JP4829757B2 - Ruled line extraction apparatus and method - Google Patents

Description

  The present invention relates to an apparatus and method for extracting ruled lines from a color image.

  FIG. 11 is a flowchart of a conventional ruled line extraction process that extracts a ruled line arranged in an image using a color image as an input. In this ruled line extraction process, first, a lightness gray image is generated from the input color image, and the lightness gray image is binarized (step 11). Then, a ruled line and cells surrounded by the ruled line are extracted from the obtained binary image (step 12). Thereby, a ruled line having a brightness difference from the background can be extracted. The lightness gray image corresponds to a so-called gray scale image.

  Further, as a method of extracting ruled lines from an image having a small density difference between the background and the ruled lines, a method of binarizing a multi-valued image by two different methods is known (see, for example, Patent Document 1 below). . In this method, a ruled line candidate region is extracted using one squashed binary image, and a ruled line candidate region is verified using the other shrunken binary image to more accurately extract a ruled line. be able to.

Prior application PCT / JP2006 / 302763 relates to a method of performing saturation correction in order to recognize a region color of a color image.
JP 2002-133426 A

The conventional ruled line extraction method described above has the following problems.
In the ruled line extraction process of FIG. 11, when the difference in brightness between the background and the ruled line is small, both the ruled line and the background become white due to binarization, and the ruled line cannot be extracted. For example, as shown in FIG. 12, light-colored cells may not be distinguished by an explicit boundary line but may be distinguished by a color change. Hereinafter, such a boundary between cells is referred to as a boundary ruled line. When a color image including such a boundary ruled line is input, both cells become white due to binarization, and the ruled line cannot be extracted.

  When the color image 21 shown in FIG. 13 is input, a lightness gray image 22 is first generated, and then a binary image 23 is generated from the lightness gray image 22 (step 11). Then, ruled lines and cells are extracted from the binary image 23, and an extraction result 24 is obtained (step 12). However, the extraction result 24 does not include the boundary ruled line included in the color image 21.

  In the ruled line extraction method of Patent Document 1, since a squashed binary image and a blurred binary image are generated from only a lightness gray image, a ruled line is obtained from a color image having the same lightness but different saturation. There is a problem that cannot be extracted properly.

  An object of the present invention is to extract a ruled line from a color image having a small difference in brightness between the background and the ruled line, such as a boundary ruled line between light-colored cells.

  FIG. 1 is a principle diagram of the ruled line extraction apparatus of the present invention. The ruled line extraction apparatus in FIG. 1 includes a storage unit 101, a first generation unit 102, a second generation unit 104, a first extraction unit 103, a second extraction unit 105, and an output unit 106.

  The storage unit 101 stores a color image 111. The generation unit 102 generates a lightness gray image 112 from the color image 111 and stores it in the storage unit 101. The extraction unit 103 extracts the first ruled line using the lightness gray image 112. The generation unit 104 generates a saturation gray image 113 of the cell region surrounded by the first ruled line, and stores it in the storage unit 101. The extraction unit 105 extracts the second ruled line using the saturation gray image 113. The output means 106 outputs information on the first and second ruled lines.

  The generation unit 102 extracts the color image 111 from the storage unit 101 and generates a lightness gray image 112 using the color information. The extraction unit 103 extracts the lightness gray image 112 from the storage unit 101, extracts a first ruled line using the lightness information, and transfers the information to the output unit 106.

  The generation unit 104 generates the saturation gray image 113 using the color information of the cell region surrounded by the first ruled line. The extraction unit 105 extracts the saturation gray image 113 from the storage unit 101, extracts the second ruled line using the saturation information, and transfers the information to the output unit 106. The output means 106 outputs the received first and second ruled line information.

  The storage unit 101 corresponds to, for example, a memory 902 in FIG. 9 described later, and the generation units 102 and 103 and the extraction units 104 and 105 correspond to, for example, a CPU (central processing unit) 901. The output unit 106 corresponds to the output device 904, for example.

  According to the present invention, by using a combination of ruled line extraction processing using a lightness gray image and ruled line extraction processing using a saturation gray image, the brightness difference from the background, which has been difficult to extract by the conventional technology, is improved. A small ruled line can be appropriately extracted.

