JP3734614B2 - Image processing method, apparatus, and recording medium - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing method , apparatus, and recording medium with improved ruled line identification accuracy.
[0002]
[Prior art]
As a method for recognizing dotted ruled lines constituting a table, for example, there is a ruled line recognition method described in JP-A-7-230525. In this method, a rectangle having a predetermined threshold value or less is extracted as a dotted line element, and a rectangle obtained by integrating elements whose intervals between the elements are within a predetermined threshold value is extracted as a dotted line. Also, the validity of the ruled line is determined based on the respective variance values of the size of the rectangle and the interval between the rectangles.
[0003]
By the way, when a ruled line or character is recognized, if the characteristics of the image to be processed are, for example, in a dark state or a faint state, the algorithm or threshold value used in the recognition process cannot handle it. Recognition accuracy may be reduced. In addition, by changing the binarization threshold value when inputting with an input device such as a scanner, the image characteristics can be corrected by correcting the density of the input image and obtaining image characteristics adapted to the recognition algorithm. It is.
[0004]
[Problems to be solved by the invention]
In the method described in the above publication, if a rectangle extraction process is performed with a predetermined threshold, for example, a rectangle constituting a character may be erroneously extracted as a broken line component. In addition, it is very time-consuming to input an image again because binary data once input has characteristics that do not match the algorithm.
[0005]
Therefore, if the characteristics of the processed image can be grasped and used as pre-processing for the recognition processing of ruled lines and characters, it is possible to change to an algorithm or threshold corresponding to the information, and to recognize it. The accuracy can be improved.
[0006]
The present invention has been made based on the above consideration,
It is an object of the present invention to improve ruled line and character identification accuracy by determining a blurred image, which is one of the characteristics of an input image, and extracting a blurred area as a pre-process for ruled line and character recognition processing. Another object of the present invention is to provide an image processing method , apparatus, and recording medium.
[0007]
[Means for Solving the Problems]
In order to achieve the above object, the invention according to claim 1 extracts a black pixel continuous component rectangle from the binarized image data, extracts a rectangle corresponding to a dotted line element from the extracted rectangle, With respect to the extracted dotted line elements, the dotted line elements within a predetermined distance are combined, and as a result of the processing, the dotted line elements that are not combined are counted, and when the count value is equal to or greater than a predetermined value, the image is It is characterized by determining that the image is a blurred image.
[0008]
In the invention of claim 2, a rectangle of black pixel continuous components is extracted from the binarized image data, a rectangle corresponding to a dotted line element is extracted from the extracted rectangle, and the extracted dotted line element is A dotted line element within a predetermined distance is combined, and the dotted line elements that are not combined as a result of the processing are counted. When the counted value is equal to or greater than a predetermined value, the image is determined to be a blurred image; For the determined faint image, the blurred area is extracted by integrating the dotted line elements that are not combined.
[0009]
The invention according to claim 3 is characterized in that a dotted line existing in the blurred area is removed as a pseudo dotted line.
[0010]
According to a fourth aspect of the present invention, when the size of the blurred area is a predetermined value or more, a warning is given that the image is a blurred image.
[0011]
According to the fifth aspect of the present invention, means for extracting a black pixel continuous component rectangle from the binarized image data, means for extracting a rectangle corresponding to a dotted line element from the extracted rectangle, and the extracted With respect to the dotted line element, means for combining dotted line elements within a predetermined distance, means for counting dotted line elements that have not been combined as a result of the combining process, and when the count value is greater than or equal to a predetermined value, the image And a means for determining that the image is a faint image .
[0012]
In the invention described in claim 6, means for extracting a black pixel continuous component rectangle from the binarized image data, means for extracting a rectangle corresponding to a dotted line element from the extracted rectangle, and the extracted With respect to the dotted line element, means for combining dotted line elements within a predetermined distance, means for counting dotted line elements that have not been combined as a result of the combining process, and when the count value is greater than or equal to a predetermined value, the image It is characterized by comprising means for determining that the image is a faint image, and means for extracting the blurred region by integrating the dotted line elements that have not been combined with the determined faint image .
[0013]
The invention described in claim 7 is a computer-readable recording medium that records a program for causing a computer to implement the image processing method according to any one of claims 1 to 4 .
[0014]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.
Before describing each embodiment of the present invention, terms used in the present invention are first defined.
Rectangle: Consecutive images in an image, or image portions that are continuous for a predetermined threshold or more (for example, a continuous black pixel portion or a continuous white pixel portion in the case of a binary image) The range enclosed by the circumscribed rectangle is defined as a rectangle so that is included in contact.
[0015]
Rectangle extraction: Extracting the position coordinates of a rectangle is defined as rectangle extraction.
