JP3024033B2 - Image processing device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for processing image data including a frame line, and more particularly to a frame line extraction process.
The present invention relates to an image processing apparatus that performs a registration process of an extracted frame line, a process of correcting deterioration of the extracted frame line, and a process of extracting an arbitrary designated region of a document from a document region designated by the frame line.

[0002]

2. Description of the Related Art Functions such as form processing of personal computers and word processors, detection of input / output positions of spreadsheet processing, and extraction of only a document portion or only a frame line of a document image specified by a frame line by a copying machine or the like are provided. To achieve this, (1) border line extraction technology,
(2) a technique for registering the extracted frame line; (3) a technique for correcting the deterioration of the extracted frame line; and (4) extracting an arbitrary designated area of the document from the document area designated by the frame line. Technology is required.

(1) The conventional frame line extraction technique extracts a line segment of the entire frame line.

(2) As a conventional technique for registering extracted frame lines, a technique for registering image data as it is is widely used, in which case the storage capacity is increased. As a method to solve this, a method of registering a start point and an end point of a registered image has been created.

(3) As a conventional technique for correcting the deterioration of the line portion of the extracted frame line, a technique using a chain code for line segment deterioration has been created.

(4) In the conventional technology for extracting an arbitrary designated area of a document from a document area designated by a frame line,
When an arbitrary designated area of a document is extracted from the image data including the document and the frame line, data of only the frame line is required separately from the image data including the document and the frame line.

(5) In the conventional line segment shift correction technique, a figure to be processed is optically converted and input to an image processing apparatus. According to the line feature amount of the input image, noise removal, embedding of a broken line, and the like are performed to obtain a line feature amount image.
A line position table is obtained from the line feature amount image.

[0008]

[Problems to be solved by the invention]

(1) In the conventional frame line extraction technology, if there is any deterioration in a line segment, it may not be possible to extract the line segment, or there is a problem that processing time is required to correct the line segment each time. Was.

(2) As a conventional technique for registering extracted frame lines, a technique for registering a start point and an end point of a registered image is not suitable for registering frame line data. Many appear and are not efficient.

(3) As a conventional technique for correcting the deterioration of the extracted frame line, a technique using a chain code for the line segment deterioration is first corrected for the line segment deterioration by extracting the line segment. The extraction has to be done, which can be time consuming.

(4) In the conventional technique for extracting an arbitrary designated area of a document from a document area designated by a frame line,
Data of only the frame line is necessary separately from the image data including the original and the frame line, which is not practical in view of convenience in use.

(5) In the conventional line segment shift correction technique, since the frame line is extracted after the correction, processing time is required depending on the quality of the line segment.

The present invention overcomes such a problem and provides an image processing apparatus for efficiently extracting a frame line from image data including the frame line, registering the frame line with a small amount of data, and reproducing the frame line. The purpose is to do.

[0014]

SUMMARY OF THE INVENTION An image processing apparatus according to the present invention is an image processing apparatus for extracting a frame line from image data including the frame line and reproducing the frame line data. Frame line extracting means for extracting frame data from data, frame line registering means for registering a frame image using the specific pattern of the frame line extracted by the frame line extracting means as registration data, and frame line reproducing means for reproducing the frame image And extracted by frame line extraction means
Density histogram of the area surrounded by the specific pattern of the drawn border
Density histogram calculation means for calculating the
Density histogram calculated by stogram calculation means
And a document designation area extracting means for registering and extracting an area having a predetermined value or more as a document area.

[0015]

[0016]

[0017]

[0018]

The frame line extracting means searches a specific pattern of the frame line from the image data including the frame line, and extracts the frame line. The frame line registration unit registers a specific pattern of the frame line extracted by the frame line extraction unit. The frame reproduction unit reads out a specific pattern of the registered frame and reproduces the frame. Concentration histo
The gram calculating means includes a frame line extracted by the frame line extracting means.
The density histogram of the area surrounded by the specific pattern
Calculate. The manuscript designated area extracting means is a density histogram meter.
The density histogram calculated by the
The upper area is registered and extracted as a document area .