The best mode for carrying out the present invention will be described below in detail with reference to the drawings.
In the present embodiment, in addition to the ruled line extraction process using the lightness gray image and the ruled line extraction process using the saturation gray image, the ruled line extraction process using the saturation correction image is further used in combination. Thereby, it becomes possible to extract a ruled line with a small brightness difference from the background more precisely.

  FIG. 2 is a flowchart of such ruled line extraction processing. The ruled line extraction device first generates a lightness gray image from a color image input by a scanner or the like, binarizes the lightness gray image, and generates a binary image (step 201). Then, a ruled line and a cell surrounded by the ruled line are extracted from the binary image (step 202). When there is a certain brightness difference between the background and the ruled line, the ruled line is extracted by this processing.

  Next, a saturation gray image is generated for each cell area extracted in step 202, and the saturation gray image is binarized to generate a binary image (step 203). Then, a ruled line and a cell surrounded by the ruled line are extracted from the binary image (step 204). When there is no brightness difference between the background and the ruled line but there is a chroma difference, the ruled line is extracted by this processing.

Next, it is checked whether or not the processing of all cells has been completed (step 205). If there is an unprocessed cell, the processing of step 203 and step 204 is repeated.
When all the cells have been processed, a saturation correction image is generated for each cell region extracted in step 204, and the saturation correction image is binarized to generate a binary image. (Step 206). Then, a ruled line and a cell surrounded by the ruled line are extracted from the binary image (step 207). However, for the cells for which the ruled lines and the cells are not extracted in step 204, the process is performed on the cell area extracted in step 202.

  Although the brightness difference and saturation difference between the background and the ruled line are both small, if the hue is different, the hue difference can be clearly recognized by appropriately correcting the saturation. Therefore, ruled lines are extracted by this processing.

  Next, it is checked whether or not all the cells have been processed (step 208). If there is an unprocessed cell, the processing in step 206 and step 207 is repeated. When all the cells have been processed, the ruled line extraction process ends.

  In this way, ruled lines having a small brightness difference from the background can be extracted by using binary images based on different color attributes in combination. Depending on the purpose of the ruled line extraction process, the processes in steps 206 to 208 can be omitted.

  FIG. 3 shows an example of a boundary ruled line between cells having a brightness difference. The color image 301 includes boundary ruled lines. The hue gray image 302 and the saturation gray image 303 cannot recognize the boundary ruled lines, but the lightness gray image 304 can recognize them. In such a case, a boundary ruled line is extracted by the process of step 202.

  FIG. 4 shows an example of a border ruled line between cells having a saturation difference. The color image 401 includes boundary ruled lines. The hue gray image 402 and the lightness gray image 404 cannot recognize the boundary ruled lines, but the saturation gray image 403 can recognize them. In such a case, a boundary ruled line is extracted by the process of step 204.

  FIG. 5 shows an example of a boundary ruled line between cells having a hue difference. The color image 501 includes boundary ruled lines. The saturation gray image 503 and the lightness gray image 504 cannot recognize the boundary ruled lines, but the hue gray image 502 can recognize them. In such a case, a boundary ruled line is extracted by the processing in step 207.

In step 201 in FIG. 2, a lightness gray image is generated by converting the RGB value of the color image into lightness I (0 ≦ I ≦ 1) by the following equation.

I = 0.299 × R + 0.587 × G + 0.114 × B (1)

In step 203, it is checked whether or not there is a ruled line with a small brightness difference that cannot be extracted by the conventional processing in each cell. First, the RGB values of the pixels in the cell are extracted and converted to saturation S (0 ≦ S ≦ 2 ^1/2 ) by the following equation.

Next, the saturation S is converted into a discrete value of 0 to 255 by normalization, and a saturation gray image is generated using the discrete value.

  In step 206, for example, the saturation correction image of each cell is generated from the color image using the saturation correction processing described in PCT / JP2006 / 302763 of the prior application. In this process, in a two-dimensional space with brightness and saturation as coordinate values, the boundary line between an achromatic color and a chromatic color is used as a threshold value, and the saturation of each pixel is compared with the threshold value. The calculation is performed to increase the value and to decrease the saturation sharply on the achromatic side.

