JPH07192085A - Document picture inclination detector - Google Patents

Abstract

PURPOSE:To reduce the troublesomeness of processing document without lowering the accuracy at the time of detecting the inclined angle of document pictures. CONSTITUTION:This detector is constituted of a picture input means 1, a means 2 for referring to inputted document picture data and detecting the connection components of picture elements, the means 3 for extracting the connection components for constituting characters among the extracted connection components of the picture elements, the means 4 for connecting the adjacent connection components each other among the extracted connection components and outputting them as tentative character rows, the means 5 for detecting an outermost contour point in contact with a straight line inclined by respective predetermined angles from directions horizontal and vertical to the estimated tentative character row, the means 6 for calculating the inclination of the tentative character row from the position relation of the detected angle points and the means 7 for referring to the inclined angle from the respective tentative character rows and obtaining the inclined angle of the document picture data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a device for detecting the inclination of a document image.

[0002]

2. Description of the Related Art In order to read a general printed document, it is first necessary to capture it as a quantized document image by using an image input device such as an image scanner. Occurs. Therefore, in the character line extraction method using projection, there is a problem that the character lines cannot be detected in a document having a narrow character line interval, and a process for detecting the inclination of the input image is required. ing.

Conventionally, in such a document image, the inclination is detected by utilizing the fact that the characters are regularly arranged in the document area, which is a main component thereof.

For example, Akiyama et al., Entitled "Paper Component Extraction Method Not Relying on Format Designation Information", has been published in the journal of electronic telecommunications science D, vol. J66-D, No. 1, pp. 111-
As described in 118, a first method has been proposed in which a histogram is created in the angle θ direction while sequentially changing the tilt angle θ, and the angle θ at which peaks and valleys on the histogram are prominent is obtained as the tilt angle. ing.

Also, Nakano et al., Entitled "One Method for Correcting Tilt of Document Image," D, vo
l. J69-D, No. 1, pp. 1833-1834
As described in, paying attention to the fact that the reference line of the character string is almost constant, the slope of the character string is detected by Hough transforming the lower end coordinate value of the character block and detecting the peak value in the Hough space. A second method for estimating is proposed.

[0006]

However, in the first method, it is necessary to calculate a histogram for each pixel, and there is a problem in terms of processing time. Further, in the second method, since the lower ends of all the characters are not always on the reference line, it is inevitable that an error is included. Further, the Hough transform process requires the same or more process steps as the process of calculating the histogram.

In order to solve these problems, an object of the present invention is to propose an inclination detecting device which requires a small number of processing steps without degrading accuracy.

[0008]

In order to solve the above problems, the present invention provides an image inputting means for optically scanning a document to obtain document image data, and a document image input by the image inputting means. Connected component extracting means for detecting connected components of pixels by referring to data; character component extracting means for extracting connected components that form a character among connected components of pixels extracted by the connected component extracting means; Of the connected components extracted by the character component extraction means, the adjacent connected components are integrated and output as a temporary character line, and the character line estimation means and the temporary character line estimated by the character line estimation means An angle point detecting means for detecting the outermost contour points that are in contact with straight lines inclined by a predetermined angle from the horizontal direction and the vertical direction, respectively, and an angle detected by the angle point detecting means. From the positional relationship of the above, the inclination of the tentative character line is calculated by the character line inclination angle calculating means, and the tentative character lines obtained from the character line estimating means are obtained by the character line inclination angle calculating means. And a document image inclination angle detecting means for obtaining the inclination angle of the document image data with reference to the obtained inclination angle.

[0009]

In the present invention, the processing of using the coordinate value obtained from each character is not performed. First, as shown in FIG. 1A, connected pixel components that form individual characters are extracted. Next, as shown in FIG. 1B, the connected pixel components that are determined to form the same line due to the condition that the positions are close to each other are integrated, and the envelope of the character line obtained by the integrated character is considered. To obtain the angular coordinates of the envelope, see Fig. 1.
The coordinates of the corner points in the outline of the character line integrated as shown in (c) are extracted by the method proposed in the present invention, and the inclination of the character line is extracted from the positional relationship of the corner points as shown in FIG. 1 (d). Estimate the angle. By using this processing method, it is possible to perform the tilt angle detection processing with a smaller number of steps as compared with the related art.