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 13 is a block diagram of an image processing apparatus according to an embodiment of the present invention. The image data including the frame line is stored in the image memory 136. When image data is read from the input device 134, the input device 1 is read via the system bus 137 by a read command from the CPU 131.
The image data read from the memory 34 enters the image memory 132 once and is stored in the image memory 136.

Each process of this embodiment is executed by the CPU 131.
Through the system bus 137, the contents of the image memory 136 are read and processed. The image data as a processing result is transmitted from the CPU 131 to the system bus 13.
7 and is stored in the image memory 136 again. The image memory 136 has a sufficient capacity to store a plurality of image data.

The output device 135 is for displaying. When a display instruction is given from the CPU 131,
The data is output to the image memory 133 through the system bus 137 and displayed on the display device 135. The display device 135 and the image memory 133 can be realized without the present invention. The pattern matching device 138 is used particularly for pattern matching when extracting a specific pattern of a frame line. High-speed processing can be performed by adding this device. The calculation can be performed by the CPU 131 without this device. FIG. 13 shows an example for realizing the present invention, and realization is possible with other system configurations. The CPU performs frame line extraction processing, frame line registration processing, frame line reproduction processing, line segment deterioration correction processing, original designated area extraction processing,
A line shift correction process is executed.

In this embodiment, a frame line is extracted by not extracting all line segments of the frame line but extracting a specific pattern of the frame line. Therefore, the frame line can be extracted without performing the correction process each time the line segment is deteriorated, which is a conventional problem, and the processing time is reduced. The frame line registration process realizes registering the image data of the frame line with a small capacity by using the specific pattern extracted in the frame line extraction process as the frame registration data.

The frame line in the image data always includes the specific pattern shown in FIG. The hatched portion of the specific pattern shown in FIG.
0. The image data 11 of the background is shown in white. The start point, the middle point (branch point), and the end point of the specific pattern can be determined based on the shape. The center point of the extracted specific pattern is set as the coordinate value of the specific pattern.

The specific patterns 1, 2, and 3 represent the starting points of the frame lines, the specific patterns 4, 5, and 6 represent the ending points,
The specific patterns 7, 8, and 9 are intermediate points (branch points).
Is represented. When data is read out in the raster scan direction, the extraction order is specific pattern 1 → 7 → 4, specific pattern 2 → 8 → 5, specific pattern 3
→ 9 → 6.

The specific pattern of the frame line is to arbitrarily set the size of the block by designating the line thickness.

FIG. 1 is a flowchart showing the frame line extraction processing of this embodiment.

First, in step S1, data is read out from image data including a frame line in raster scan order.

Next, in step S2, it is determined whether the read data is a specific pattern of a frame line. If the read data is not the specific pattern, the process returns to step S1. If the read data is the specific pattern, the process proceeds to step S3. In step S3, it is determined whether the read data is the start point. If it is the start point, the process proceeds to step S4, and if it is not the start point, the process proceeds to step S5.

In step S4, the start point of the frame line is registered.

In step S5, it is determined whether or not the point is an intermediate point (branch point). If it is an intermediate point (branch point), step S
Proceed to 6. If it is not an intermediate point (branch point), the process proceeds to step S7.

In step S6, an intermediate point (branch point) of the frame line is registered.

In step S7, the end point of the frame line is registered.

The above flowchart shows one frame line element in one line of image data including a frame line. When there are a plurality of frame line elements, it is necessary to search until the end of one line of this image data. Repeat the above work,
By performing the processing up to the last line of the image data including the frame line, a specific pattern in the image data can be extracted.

At the time of extraction, a frame line can be extracted irrespective of deterioration of the frame image other than the specific pattern itself.

In the above description, an example is described in which a specific pattern indicating a start point is first extracted. However, a specific pattern may be extracted from an intermediate point (branch point).

FIG. 3 shows an example of frame image data. This data is composed of frame line image data 31 and background image data 32.

FIG. 4 explicitly shows a specific pattern of a frame line to be extracted in FIG. 3 of the frame image data example. The pattern 41 is the specific pattern 1 and the pattern 4 in FIG.
2 corresponds to the specific pattern 7 in FIG. 2, pattern 43 corresponds to the specific pattern 7 in FIG. 2, and pattern 44 corresponds to the specific pattern 4 in FIG.

