JPH02263272A - Document picture processor - Google Patents

Abstract

PURPOSE:To integrate one or more areas in a document picture and to facilitate an area segmenting process by obtaining a coupling component of black picture elements after excluding the blank areas out of an input document picture which can include a rectangle set previously. CONSTITUTION:A document picture processor consists of a picture input device 5 which reads the documents as the binary pictures, a picture memory 6 which stores temporarily the input pictures, a picture processor 15 including a run extracting means 12 which extracts the runs longer than the prescribed value, a picture logical arithmetic means 13 which performs the logical arithmetic of pictures, and a contour tracking means 14 which tracks the contours of pictures, etc. Then the blank areas are detected out of a document picture which can include a rectangle set previously, and the means 13 turns all picture elements into black ones except the blank areas to integrate the areas of the document picture. As a result, the blank between stages is ignored within a stage set and the areas of the stage set are integrated into a single area. Then an area segmenting process is facilitated.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a document image processing device, and more particularly to a document image processing device that easily cuts out regions by integrating regions.

(Prior art) Conventional methods for cutting out regions in document images utilize the projection of the document, obtain vertical and horizontal projections, and focus on the presence or absence of black pixels in each row and column of pixels. Methods (e.g., edited by Arata Hashimoto, published by Telecommunications Association,
[Refer to Introduction to Character Recognition JP59-60].

For example, in a document image l as shown in FIG. 2 (diagonal lines represent characters, etc.), the position of each area can be determined by its projection 2. That is, each area can be cut out.

(Problem to be Solved by the Invention) However, in a document image 1' having a frame 3 and columns 4 as shown in FIG. It is difficult to determine the location.

An object of the present invention is to provide a document image processing device that can integrate regions in a document image and easily cut out regions regardless of the presence or absence of columns and frames.

(Means for Solving the Problems) The document image processing device of the present invention includes an image input means (FIG. 1, 5) for inputting a document as a binary image, and an input image storage means (6) for storing the input image. and processed image storage means (7 to 7) for storing new images obtained as a result of processing the images.
10), blank area detection means (12) for detecting a blank area in a document image that can include a preset rectangle, and image logic operation means (13) for integrating areas in the document image based on the detected blank area. ).

Further, a document image processing device according to another aspect of the present invention includes image input means (FIG. 2, 5) for inputting a document as a binary image.
and input image storage means (6) for storing the input document image.
), a processed image storage means (7, 8, 91, 92, 10) for storing an image obtained as a result of processing the input image, and a white run for extracting a white run in the input document image that is longer than a preset value. an extraction means (12') and a first setting value (lyl
) longer than the second setting value (lxl) and the corresponding part of the horizontal white run in the input document image that is longer than the second setting value (lxl), and the vertical direction longer than the third setting value (ly2) image logical operation means for integrating regions in an input document image by turning all pixels to black except for a portion where a white run of 1 and a white run in a horizontal direction longer than a fourth setting value (lx2) coincide with each other; It is characterized by comprising:

(Operation) The document image processing device according to the first aspect of the present invention detects a blank area in a document image that can include a preset rectangle using a blank area detecting means (12), and detects a blank area in a document image that can include a preset rectangle, and Areas in a document image are integrated by performing processing to make all pixels except blank areas black. By this process, blank areas smaller than the preset rectangle are excluded from the blank areas, so the blank spaces between columns in column set 4 are
) are ignored, and the column areas are merged into one, as shown by the dashed line. Furthermore, frame 3 is detected as a remaining character line image area, etc. as an area other than a blank area.

By integrating the regions in this way, the process of cutting out the regions can be easily executed.

Further, in the document image processing device according to the second aspect of the present invention, in the image logic operation means for performing logical operations such as logical AND and OR of images and black and white inversion, settings in the document image extracted by the white run extraction means are provided. An operation is performed to make all pixels black except for the matching part of the vertical and horizontal white runs that are longer than the value. This allows regions in the document image to be integrated,
To easily execute a process of cutting out an area.

(Description of Embodiments) First Embodiment FIG. 1 is a block diagram showing the structure of a document image processing device according to a first embodiment to which the present invention is applied. Image input device to read! ! 5
and an image memory 6 for temporarily storing manually generated images.
Image memory 7 having the same memory size as image memory 6
to 10, and a control device 11 that controls the entire device;
Run extraction means 1 for extracting runs longer than a preset value
2, an image processing device 15 consisting of an image logic operation means 13 that performs logical operations on the image, and a contour tracking means 14 that traces the contour of the image, and an input device that inputs commands and the like! 1!
1B, a display 17 for displaying commands etc. input from the human-powered device 16 and images stored in the image memories 6 to 10, a file device 18 for storing image data, and an image memory 6%N L, 10. It also includes an image output device 19 that prints the stored image.