Single element dotted line: A dotted line having one dotted line element, that is, synonymous with a dotted line element that cannot be combined.
Dotted line: A combination of dotted line elements.
[0016]
<Example 1>
When determining whether or not the image is a blurred image based on the binary input data, a characteristic of the blurred image is that the ruled lines and characters are blurred and discontinuous, and thus the number of rectangles is large. By the way, even if an image includes dotted lines and halftone dots, the number of rectangles increases, and simply counting the number of rectangles makes it possible to determine whether the image is a blurred image or a clear image including many dotted lines. Can not. Therefore, if the dotted line element rectangle and the rectangle due to the blur can be distinguished, it can be determined that the image is a blur image when the number of rectangles due to the blur is larger than the threshold value.
[0017]
Therefore, in the first embodiment, using the feature that the number of rectangles in the blurred image is large, rectangles that are discontinuous due to blurring are distinguished from dotted line elements and halftone dot element rectangles, and the number of rectangles generated by blurring is counted. Then, based on the count value, it is determined whether the image is a blurred image.
[0018]
FIG. 1 shows the configuration of Embodiment 1 of the present invention. FIG. 2 shows a process flowchart of the first embodiment of the present invention. An image such as a document is input by the binary image input unit 1 such as a scanner and stored in the binary image memory 2 (step 101). The rectangle extraction unit 3 extracts the rectangle of the black pixel continuous component from the binary image memory 2 and stores it in the rectangle memory 4 (step 102).
[0019]
The dotted line extraction unit 5 includes a dotted line element selection unit 6 and a dotted line element combination processing unit 8. The dotted line element selection unit 6 selects a rectangle having a size corresponding to a dotted line element from the rectangles held in the rectangular memory 4, It is stored in the dotted line element memory 7 (step 103).
[0020]
Subsequently, the dotted line element combination processing unit 8 combines the dotted line elements located within a certain value in the processing direction. That is, when connecting the dotted line elements arranged in the horizontal direction, the dotted line elements existing within a certain value in the horizontal direction (processing direction) are combined. The dotted lines combined and extracted are held in the ruled line memory 9 (steps 104 and 109). Also, the dotted line elements that have not been combined with two or more dotted line elements by the dotted line element combination processing unit 8 are held in the ruled line memory 9 as single element dotted lines (steps 104 and 105).
[0021]
A single element dotted line is a dotted line that could not be combined by the dotted line element combination processing, that is, a rectangle selected as the dotted line element size by the dotted line element selection unit 6, but determined not to be a dotted line element rectangle , Consider a rectangle with blur. Here, the number of single element dotted lines can be assumed to be the number of rectangles generated by blurring. The faint document determination unit 10 determines a faint document as follows. The faint document determination unit 10 includes a counter 11 that counts a single element dotted line, a marking unit 12 that gives a marking to a faint document, and a document type memory 13.
[0022]
As a result of the counter 11 counting the single element dotted lines (step 106), if the number of single element dotted lines is equal to or greater than a predetermined number, the image being processed is determined to be a blurred image (steps 107 and 108). Information indicating that the input image is a faint image is added and stored in the document type memory 13.
[0023]
<Example 2>
In the first embodiment described above, information regarding which part of the image is blurred in the image determined to be a blurred image cannot be obtained. If the blurred area in the image is known, it is possible to provide information that is effective for subsequent processing, for example, processing for executing the blurred correction process on the extracted blurred area.
[0024]
Therefore, this embodiment is an embodiment in which information that is effective for other processing is provided by extracting a blurred region in an image.
[0025]
FIG. 3 shows a configuration of the second embodiment of the present invention, and FIG. 4 shows a process flowchart of the second embodiment of the present invention. The configuration up to the dotted line element combination processing unit 28 and the ruled line memory 29 is the same as in the first embodiment. In the present embodiment, a single element dotted line coupling unit 31, a blurred region extraction processing unit 30 including a blurred region integration unit 32, and a blurred region memory 33 are provided.
[0026]
The single element dotted line combining unit 31 combines the single element dotted lines that are located within a predetermined range with the single element dotted lines held in the ruled line memory 29 to thereby reduce the faint area. Then, the faint areas are extracted by causing the faint area integration unit 32 to similarly bond and grow the faint areas (step 207).
[0027]
5, 6 and 7 are diagrams for explaining the generation of the blurred area. In FIGS. 5, 6, and 7, A, B, and C are single element dotted lines, which are dotted line elements that cannot be combined with two or more dotted line elements. First, assuming that A in FIG. 5 is a reference for processing, a range that is widened by an arbitrary value (width) centering on A is defined as A, and the range A is a processing range that is combined with A. Further, in FIG. 5, B exists within the processing range of A, so A and B are combined to form a faint area α as shown in FIG.