Referring to FIG. 4, image data including a frame line is read out in a raster scan order to search for a specific pattern.
First, the specific pattern 41 representing the start point of the frame line is extracted. The coordinates of the extracted start point are registered. The image data including the frame line is read out in order, and the intermediate point (branch point)
Alternatively, an end point pattern is extracted. In the example shown in the figure, an intermediate point (branch point) 42 is extracted. Register the extracted midpoint (branch point). Similarly, another intermediate point (branch point) 43
And the end point 44 are extracted and registered. When the end point 44 has been extracted, the extraction of one frame segment has been completed. Thus, the specific pattern of the frame line in FIG. 3 is extracted as shown in FIG.

FIG. 5 is a pattern showing the frame line itself. This appears when data is read from image data including a frame line in order to extract a specific pattern of the frame line. In the case of the present embodiment, the frame line can be extracted regardless of the elements in FIG. 5, but when the image data is read and a specific pattern is searched, it is confirmed that the frame line is a line segment of the frame line. May be read. However, since this data is irrelevant to the extraction of the elements of the frame line, this data may be degraded, and it can be seen that this extraction method is effective.

With the above method, the start point, the middle point (branch point), and the end point of the frame line can be obtained from the image data including the frame line. It is also easy to create a frame image by connecting the obtained points of the frame.

FIG. 6 is a flowchart showing a frame line registration process for registering a frame image according to the present embodiment.

In step S61, a specific pattern of the frame line is extracted from the image data including the frame line by the above-described frame line extraction process.

The center point of the specific pattern extracted in step S62 is registered as coordinates.

The extracted patterns correspond to the specific patterns 1 to 9 in FIG. By registering the specific pattern extracted as described above together with the extracted coordinate data, it is possible to register with a much smaller capacity than simply registering as frame image data.

To reproduce the registered frame data, FIG.
Are connected as follows. The specific pattern is drawn at the center point coordinates of each specific pattern, and the specific pattern 1 → 7 → 4, the specific pattern 3 → 9 → 6, the specific pattern 2 in the horizontal direction from the central point of each specific pattern.
Connect straight lines between → 8 → 5 pattern center points. Similarly, a straight line is connected between the pattern center points of the specific patterns 1 → 3 → 2, the specific patterns 4 → 6 → 5, and the specific patterns 7 → 9 → 8 in the vertical direction from the specific patterns. By this processing, FIG. 3 can be created from FIG. 4 in reverse to the frame line extraction processing.

By the above processing, an image can be registered efficiently.

This embodiment is characterized in that a frame line is extracted by searching and extracting a specific pattern of the frame line from image data including the frame line. According to this method, even if a line segment other than the specific pattern of the frame data has image deterioration, the line segment can be extracted by detecting the specific pattern. Since only a specific pattern is extracted, it is not necessary to perform a separate process of determining whether the image is a line segment or noise due to deterioration of an image or the like, so that high-speed processing is possible. By using the specific pattern of the frame line extracted by the frame line extraction process as data for frame line registration, it is possible to reproduce the frame line image with a smaller data amount than registering the entire frame line image data. is there. Further, since the coordinates of the specific pattern are extracted, there is a feature that the frame line image can be easily used for data processing of a form or the like as it is.

In the line part deterioration correction processing, the line segment data of the frame line is corrected from the specific pattern extracted in the frame line extraction processing. Therefore, line segments can be extracted without performing deterioration correction by the chain code, and the deterioration of the frame data can be corrected, so that high-speed processing can be performed.

FIG. 7 is a flowchart showing a line portion deterioration correction process for correcting a degraded frame line.

In step S71, a specific pattern of a frame line is extracted from image data by frame line extraction.

In step S72, the shape of the extracted specific pattern is determined, and the center points are connected by straight lines in the vertical and horizontal directions.

The processing for connecting the specific pattern is the same as the processing for reproducing the registered frame data. By connecting the straight lines, it is possible to realize the correction of the line segment of the degraded frame line.