  The ruled line extraction apparatus first compares the saturation S with a preset threshold value α. Since the threshold value α is set according to the lightness I, information on the color image of the correction target area is necessary when performing the saturation correction process.

Next, an intermediate correction value S ′ is calculated according to the comparison result. Specifically, when the saturation S is smaller than the threshold value α, it is determined that the color is achromatic, and the intermediate correction value S ′ is calculated by the following equation.

S ′ = (S−α) / α (5)

On the other hand, when the saturation S is larger than the threshold value α, it is determined that the color is a chromatic color, and an intermediate correction value S ′ is calculated by the following equation.

S ′ = (S−α) / (1−α) (6)

Next, using the calculated intermediate correction value S ′, the corrected saturation S ″ is calculated by the following equation.

Here, A and B in the equation (7) are calculated by the following equations.

A = (e ^T −1) / (e ^T +1) (8)
B = 1 / (1 + e ^T ) (9)

The parameter T in the equations (8) and (9) is set to T = 10, for example.

  FIG. 6 shows the relationship between the saturation S ″ before correction and the saturation S ″ after correction. In the graph 610, the horizontal axis represents the intermediate correction value S ′, and the vertical axis represents the saturation S ″ after correction. Represents. The intermediate correction value S ′ takes a value in the range of −1 to 1, and the corrected saturation S ″ takes a value in the range of 0 to 1. A straight line 620 below the graph 610 indicates a saturation before correction. Degree of degree S. S ′ = − 1 corresponds to S = 0, and S ′ = 1 corresponds to S = 1.

  Expression (7) is represented by a sigmoid curve 601. The rate of change of the saturation S ″ after correction with respect to the saturation S varies depending on the value of the intermediate correction value S ′, and decreases as the intermediate correction value S ′ approaches −1 or 1, and increases rapidly near 0. .

  Substituting the intermediate correction value S ′ of the equation (5) into the equation (7) makes the corrected saturation S ″ larger than the saturation S, and substituting the equation (6) into the equation (7) The saturation S ″ is smaller than the saturation S. That is, when the saturation S is smaller than the threshold α, the corrected saturation S ″ is a value larger than the saturation S, and when the saturation S is larger than the threshold α, the corrected saturation S ″ is The value is smaller than the saturation S. As a result, the saturation of the achromatic color is brought close to black (saturation S = 0), and the saturation of the chromatic color is emphasized so as to make the hue more clear.

  For example, if α = 0.18 and S = 0.3, as shown by the broken line, S ′ = 0.15 from the equation (6) and S ″ = 0.8 from the equation (7). The saturation S ″ after the correction is larger than the saturation S before the correction, and the saturation is emphasized.

Next, a specific example of the ruled line extraction process of FIG. 2 will be described with reference to FIGS.
When a color image 701 as shown in FIG. 7 is input, a lightness gray image 702 is generated, and the lightness gray image 702 is binarized (step 201). Then, ruled lines and cells are extracted from the obtained binary image (step 202). In this case, a boundary ruled line is included in the area 711 of the color image 701, but the boundary ruled line is not extracted.

  Therefore, a saturation gray image 703 is generated using the RGB values of the region 711 of the color image 701 corresponding to the extracted cell 712, and a binary image 704 is generated from the saturation gray image 703 (step 203). ). Then, an attempt is made to extract ruled lines and cells from the binary image 704 (step 204), but the boundary ruled lines cannot be extracted.

  Next, a saturation correction image 705 is generated using the RGB values of the region 711, and a binary image 706 is generated from the saturation correction image 705 (step 206). Then, ruled lines and cells are extracted from the binary image 706 (step 207). In this case, the hue correction between the cells in the region 711 is clarified by the saturation correction, and the boundary ruled line can be extracted.

  When a color image 801 as shown in FIG. 8 is input, a lightness gray image 802 is generated, and the lightness gray image 802 is binarized (step 201). Then, ruled lines and cells are extracted from the obtained binary image, and an extraction result 803 is obtained (step 202). In the extraction result 803, two cells 821 and 822 are extracted.

  Next, saturation gray images 804 and 805 for each cell are generated using the RGB values of the area of the color image 801 corresponding to each of the cells 821 and 822, and the saturation gray images 804 and 805 are respectively binary. (Step 203). Then, ruled lines and cells are extracted from the obtained binary image, and extraction results 806 and 807 are obtained (step 204). In the extraction result 807, three cells 823, 824, and 825 are extracted.

