel line is within the search angle range R in S408, the search interval D is added to the angle X in S409, and the process returns to S402. If it is out of range, go to S410.

  A circle is determined in S410, and if it is determined to be a circle, rotation is not performed. When the document area pixel 502 is a perfect circle, the minimum ferret diameter sum and the maximum ferret diameter sum are equal. Therefore, the obtained maximum ferret diameter sum FERET_SUM_MAX is compared with the minimum ferret diameter sum FERET_SUM_MIN, and when the difference is equal to or smaller than the threshold value, it is determined that the document area pixel 502 is circular, and the parallel lines at the time of the minimum ferret diameter sum Is set to 0 (S411).

  In step S412, a parallel line having an angle A and in contact with the document area pixel 502 and a parallel line perpendicular to the parallel line having the angle A and in contact with the document area pixel 502 are created, and their four intersections are obtained. A rectangle having the obtained four points as vertices is determined as a rectangular region of the document region pixel 502.

<Embodiment 3>
An object of the present invention is to supply a recording medium recording software program codes for realizing the functions of the above-described embodiments to a system or apparatus, and store the computer (or CPU or MPU) of the system or apparatus in the recording medium. Needless to say, this can also be achieved by reading and executing the program code. In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiment, and the storage medium storing the program code constitutes the present invention.

  As a storage medium for supplying the program code, for example, a flexible disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a CD-R, a magnetic tape, a nonvolatile memory card, a ROM, a DVD, or the like is used. it can.

  Further, by executing the program code read by the computer, not only the functions of the above-described embodiments are realized, but also an operating system (OS) running on the computer based on the instruction of the program code. It goes without saying that a case where the function of the above-described embodiment is realized by performing part or all of the actual processing and the processing is included.

  Furthermore, after the program code read from the storage medium is written to the memory provided in the function expansion board inserted into the computer or the function expansion unit connected to the computer, the function is based on the instruction of the program code. It goes without saying that the CPU of the expansion board or function expansion unit performs part or all of the actual processing, and the functions of the above-described embodiments are realized by the processing.

301: Document table 302: Document 502: Candidate pixel group

Claims (7)

In an image processing apparatus including a document (302) on which an object to be read is described, and a unit that reads a document placed on a document table (301) and obtains a document table image,
Candidate pixel group extraction means (S2) for extracting a candidate pixel group (502) estimated as the object from the read document table image;
Parallel line creation means (S302) for creating a plurality of parallel lines in contact with the candidate pixel group;
Of the parallel lines created by the parallel line creating means, a rectangular area surrounded by a parallel line (802) having the smallest parallel line distance and a parallel line (803) orthogonal thereto is determined as a document area. Means (S312);
Clipping means for cutting out the area determined by the document area determining means from the document table image;
An image processing apparatus comprising: a rotation unit configured to rotate the document area cut out by the clipping unit by an angle of a parallel line surrounding the document area with respect to the document table. The image processing apparatus according to claim 1, wherein the parallel line creating unit creates parallel lines in contact with the candidate pixel group by changing an angle with respect to the document table. The document area determining means compares the parallel lines created by the parallel line creating means with the maximum parallel line distance and the minimum parallel line distance (S310). The image processing apparatus according to claim 1, wherein a rectangular area surrounded by the parallel lines and a parallel line perpendicular thereto is determined as a document area (S 311). In an image processing apparatus including a document (302) on which an object to be read is described, and a unit that reads a document placed on a document table (301) and obtains a document table image,
Candidate pixel group extraction means (S2) for extracting a candidate pixel group estimated as the object from the read document table image;
Parallel line creation means (S402) for creating a plurality of sets of parallel lines in contact with the candidate pixel group and parallel lines orthogonal to the candidate pixel group and in contact with the candidate pixel group;
A document area determining means (S412) for determining, as a document area, a rectangular area surrounded by a set of parallel lines that minimizes the sum of the parallel line distances among a set of parallel lines created by the parallel line creating means;
Clipping means for cutting out the area determined by the document area determining means from the document table image;
An image processing apparatus comprising: a rotation unit configured to rotate the document area cut out by the clipping unit by an angle of a parallel line surrounding the document area with respect to the document table. The image processing apparatus according to claim 4, wherein the parallel line creation unit creates a set of parallel lines in contact with the candidate pixel group by changing an angle with respect to the document table. The document area determination unit compares the parallel line set created by the parallel line creation unit with the largest and smallest of the parallel line distances (S410), and the difference is within a threshold value. 5. The image processing apparatus according to claim 4, wherein a rectangular area surrounded by parallel lines having an angle of 0 and parallel lines orthogonal thereto is determined as a document area (S 411). In a program for reading a document (302) on which an object to be read is written, the document placed on the document table (301) and obtaining a document table image, the computer executes the procedure.
A candidate pixel group extraction procedure (S2) for extracting a candidate pixel group estimated as the object from the read document table image;
A parallel line creation procedure (S402) for creating a plurality of pairs of parallel lines in contact with the candidate pixel group and parallel lines orthogonal to and in contact with the candidate pixel group;
A document region determination procedure (S412) for determining, as a document region, a rectangular region surrounded by a set of parallel lines that minimizes the sum of the parallel line distances among a set of parallel lines created by the parallel line creation unit;
Clipping means for cutting out the area determined by the document area determining means from the document table image;
A program for causing a computer to execute a rotation procedure for rotating a document area cut out by the clipping means by an angle of a parallel line surrounding the document area with respect to the document table.