FIG. 8 explicitly shows a degraded frame line and a specific pattern. Patterns 81, 82, 83, 86, 87, and 88 are extracted by the frame line extraction processing. In order to correct the deteriorated image shown in FIG. 8, a straight line is connected between the center points of the specific patterns. The connecting process is the same as the process of reproducing the registered frame line data. By performing this process, the degraded portions 84 and 85 can be corrected in FIG. The correction process can be performed by connecting the center of the specific pattern with a straight line or by replacing the specific pattern with this pattern when searching for the specific pattern of the frame line since there is the pattern of FIG. .

With the above processing, the frame line data can be corrected.

The connection of the frame data can be determined by the specific pattern of the frame extracted by the frame extraction processing.
Even when the image of the frame line other than the specific pattern itself is deteriorated, the frame line data can be corrected from the coordinates of the specific pattern. Further, since the correction is performed using the extracted specific pattern, the correction can be realized at a higher speed than the correction using a chain code or the like.

In the original designated area extracting process, first, only the frame line is extracted from the image data including the original and the frame line, and the data is subjected to arithmetic processing on the original data, that is, the image data including the original and the frame line. Thus, it is possible to extract an arbitrary designated area of the document.

FIG. 9 is a flowchart of an original designated area extracting process for extracting an arbitrary designated area of the original from the original area designated by the frame line from the image data including the original and the frame line.

In step S91, a specific pattern of a frame line is extracted by a frame line extraction process.

In step S92, a frame image is created from the extracted specific pattern by the frame reproduction process described above.

In step S93, a logical operation is performed on the created frame line image and the original image data, ie, the original and the image data including the frame line.

In step S94, the image in the area not surrounded by the frame is erased to obtain image data of only the original.

In step S95, a density histogram is calculated for the data of only the extracted original document for each area surrounded by the specific pattern extracted in step S91.

In step S96, step S95
If the density histogram obtained in step 2 becomes equal to or larger than the value specified arbitrarily, the area is registered as an area where the document is written. As the data to be registered, a specific pattern and its coordinates are registered.

FIG. 10 shows an example of image data including a document and a frame line. Document data 101 is included in frame line data 102.

FIG. 11 shows frame line image data obtained by the above processing from the image data shown in FIG.

FIG. 12 shows the image data shown in FIG.
In the document area specified by the frame line obtained from the image data shown in FIG. 1, the area 121 where an arbitrary specified area of the document is extracted is explicitly shown.

By the above processing, it is possible to extract an arbitrary designated area of the original specified by the frame line from the image data including the original and the frame line.

The frame line data can be corrected by the line partial deterioration correction processing. Here, the method of correcting the deterioration has been described, but when there is a deviation in the line segment, the deviation is corrected for the extracted specific pattern. Next, a description will be given of a line segment shift correction process for creating frame line data corrected using a specific pattern of the frame line including the shift correction.

FIG. 14 is a flowchart showing a line segment shift correction process for correcting the shift and deterioration of the frame line data and creating the frame line data according to a specific pattern of the frame line data.
A specific pattern of the frame line is extracted from the image data including the frame line by the frame line extraction process.

In step S141, if there is a deviation in the center coordinates of the extracted specific pattern, it is corrected.

In step S142, step S1
The specific patterns extracted in 41 are connected by straight lines.

FIG. 15 shows an example in which a specific pattern of a frame line is extracted. Patterns 151 to 154 are extracted specific patterns, and pattern 152 is a pattern 151, 15
This is an example in which the position is shifted as compared with 3, 154.

FIG. 16 is an example explicitly showing that a frame image is created by correcting the specific pattern of FIG. Patterns 151 to 154 are specific patterns of the frame lines. Patterns 155 to 157 are straight lines connecting the specific patterns. The specific pattern of the pattern 152 is the pattern 15
It can be seen that the coordinates have been corrected to the coordinates connected by the straight lines 1, 153 and 154. By performing positional correction on the specific pattern of the frame line extracted and shifted in this way, the shift can be corrected.

The horizontal and vertical positions of each specific pattern may be corrected by taking the average position of each specific pattern or the position where the most frequently appearing.