Next, an example of a procedure for integrating regions in a document image input by the above-mentioned apparatus and cutting out the regions will be described in detail.

In cutting out the area of the document image 1' shown in FIG.
There is a case where each area is completely separated and cut out as shown by the broken line in Figure 4(a), and a case where the layout of the document is strongly related from a structural point of view as shown by the broken line in Figure 4(b). A plurality of estimated regions may be cut out as one region. In the present invention, the unit of the region to be extracted can be changed by simply changing the length of the run to be extracted. In this embodiment, an example will be explained in which the regions arranged in columns as shown in FIG. 4(b) are integrated into one region.

FIGS. 5(a), 5(b), and 5(c) are processing flows for integrating regions, and .circlein. through O in the drawings represent main processing steps. The processing procedure of the first embodiment of the present invention will be explained along this processing flow.

Processing step (2): A document is input as a binary image using the image input device 15, and the input document image is stored in the image memory 6.

Processing step (2): All pixels in the image memory 8 and the image memory 9 are made white.

Processing step (2): Make all pixels in the image memory 7 white.

Processing step (2): The +1) input document image stored in the image memory 6 is sequentially scanned in the vertical direction.

Processing step ■: If there is a white run during scanning, compare the length of the white run with a preset value 1yl, and if the length of the white run is greater, proceed to processing step ■; otherwise, Proceed to processing step ■. The above lyl is a value determined empirically.

Processing step (2): The pixels in the image memory 7 corresponding to the same position as the white run extracted in processing step (2) are made black.

Processing step ■: Determine whether processing step ■ has ended. If it has ended, proceed to processing step ■; otherwise, proceed to processing step ■. Here, the image stored in the image memory 7 and increased by +1) when it is determined that the process has ended is as shown in FIG. 6(a).

Processing step ■: The image stored in the image memory 7 is scanned in the horizontal direction (.

Processing step ■: If there is a black run during scanning, compare the length of the black run with a preset value 1xl, and if the length of the black run is greater, proceed to the processing step [phase] and perform other steps. If so, proceed to processing step ■. The above Ixl is a value determined empirically.

Processing step B [phase]: Pixels in the image memory 8 corresponding to the same position as the black run extracted in processing step (2) are made black.

Processing step ■: Determine whether processing step ■ has ended. If it has ended, proceed to processing step @; otherwise, proceed to processing step ■. Here, the image stored in the image memory 8 when it is determined that the process has ended is as shown in FIG. 6(b). The black area of this image 21 has a width of 1 xi and a height of lyl
In this embodiment,
Set Ixl and ly values to detect vertically long blank areas.
It is set to l.

Processing step @: Same as processing step ■.

Processing step [phase]: The input document image stored in the image memory 6 is sequentially scanned in the horizontal direction (.

Processing step 0: If there is a white run during scanning, compare the length of the white run with a preset value of 1x2, and if the length of the white run is greater, proceed to processing step [phase] and perform other steps. If so, proceed to processing step [phase]. The above lx2 is a value determined empirically.

→Processing step [phase]: The pixels in the image memory 7 corresponding to the same position as the white run extracted in processing step (2) are made black.

Processing step [phase]: Determine whether processing step @ has ended. If it has ended, proceed to processing step (2); otherwise, proceed to processing step @. Here, the image stored in the image memory 7 when it is determined that the process has ended is as shown in FIG. 6(C).

Processing Step 02 The image stored in the image memory 7 is scanned in the vertical direction.

If there is a black run during the processing step @2 scanning, compare the length of the black run with a preset value 1y2, and if the length of the black run is greater, proceed to the processing step [Stock] and proceed to the other If so, proceed to the processing step. The above ly2 is a value determined empirically.

Processing step @: The pixels in the image memory 9 corresponding to the same position as the black run extracted in the processing step [phase] are made black.

Processing step @: Determine whether processing step O has ended. If it has ended, proceed to processing step O; otherwise, proceed to processing step O. The image stored in the image memory 8 when it is determined that the process has ended is as shown in FIG. 6(d). The black area of this image 22 indicates a blank area in the input document stored in the image memory 6 that can contain a rectangle with a width of 1×L and a height of ly2. In this embodiment, values for detecting a horizontally long blank area are set for Ix2 and ly2.

Processing step O: The image obtained by calculating the logical sum of the image memory 8 and the image memory 9 is inverted in black and white, and the result is stored in the image memory 10. The image stored in the image memory 10 becomes as shown in FIG. 6(e), and the purpose of this processing, which is the integration of regions, is achieved.

