Linear line detection apparatus using projection image of character string including linear line

• US 5,822,455 A
• Filed: 12/12/1995
• Issued: 10/13/1998
• Est. Priority Date: 12/12/1994
• Status: Expired due to Fees

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Sano, Tikara, Nakamura, Yoshikatu
• Primary Examiner(s): Couso, Yon J.
• Assistant Examiner(s): NGUYEN, HA T
• Application Number:
  US08/570,973
• Time in Patent Office: 1,036 Days
• Field of Search: 382/170, 382/171, 382/172, 382/174, 382/177, 382/179, 382/202, 382/101, 382/102
• US Class Current: 382/202
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/155 Removing patterns interfering with the pattern to be recognised, such as ruled lines or underlines

Abstract

A line image detection apparatus having an image input function for inputting an image on a document. The image includes a rectangular character string and a linear line drawn along the character string. The line detection apparatus has a function for detecting a character string image from the image, a function for forming a projection image on the basis of the distance from a lower end of the character string image to each first black pixel found in the character string image in the vertical direction, and a function for comparing the distance from the lower end of the character string image to each first black pixel, with a predetermined threshold value, thereby to define a plurality of "slice 0" areas included in the projection image. The distance, in each of the "slice 0" areas, from the lower end of the character string image to each first black pixel is lower than the predetermined threshold value. The apparatus also includes a function for determining an interval distribution of the interval between each adjacent pair of the "slice 0" areas determined by an area determination means, and a function for detecting the linear line on the basis of the interval distribution.

Claims

1. A line image detection apparatus comprising:

means for imaging at least a portion of a document to form an input image, the input image including an image of a character string and a straight line being under the character string, the image of the character string and the straight line being constituted by black pixels;

means for detecting and recognizing, from the input image, the image of the character string and the straight line, the image of the character string and the straight line being located in an area of the input image being between a lower end and an upper end of the input image;

means for forming a projection image based on a distance from the lower end of the input image to each first black pixel found in the input image in a vertical direction;

means for comparing the distance with a predetermined threshold value;

means for defining a plurality of "slice 0" areas having the distance being lower than the predetermined threshold value;

means for determining an interval distribution of a corresponding interval between each adjacent pair of the "slice 0" areas determined by the defining means; and

means for detecting the straight line in the input image based on the interval distribution.

2. The line image detection apparatus according to claim 1, wherein the interval distribution determining means includes:

means for determining a center point of each of the "slice 0" areas; and

means for detecting an interval between each adjacent pair of the center points.

3. The line image detection apparatus according to claim 1, wherein the straight line detecting means includes:

means for preparing an interval histogram concerning the corresponding intervals between each of the adjacent pair of the "slice 0" areas; and

means for determining that the detected straight line is a broken line, when in the interval histogram, a ratio of a frequency of one interval value to a sum of frequencies of all interval values is higher than a predetermined value.

4. The line image detection apparatus according to claim 2, wherein the straight line detecting means includes means for determining that the detected straight line is a broken line, when a ratio of a sum of frequencies of two interval values, corresponding to adjacent intervals between adjacent pairs of the center points, to a sum of frequencies of all interval values is higher than a predetermined value.

5. The line image detection apparatus according to claim 1, wherein the straight line detecting means includes:

means for preparing an interval histogram concerning the interval between each of the adjacent pair of the "slice 0" areas; and

means for determining that the detected straight line is a broken line, when in the interval histogram, frequency of one interval value is 80% or more of a sum of frequencies of all interval values.

6. The line image detection apparatus according to claim 1, further comprising:

means for obtaining the distance from the lower end of the input image to each first black pixel, and preparing a distance histogram concerning the distance; and

means for determining that the detected straight line is a solid line, when in the distance histogram, a ratio of a frequency of one distance value to a sum of frequencies of all distance values is higher than a predetermined value.

7. The line image detection apparatus according to claim 6, wherein the solid line determinating means includes means for determining that the detected straight line is a solid line, when in the distance histogram, a ratio of a sum of frequencies of two continuous distance values to the sum of the frequencies of all distance values is higher than a predetermined value.

8. The line image detection apparatus according to claim 6, wherein the solid line determinating means includes means for determining that the detected straight line is a solid line, when in the distance histogram, the frequency of one distance value is 90% or more of the sum of the frequencies of all distance values.

9. The line image detection apparatus according to claim 8, wherein the solid line determinating means includes means for determining that the detected straight line is a solid line, when in the distance histogram, the sum of the frequencies of two continuous distance values is 90% or more of the sum of the frequencies of all distance values.

10. The line image detection apparatus according to claim 1, further comprising:

means for preparing a straight line width histogram concerning a width of each portion of the straight line in the projection image, when the straight line detection means has detected the straight line; and

removal means for detecting a position and the width of each portion of the straight line based on the straight line width histogram so as to remove the straight line.

11. The line image detection apparatus according to claim 10, further comprising means for recognizing characters in the input image from which the straight line is removed by the removal means.

12. The line image detection apparatus according to claim 1, wherein the straight line detecting means includes:

means for forming a second projection image based on a second distance from the upper end of the input image to each first black pixel found in the input image;

means for comparing the second distance with a second predetermined threshold value, so as to define a plurality of second "slice 0" areas having the second distance being lower than the second predetermined threshold value;

means for determining a center point of each of the second "slice 0" areas;

means for detecting a second interval between each adjacent pair of the center points of the second "slice 0" areas;

means for preparing a first interval histogram concerning the intervals of the "slice 0" areas;

means for preparing a second interval histogram concerning the second intervals; and

means for determining that the detected straight line is a broken line, when a ratio of a frequency of a first interval value which is highest in the first interval histogram, to a frequency of all interval values is higher than a predetermined value, and when the first interval value is lower than a second interval value with a highest frequency in the second interval histogram.