[0028]
After the faint area is extracted, the combined range a is set as shown in FIG. 6 with the faint area α as a reference. Since C exists in the set coupling range, when C and the faint area are coupled in the same manner as described above, a faint area α is formed as shown in FIG. Such processing is repeated until there is no single element dotted line within the combined range. When there is no single element dotted line within the combined range, the faint area extraction process based on A is completed.
[0029]
The single element dotted line coupling unit 31 executes the above processing for all single element dotted lines. The extracted faint area is held in the faint area memory 33. Next, the faint area integration unit 32 executes an integration process on the small faint areas generated by the single element dotted line coupling unit 31. Based on the small blurred area data held in the blurred area memory 33, the small blurred areas are combined as in the processing of the combining unit 31. That is, a process of replacing the single element dotted line combining process in the combining unit 31 with a faint area is performed. The blurred area integrated here is held in the blurred area memory 33 again.
[0030]
<Example 3>
If the blurred area process of the second embodiment is executed for all images, it takes a long time. In this embodiment, therefore, the blur area extraction process of the second embodiment is executed only for the image determined to be a blur image in the first embodiment, and the processing time is shortened.
[0031]
FIG. 8 shows a configuration of the third embodiment. FIG. 9 shows a process flowchart of the third embodiment. The configuration of the third embodiment is a combination of the configurations of the first and second embodiments. In other words, since the type of image being processed is held in the document type memory 48 by the processing of the first embodiment, when information indicating that the image is faint is obtained when the memory 48 is referred to (step 50). 308), the blurred area extraction process 49 of the second embodiment is executed (step 309).
[0032]
<Example 4>
In the dotted line extraction process, a dotted line is formed by extracting a dotted line element size rectangle and combining the dotted line element rectangles positioned within the threshold value. However, a false dotted line is generated by erroneously combining discontinuous rectangles in a blurred area in the image.
[0033]
Therefore, in this embodiment, by extracting a faint area in an image by a simple method, the dotted line extracted in the faint area is regarded as a pseudo dotted line and removed from the dotted line information, and the accuracy of the dotted line recognition is improved. It is an example.
[0034]
FIG. 10 shows the configuration of the fourth embodiment, in which a pseudo dotted line removing unit 51 is further added to the configuration of the third embodiment, and other components are the same as those of the third embodiment. FIG. 11 shows a process flowchart of the fourth embodiment. The blurred area extracted in the second embodiment is held in the blurred area memory 50, and the dotted line extracted by the dotted line extraction unit 45 is held in the ruled line memory 46.
[0035]
The pseudo dotted line removing unit 51 compares the position coordinates of the dotted line in the ruled line memory 46 with the blurred area in the blurred area memory 50 and is in contact with the dotted line (pseudo dotted line) or the blurred area existing in the blurred area. The dotted line (pseudo dotted line) is removed from the ruled line memory 46 (step 410).
[0036]
<Example 5>
In order to improve the recognition accuracy of ruled lines and characters in an image with a remarkable degree of fading, a method of correcting the recognition processing algorithm can be considered, but a required time is required until the algorithm is developed.
[0037]
Therefore, in the following fifth and sixth embodiments, in order to deal with a problem of a blurred image, in an image in which the number or area of the blurred areas extracted in the second embodiment is larger than a predetermined value, the user is blurred. Processing to warn that the image is present. The user inputs the image again for the image for which the warning is given or corrects the image.
[0038]
FIG. 12 shows the configuration of the fifth embodiment. When the document type memory 48 is referred to and the information indicating that the document is a blurred document is obtained, the blurred image warning unit 52 issues a warning that the document is a blurred image. For example, the user is notified using a method such as displaying a warning on the display screen or notifying by voice.
[0039]
<Example 6>
FIG. 13 shows the configuration of the sixth embodiment. Since the blurred area extracted in the second embodiment is held in the blurred area memory 50, the blurred area calculation unit 53 calculates the total area of the blurred areas or the total number of blurred areas from the blurred area memory 50. When the predetermined threshold value is exceeded, the blurred image warning unit 52 warns that the image is a blurred image, as in the fifth embodiment.
[0040]
<Example 7>
FIG. 14 shows the configuration of the seventh embodiment and is an embodiment realized by software. When the present invention is realized by software, a computer system including a CPU, a memory, a display device, a hard disk, a keyboard, a CD-ROM drive, a mouse, and the like is prepared as shown in FIG. A computer-readable recording medium such as a CD-ROM records a program for realizing the image processing function and processing procedure of the present invention. An image such as a document to be processed is stored in, for example, a hard disk. Then, the CPU reads a program for realizing the processing functions and processing procedures described above from the recording medium, determines whether or not the image read from the hard disk or the like is a faint image, and displays the determination result on a display or the like. To do.