[0010]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 3 shows an outline of processing in the apparatus realized by the present invention.

In FIG. 3, 1 is an image input means,
The document image is read by optical means or other means. Alternatively, it also includes a process of reading an image that has been electronically stored in a recording medium in advance in a magnetic or other form.

Reference numeral 2 denotes a connected component extracting means, which refers to the document image data input by the image input means 1 to detect a connected component of black pixels. Specific processing methods include known methods such as a method using labeling and a method using contour tracking processing.

Reference numeral 3 denotes a character component extracting means, which selects, from the connected components of the black pixels obtained by the connected component extracting means 2, a connected component which is determined to form a character.
As the selection criterion, the size of the connected component, the number of black pixels with respect to the area of the circumscribed rectangle of the connected component (that is, the density of black pixels), and other texture features can be used.
For example, when the size of the connected component is within a predetermined range, it can be determined that the connected component is a character.

Numeral 4 is a character line estimating means, which makes the same line from the connected components constituting the characters extracted by the character component extracting means 3 because of their positional proximity, similarity in size, and the like. The tentative character line is estimated by integrating the character elements that are considered to be composed.

Reference numeral 5 is a corner point detecting means, which is a means for detecting the corner points of the connected black pixel components forming the temporary character line obtained by the character line estimating means 4, and is a feature of the present invention.

Reference numeral 6 is a character line inclination angle calculating means for calculating the provisional character line inclination angle from the positional relationship of the coordinates of the corner points obtained by the corner point detecting means 5.

Reference numeral 7 denotes a document image inclination angle detecting means, which outputs a value which is considered to be most appropriate from the calculated values of the inclination angle of each of the temporary character lines as the inclination angle of the document image.

Each means will be described in detail below.

The image input means 1 captures a document image in a binary (bit image) format by using an optical image capturing device or another image capturing device. It also includes a process of reading from an image storage device that performs recording by a magnetic or other recording method.

The connected component extraction means 2 refers to the document image data input by the image input means 1 and extracts black pixels to be connected. As a method for determining connectivity, known definitions such as connectivity in four directions in the neighborhood and connectivity in eight directions can be used. Pixels to be connected are extracted based on this definition of connectivity. The extraction of the connected pixels can be performed by using a known technique such as labeling processing and contour tracking processing used in general image processing.

The character component extracting means 3 selects a connected component forming a character from the connected components of the black pixels obtained by the connected component extracting means 2. As selection criteria,
The size of the connected component, the number of black pixels with respect to the area of the circumscribed rectangle of the connected component (that is, the density of black pixels), and other texture features can be used. As an example, in a printed document, the size of the characters in the body area is constant, so it is possible to make a rough guideline for the determination. In a typical printed document, the size of the characters is about 5mm square,
If this is taken in with a resolution of 400 dpi (about 16 lines / mm), it will be about 60 dots square for a character. Therefore, it is possible to determine that a connected component extremely larger than this is not a text character. Similarly, it is possible to judge that a connected component that is too small is noise.

The character line estimation means 4 calculates from the connected components which form the characters extracted by the character component extraction means 3
Temporary character lines are estimated by integrating the character elements that are considered to form the same line because they are close to each other in position or have similar sizes. When the text is written horizontally, it is considered that connected components located at substantially the same coordinates in the horizontal direction and having a sufficiently small horizontal interval form the same line.
The determination as to whether or not they are located at the same coordinates can be made based on the positional relationship between feature points such as end points and centroids.
In addition to this, a method of making a determination from the height of the connected component and the degree of overlap in the horizontal direction may be considered. Further, when the text is written vertically, the same processing can be performed by considering the horizontal direction and the vertical direction in exchange. Whether or not they are sufficiently close to each other can be determined based on the positional relationship between feature points such as end points and center of gravity.

The corner point detecting means 5 is the character line estimating means 4 described above.
It is a means for detecting the corner points of the connected black pixel components forming the temporary character line obtained by the above, and is the point of the present invention.