Even though the specific patterns are not arranged so as to be connected horizontally and vertically, if the specific patterns are connected by a straight line, there is a possibility that the patterns are totally displaced. A method of performing coordinate conversion such as rotation on the frame line data is available. When only one specific pattern is rotated and extracted, the specific pattern can be subjected to rotation coordinate transformation with reference to the position of another specific pattern.

In order to create frame data from the extracted specific patterns, the specific patterns in FIG. 2 are connected as follows. A specific pattern is drawn at the center point coordinates of each specific pattern, and the specific pattern 1 → 7 → 4 and the specific pattern 3 → in the horizontal direction from the center point of each specific pattern, respectively.
A straight line connects the pattern center points of 9 → 6 and specific pattern 2 → 8 → 5. Similarly, the specific pattern 1 → 3 → 2 and the specific pattern 4
A straight line connects the pattern center points of → 6 → 5 and specific pattern 7 → 9 → 8. The above is a description of the displacement. To correct deterioration such as line breakage, frame line data may be simply created by the above method. The line portion deterioration correction processing can correct the graphic in FIG.

[0078]

According to the image processing apparatus of the present invention, there is provided a frame line extracting means for extracting a specific pattern of a frame line from image data, and a frame pattern extracted by the frame line extracting means as registration data. A frame line registration unit that registers an image, a frame line reproduction unit that reproduces a frame image, and the frame line extraction unit.
The density of the area surrounded by the specific pattern
Density histogram calculation means for calculating the degree histogram
And the density calculated by the density histogram calculation means.
Area where the degree histogram is equal to or greater than a predetermined value
Since there is provided a document designated region extracting means for registering and extracting, it is not necessary to separately determine whether the image is a line segment or noise due to deterioration of an image, and high-speed processing is possible. In frame line registration, image data of a frame line can be registered with a small data amount. Further, the process of extracting an arbitrary designated area of the document from the document area designated by the frame line can be performed at high speed.

[0079]

[0080]

[0081]

[Brief description of the drawings]

FIG. 1 is a flowchart illustrating a method for extracting a frame line from image data including the frame line.

FIG. 2 is a specific pattern to be extracted from image data including a frame line.

FIG. 3 is an example of image data of a frame line.

FIG. 4 explicitly shows a specific pattern of a frame line to be extracted from the image data of the frame line.

FIG. 5 is a pattern showing a frame line itself.

FIG. 6 is a flowchart for registering a frame image.

FIG. 7 is a flowchart illustrating a method for correcting deterioration of a frame line.

FIG. 8 explicitly shows a degraded frame line and a specific pattern of the frame line.

FIG. 9 is a flowchart illustrating a method for extracting an arbitrary designated area of a document from a document area specified by a frame line from image data including the document and the frame line.

FIG. 10 is an example of image data including a document and a frame line.

FIG. 11 shows image data of a frame line.

FIG. 12 explicitly shows an arbitrary designated area of a document among document areas designated by frame lines.

FIG. 13 is a system block diagram showing an embodiment of the present invention.

FIG. 14 is a flowchart illustrating a method of creating image data by correcting a frame line for deterioration and deviation.

FIG. 15 is an example of a specific pattern of a frame line having a displacement.

FIG. 16 is an example in which a specific pattern of a displaced frame line is corrected to generate frame line data.

[Explanation of symbols]

 1-9 Specific pattern 131 CPU 132, 133, 136 Image memory 134 Input device 135 Output device 137 System bus 138 Pattern matching device

Claims (1)

(57) [Claims] 1. An image processing apparatus for extracting a frame line from image data including the frame line and reproducing the frame line data, comprising: a frame line extracting means for extracting a specific pattern of the frame line from the image data; and borders registration means for registering the border image specific pattern of the extracted frame line by line extracting means as registration data, and borders reproducing means for reproducing the border image, the border extract
Enclosed by a specific pattern of the frame line extracted by the output means
Density histogram meter that calculates the density histogram of the area
Calculating means and the density histogram calculating means.
The area where the density histogram has a predetermined value is the original area.
An image processing apparatus, comprising: a document designation area extracting unit for registering and extracting a document.