13. A line image detection apparatus comprising:

means for imaging at least a portion of a document to form an input image, the input image including an image of a character string and a straight line being under the character string, the image of the character string and the straight line being constituted by black pixels;

means for detecting and recognizing, from the input image, the image of the character string and the straight;

means for obtaining a distance from each of a plurality of points of a lower end of the input image, to each first black pixel in a vertical direction, and preparing a distance histogram concerning the distance; and

means for determining that the detected straight line is a solid line, when in the distance histogram, a ratio of a frequency of a distance value, which is highest, to the sum of frequencies of all distance values higher than a predetermined value.

14. The line image detection apparatus according to claim 13, wherein the solid line determining means includes means for determining that the detected straight line is a solid line, when in the distance histogram, the frequency of the distance value, which is highest, is 90% or more of the sum of the frequencies of all distance values.

15. The line image detection apparatus according to claim 13, further comprising:

means for preparing a solid line width histogram concerning a width of each portion of a solid line in the input image, when the solid line determining means has determined that the solid line exists; and

solid line removal means for detecting a position and the width of each portion of the solid line based on the solid line width histogram, and removing the solid line.

16. The line image detection apparatus according to claim 15, further comprising means for recognizing characters in the input image from which the solid line is removed by the solid line removal means.

17. A line image detection apparatus comprising:

means for imaging at least a portion of a document to form an input image, the input image including an image of a character string and a straight line being under the character string, the image of the character string and the straight line being constituted by black pixels;

means for detecting and recognizing, from the input image, an image of the character string and the straight line;

means for projecting only a first area and a second area, the first area being an area of a left portion of the input image including only a portion of the straight line up to a left and of the input image, the second area being an area of a right portion of the input image and including only another portion of the straight line up to a right end of the input image, so as to form projection images;

means for comparing a vertical width of projection data of each of the first and the second areas with a vertical width of the image of the character string;

means for determining that the input image includes the straight line, when a ratio of the vertical width of the projection data to the vertical width of the image of the character string is lower than a predetermined value;

means for detecting an inclination of the straight line based on the projection data, when the means for determining that the input image includes the straight line has determined that the straight line exists; and

means for correcting the inclination of the straight line.

18. The line image detection apparatus according to claim 17, further comprising:

means for detecting a position and a width of the straight line included in the input image based on the projection data; and

means for removing the straight line when the detected width is narrower than a predetermined width.

19. The line image detection apparatus according to claim 18, further comprising:

means for recognizing each character included in the input image with the straight line removed by the straight line removing means.

20. A method of detecting a line comprising:

imaging at least a portion of a document, including a character string and a straight line being under the character string, to form an input image;

the character string and the straight line in the input image being constituted by black pixels;

detecting and recognizing, from the input image, an image of the character string and the straight line, the character string and the straight line being located in an area of the input image being between a lower end and an upper end of the input image;

forming a projection image based on a distance from the lower end of the input image to each first black pixel found in the input image in a vertical direction;

comparing the distance with a predetermined threshold value;

defining a plurality of "slice 0" areas having the distance being lower than the predetermined threshold value;

determining and interval distribution of a corresponding interval between each adjacent pair of the "slice 0" areas determined by the defining step; and

detecting the straight line in the input image based on the interval distribution.

21. A method of detecting a line comprising:

imaging at least a portion of a document, including a character string and a straight line being under the character string, to form an input image;

the character string and the straight line in the input image being constituted by black pixels;

detecting and recognizing, from the input image, an image of the character string and the straight line;

obtaining a distance from each of a plurality of points of a lower end of the input image to each first black pixel in a vertical direction, and preparing a distance histogram concerning the distance; and

determining that the detected straight line is a solid line, when in the distance histogram, a ratio of a frequency of a distance value, which is highest, to the sum of frequencies of all distance values is higher than a predetermined value.

22. The method of detecting a line according to claim 21, wherein the step for determining that the detected straight line is the solid line comprises determining that the detected straight line is a solid line, when in the distance histogram, the frequency of the distance value, which is highest, is 90% or more of the sum of the frequencies of all distance values.

23. The method of detecting a line according to claims 21, further comprising:

preparing a solid line width histogram concerning a width of each portion of a solid line in the input image, when the step of determining that the detected straight line is the solid line has determined that the solid line exists; and

detecting a position and the width of each portion of the solid line based on the solid line width histogram and removing the solid line.

24. The method of detecting a line according to claim 23, further comprising:

recognizing characters in the input image from which the solid line is removed in the step of detecting the position and the width of each portion of the solid line.

25. A method of detecting a line comprising:

imaging at least a portion of a document, including a character string and a straight line being under the character string, to form an input image;

the character string and the straight line in the input image being constituted by black pixels;

detecting and recognizing, from the input image, an image of the character string and the straight line;

projecting only a first area and a second area, the first area being an area of a left portion of the input image and includes only a portion of the straight line up to a left end of the input image, the second area being an area of a right portion of the input image and includes only another portion of the straight line up to a right end of the input image, so as to form projection images;

comparing a vertical width of projection data of each of the first and the second areas with a vertical width of the image of the character string;

determining that the input image includes the straight line when a ratio of the vertical width of the projection data to the vertical width of the image of the character string is lower than a predetermined value;

detecting an inclination of the straight line based on the protection data, when the determining step determines that the input image includes the straight line; and

correcting the inclination of the straight line.