An algorithm for detecting a corner point will be described below with reference to the drawings.

As shown in FIG. 2, the corner point detection algorithm detects corner points by detecting, as corner points, points contacting a temporary character line when pressed from outside with a straight line of ± θ. Among the detected corner points, the corner points that are in contact with the straight line inclined by ± θ from the vertical direction and the corner points that are in contact with the straight line inclined by ± θ from the horizontal direction are paired. However, since the corner points of adjacent sides often coincide with each other, in this case, the coordinates of the corner points existing on the same side are appropriately paired and sent to the next character line inclination angle detecting means.

An example of a specific processing procedure will be described. The contour tracking processing is performed on the character lines integratedly extracted by the character line extracting means 4. At this time, the value of the intercept is calculated for each tracked coordinate, assuming that a straight line of ± θ degrees passes through the coordinate. Then, the contour coordinates when the value of the intercept becomes maximum or minimum are stored. When the contour tracking process is completed, the coordinate values of the corner points can be obtained.

In some cases, a plurality of contact points may occur on a certain straight line and the coordinates of the corner points may not be uniquely determined. In this case, selection is made so that the distance between the corner points becomes the largest.

The character line inclination angle calculating means 6 determines from the positional relationship of the coordinates of the corner points obtained by the corner point detecting means 5,
The inclination angle of the temporary character line is detected. In the method of calculating the tilt angle, the circumscribing frame of the character line is detected from the coordinates of the corner points, and the tilt angle is calculated from the direction of the frame line. Although four frame lines can be obtained, the accuracy increases as the distance in the row direction increases, so a method of estimating the angle from two corner points having the longest distance in the row direction can be given as an example. You may average using both the upper and lower frame lines.

The document image tilt angle detection means 7 outputs a value which is considered to be most appropriate from the calculated values of the tilt angles of the respective temporary character lines as the tilt angle of the document image. A method such as maximum likelihood estimation or averaging can be used as the most appropriate angle estimation method.

[0031]

As described above, according to the present invention, since only the coordinates of both end points of a temporary character line are detected, the number of processing steps is reduced particularly in the processing such as linear approximation as compared with the conventional method using the Hough transform. can do. Since the inclination angle detection process for one character line can be performed with a small number of processing steps, according to the conventional method, the present invention can obtain the inclination angles for a plurality of lines within the processing time for one line. It is possible to reduce the number of processing steps as a whole without lowering the accuracy of the estimated value of.

[Brief description of drawings]

FIG. 1 is a diagram showing an outline of processing of the present invention,
(A) shows the stage where the connected pixel component is extracted, and (b).
Shows the stage where the adjacent connected pixel components are integrated to generate a character line, (c) shows the stage where the corner points of the character line are extracted, and (d) estimates the tilt angle θ from the corner points. The results are shown.

FIG. 2 is a diagram illustrating a corner point detection method according to the present invention.

FIG. 3 is a diagram showing an embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 image input means 2 connected component extraction means 3 character component extraction means 4 character line estimation means 5 corner point detection means 6 character line inclination angle calculation means 7 document image inclination angle detection means

Claims (1)

[Claims] 1. An image input unit for optically scanning a document to obtain document image data, and a connected component extracting unit for referring to the document image data input by the image input unit to detect connected components of pixels. Among the connected components of the pixels extracted by the connected component extraction unit, a connected component that extracts a connected component that forms a character, and among the connected components extracted by the character component extraction unit, adjacent connected components A character line estimating unit that integrates each other and outputs as a temporary character line, and a temporary character line estimated by the character line estimating unit into straight lines inclined by a predetermined angle from the horizontal direction and the vertical direction, respectively. Detects the outermost contour points that touch,
It is obtained from the character line inclination angle calculating means for calculating the inclination of the temporary character line from the positional relationship between the corner point detecting means and the corner points detected by the corner point detecting means, and the character line estimating means. A document image inclination angle detection unit that obtains the inclination angle of the document image data by referring to the inclination angle obtained by the character line inclination angle calculation unit from each temporary character line. Document image tilt detection device.