[0041]
【The invention's effect】
As described above, according to the present invention , it is determined whether or not the input image is faint as pre-processing for ruled line and character recognition, so that the accuracy of recognition of ruled lines and characters as subsequent processing is improved. Can give useful information for.
[0042]
According to the present invention , since the blurred area in the input image is extracted, the recognition accuracy of ruled lines and characters can be further improved by using the area information.
[0043]
According to the present invention , the processing time is shortened because the processing for extracting the blurred region is performed only on the image determined to be a blurred image.
[0044]
According to the present invention , since there is a high probability that the dotted line existing in the blurred area is a pseudo-dotted line, by removing the dotted line extracted in the blurred area, the removal rate of the pseudo-dotted line in the image is improved, and therefore the dotted line. The accuracy of ruled line extraction is improved.
[0045]
According to the present invention , it is possible to alert the user that the input image is a faint image, and the user can execute an appropriate treatment on the faint image.
[Brief description of the drawings]
FIG. 1 shows a configuration of Embodiment 1 of the present invention.
FIG. 2 shows a processing flowchart of Embodiment 1 of the present invention.
FIG. 3 shows a configuration of Embodiment 2 of the present invention.
FIG. 4 shows a process flowchart of Embodiment 2 of the present invention.
FIG. 5 is a first diagram for explaining fading area generation;
FIG. 6 is a second diagram for explaining the generation of the blurred area.
FIG. 7 is a third diagram for explaining fading region generation.
FIG. 8 shows a configuration of Embodiment 3 of the present invention.
FIG. 9 shows a process flowchart of Embodiment 3 of the present invention.
FIG. 10 shows a configuration of Example 4 of the present invention.
FIG. 11 shows a process flowchart of Embodiment 4 of the present invention.
FIG. 12 shows a configuration of Example 5 of the present invention.
FIG. 13 shows a configuration of Embodiment 6 of the present invention.
FIG. 14 shows a configuration of Example 7 of the present invention.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Image input part 2 Binary image memory 3 Rectangle extraction part 4 Rectangular memory 5 Dotted line extraction part 6 Dotted line element selection part 7 Dotted line element memory 8 Joining process part 9 Ruled line memory 10 Blurred original determination part 11 Counter 12 Marking part 13 Original type memory

Claims (7)

  A black pixel continuous component rectangle is extracted from the binarized image data, a rectangle corresponding to a dotted line element is extracted from the extracted rectangle, and a dotted line element within a predetermined distance is extracted from the extracted dotted line element. The image processing method is characterized in that the combination processing is performed, the dotted line elements that are not combined as a result of the processing are counted, and the image is determined to be a blurred image when the count value is equal to or greater than a predetermined value.   A black pixel continuous component rectangle is extracted from the binarized image data, a rectangle corresponding to a dotted line element is extracted from the extracted rectangle, and a dotted line element within a predetermined distance is extracted from the extracted dotted line element. And the dotted line elements that are not combined as a result of the processing are counted, and when the counted value is equal to or greater than a predetermined value, it is determined that the image is a blurred image, and for the determined blurred image, An image processing method, wherein a blurred area is extracted by integrating the dotted line elements that are not combined. The image processing method according to claim 2, wherein a dotted line existing in the blurred area is removed as a pseudo dotted line. 3. The image processing method according to claim 2 , wherein when the size of the blurred area is equal to or larger than a predetermined value, a warning is given that the image is a blurred image. Means for extracting a black pixel continuous component rectangle from the binarized image data; means for extracting a rectangle corresponding to a dotted line element from the extracted rectangle; and for the extracted dotted line element within a predetermined distance. Means for combining the dotted line elements at the position, means for counting the dotted line elements that have not been combined as a result of the combination process, and determining that the image is a blurred image when the count value is equal to or greater than a predetermined value. And an image processing apparatus. Means for extracting a black pixel continuous component rectangle from the binarized image data; means for extracting a rectangle corresponding to a dotted line element from the extracted rectangle; and for the extracted dotted line element within a predetermined distance. Means for combining the dotted line elements at the position, means for counting the dotted line elements that have not been combined as a result of the combination process, and determining that the image is a blurred image when the count value is equal to or greater than a predetermined value. An image processing apparatus comprising: means for extracting a blurred area by integrating the dotted line elements that have not been combined with the determined blurred image. A computer-readable recording medium on which a program for causing a computer to realize the image processing method according to claim 1 is recorded.

Images