  Next, saturation correction images 808, 809, and 810 are generated using the RGB values of cells 823, 824, and 825, and the saturation correction images are binarized (step 206). Then, ruled lines and cells are extracted from the obtained binary image, and extraction results 811, 812, and 813 are obtained (step 207). In the extraction result 813, three cells 826, 827, and 828 are extracted.

  Note that the saturation correction process in step 206 is not limited to the process described above, and the saturation may be corrected by another method. For example, the saturation correction image may be generated by directly correcting the saturation gray image generated in step 203 by using an appropriate calculation without using the input color image.

  The ruled line extraction apparatus of the present embodiment is configured using, for example, an information processing apparatus (computer) as shown in FIG. 9 includes a CPU (central processing unit) 901, a memory 902, an input device 903, an output device 904, an external storage device 905, a medium driving device 906, and a network connection device 907, which are connected via a bus 908. Are connected to each other.

  The memory 902 includes, for example, a read only memory (ROM), a random access memory (RAM), and the like, and stores programs and data used for processing. The CPU 901 performs the above-described ruled line extraction process by executing a program using the memory 902.

  In this case, the input color image is stored in the memory 902 as processing target data, and a lightness gray image, a saturation gray image, a saturation correction image, and their binary images are generated in the memory 902.

  The input device 903 is, for example, a keyboard, a pointing device, and the like, and is used for inputting instructions and information from an operator. The output device 904 is, for example, a display, a printer, a speaker, or the like, and is used to output an inquiry to an operator or a processing result.

  The external storage device 905 is, for example, a magnetic disk device, an optical disk device, a magneto-optical disk device, a tape device, or the like. The information processing apparatus stores programs and data in the external storage device 905, and loads them into the memory 902 and uses them as necessary.

  The medium driving device 906 drives a portable recording medium 909 and accesses the recorded contents. The portable recording medium 909 is an arbitrary computer-readable recording medium such as a memory card, a flexible disk, an optical disk, and a magneto-optical disk. The operator stores programs and data in the portable recording medium 909 and loads them into the memory 902 and uses them as necessary.

  The network connection device 907 is connected to a communication network such as a local area network (LAN) and performs data conversion accompanying communication. Further, the information processing apparatus receives programs and data from an external apparatus via the network connection apparatus 907 as necessary, and loads them into the memory 902 for use.

  FIG. 10 shows a method for providing a program and data to the information processing apparatus of FIG. Programs and data stored in the portable recording medium 909 and the database 1011 of the server 1001 are loaded into the memory 902 of the information processing apparatus 1002. The server 1001 generates a carrier signal that carries the program and data, and transmits the carrier signal to the information processing apparatus 1002 via any transmission medium on the communication network. The CPU 901 executes the program using the data and performs the above-described processing.

(Supplementary note 1) storage means for storing a color image;
First generation means for generating a lightness gray image from the color image and storing it in the storage means;
First extraction means for extracting a first ruled line using the lightness gray image;
Second generation means for generating a saturation gray image of the cell region surrounded by the first ruled line and storing it in the storage means;
Second extraction means for extracting a second ruled line using the saturation gray image;
A ruled line extraction apparatus comprising: output means for outputting information on the first and second ruled lines.
(Appendix 2) A color image is stored in a storage means,
A lightness gray image is generated from the color image and stored in the storage means;
Extracting a first ruled line using the lightness gray image;
A saturation gray image of the cell region surrounded by the first ruled line is generated and stored in the storage means;
Extracting a second ruled line using the saturation gray image;
A program for causing a computer to execute a process of outputting information on the first and second ruled lines.
(Supplementary Note 3) The lightness gray image is binarized, the obtained first binary image is stored in the storage means, and the first ruled line is extracted using the first binary image, Binarizing the saturation gray image, storing the obtained second binary image in the storage means, and extracting the second ruled line using the second binary image, The program according to appendix 2, wherein the program is executed by a computer.
(Supplementary Note 4) A saturation correction image of a cell region surrounded by the first or second ruled line is generated, stored in the storage means, and a third ruled line is extracted using the saturation correction image. The program according to claim 2 or 3, further causing the computer to execute a process of outputting information of the third ruled line.
(Supplementary Note 5) The saturation of the cell region surrounded by the second ruled line is compared with a threshold value, and if the saturation is larger than the threshold value, an operation for increasing the saturation is performed. The program according to appendix 4, wherein if it is smaller, the computer is caused to execute the process of generating the saturation correction image by performing an operation for reducing the saturation.
(Additional remark 6) A 1st production | generation means produces | generates the brightness gray image from the color image stored in the storage means, and stores it in this storage means,
A first extracting means for extracting a first ruled line using the lightness gray image;
A second generation unit generates a saturation gray image of the cell region surrounded by the first ruled line, and stores the saturation gray image in the storage unit;
A second extracting means for extracting a second ruled line using the saturation gray image;
A ruled line extraction method, wherein the output means outputs information of the first and second ruled lines.

It is a principle figure of the ruled line extraction apparatus of this invention. It is a flowchart of the ruled line extraction process of this invention. It is a figure which shows the ruled line with a brightness difference with a background. It is a figure which shows the ruled line with a background and saturation difference. It is a figure which shows the ruled line with a background and a hue difference. It is a figure which shows the saturation correction method. It is a figure which shows the 1st ruled line extraction process. It is a figure which shows the 2nd ruled line extraction process. It is a block diagram of information processing apparatus. It is a figure which shows the provision method of a program and data. It is a flowchart of the conventional ruled line extraction process. It is a figure which shows a boundary ruled line. It is a figure which shows the conventional ruled line extraction process.

Explanation of symbols

21, 301, 401, 501, 701, 801 Color image 22, 304, 404, 504, 702, 802 Light gray image 23, 704, 706 Binary image 24, 803, 806, 807, 811, 812, 813 Extraction result DESCRIPTION OF SYMBOLS 101 Storage means 102 1st production | generation means 103 1st extraction means 104 2nd production | generation means 105 2nd extraction means 106 Output means 302, 402, 502 Hue gray image 303, 403, 503, 703 Saturation gray image 601 Sigmoid curve 610 Graph 620 Line 705 Saturation correction image 711 Region 712, 821, 822, 823, 824, 825, 826, 827, 828 Cell 901 CPU
902 Memory 903 Input device 904 Output device 905 External storage device 906 Medium drive device 907 Network connection device 908 Bus 909 Portable recording medium 1001 Server 1002 Information processing device 1011 Database

Claims (5)

Storage means for storing color images;
First generation means for generating a lightness gray image from the color image and storing it in the storage means;
First extraction means for extracting a first ruled line using the lightness gray image;
Second generation means for generating a saturation gray image of the cell region surrounded by the first ruled line and storing it in the storage means;
Second extraction means for extracting a second ruled line using the saturation gray image;
A ruled line extraction apparatus comprising: output means for outputting information on the first and second ruled lines. Store the color image in the storage means,
A lightness gray image is generated from the color image and stored in the storage means;
Extracting a first ruled line using the lightness gray image;
A saturation gray image of the cell region surrounded by the first ruled line is generated and stored in the storage means;
Extracting a second ruled line using the saturation gray image;
A program for causing a computer to execute a process of outputting information on the first and second ruled lines.   A saturation correction image of the cell region surrounded by the first or second ruled line is generated, stored in the storage means, a third ruled line is extracted using the saturation correction image, and the third The program according to claim 2, further causing the computer to execute a process of outputting the ruled line information.   Comparing the saturation of the cell region surrounded by the second ruled line with a threshold value, if the saturation is larger than the threshold, an operation for increasing the saturation is performed, and if the saturation is smaller than the threshold, The program according to claim 3, wherein the computer is caused to execute the process of generating the saturation correction image by performing an operation for reducing the saturation. The first generation means generates a lightness gray image from the color image stored in the storage means, stores the lightness gray image in the storage means,
A first extracting means for extracting a first ruled line using the lightness gray image;
A second generation unit generates a saturation gray image of the cell region surrounded by the first ruled line, and stores the saturation gray image in the storage unit;
A second extracting means for extracting a second ruled line using the saturation gray image;
A ruled line extraction method, wherein the output means outputs information of the first and second ruled lines.