Furthermore, in order to find the position of the above-mentioned integrated area, for example, scan the image in FIG. It is sufficient to trace the contour of the integrated area. Any known method can be used for contour tracking, and for example, the method detailed in "Image Database", co-authored by Masao Sakauchi and Yutaka Osawa, published by Shokodo, pages 91 to 95, can be used. This completes the integration and extraction of regions in the document image.

In this example, a case has been described in which multiple areas that are estimated to have a strong relationship from the perspective of product structure are extracted as one area for document layout, but small blank areas such as between lines or characters can also be detected. By setting the above settings, a character string or sub-character pattern can be made into a black pixel connected component, and as described in the example, by tracing the outline of the black pixel connected component, the character string can be Or sub-character patterns can be similarly cut out.

As described above, according to the present invention, areas in the input document image can be integrated by excluding blank areas in the input document image that can include a preset rectangle and making them a connected component of black pixels. Therefore, processing for cutting out regions can be easily applied.

Second Embodiment FIG. 7 is a block diagram showing the structure of a document image processing apparatus according to a second embodiment of the present invention. Note that the same parts as in the first embodiment shown in FIG. 1 are given the same reference numerals, and corresponding parts are given the reference numerals with "1" added. This device reads I! documents as binary images. li image input device 5, an image memory 6 that temporarily stores input images, and an image memory 7.8.91.92.10 that has the same memory size as the image memory 6, and controls the entire device. Image processing comprising a control device 11, a white run extraction means 12' for extracting white runs longer than a preset value, an image logic operation means 13 for performing logical operations on an image, and a contour tracking means 14 for tracing the contour of the image. A device f 5', an input bag [1B, and an input device 1t for inputting commands, etc.
Commands etc. input from IB and image memories 8, 7
, 8, 91.92° 10 A display 17 that displays images stored in the 10, and a file device that holds image data @
18, and an image output device 19 for printing images stored in image memories 6.7, 8.91.92.10.

Next, an example of a procedure for integrating areas in a document image input by the above-mentioned apparatus and cutting out the integrated area will be described in detail.

In the region extraction of the document image shown in Fig. 3, the fourth
It is assumed that there is a strong relationship between the case where each area is completely separated and cut out as shown by the broken line in Figure (a), and the document layout from a structural perspective, as shown by the broken line in Figure (b). In some cases, the areas that are separated may be merged and extracted. In the present invention, the unit of the area to be extracted can be changed simply by changing the length of the white run to be extracted. In this example, the fourth
The case of ryJ(b) will be explained as an example.

FIGS. 8(a) and 8(b) are processing flows for integrating regions, and in the figures, ■ to [phase] represent main processing steps. The processing procedure of the present invention will be explained along this processing flow.

Processing step ■: Image input device! 15, the document is input as a binary image, and the input document image is stored in the image memory 6.

Processing step (2): All pixels in the image memory 7 and the image memory 8 are made white.

Processing step (2): The entire surface of the input document image stored in the image memory 6 is sequentially scanned in the vertical direction.

Processing step ■: If there is a white run during scanning, compare the length of the white run with a preset value 1yl, and if the length of the white run is greater, proceed to processing step ■; otherwise, Proceed to processing step ■.

The above lyl is a value determined empirically. (In this embodiment, lxl and Iyl are set to values that extract a vertically long blank area.) Processing step ■: The pixels in the image memory 7 corresponding to the same position as the white run extracted in processing step ■ are Make it.

Processing step ■: Determine whether processing step ■ has ended. If it has ended, proceed to processing step ■; otherwise, proceed to processing step ■. Here, the image stored in the image memory 7 when it is determined that the process has ended is as shown in FIG. 9(a).

Processing step ■: If there is a white run during scanning, compare the length of the white run with a preset value IXI, and if the run length is larger, proceed to processing step ■; otherwise, proceed to processing step Proceed to [phase]. The above 1xl is a value determined empirically.

Processing step (2): The pixels in the image memory 8 corresponding to the same position as the white run extracted in processing step (2) are made black.

Processing step @: Determine whether processing step (2) has been completed. If it has been completed, proceed to processing step (2); otherwise, proceed to processing step (2). Here, the image stored in the image memory 8 when it is determined that the process has ended is as shown in FIG. 9(b).

Processing step 7102 calculates the logical product of image memory 7 and image memory 8, and stores the result in image memory 91. The image stored in the image memory 91 is as shown in FIG. 9(c).

Processing step O: Set value 1yl to ly2, set value lx
1 is changed to lx2, respectively, and processing steps (1) to (2) are performed (the processing flow in FIG. 8 is a simplified illustration). The above set values 1y2 and lx2 are values determined empirically. (In this embodiment, 1x2 and Iy2 are set to values that allow extraction of a horizontally long blank area.

) When this processing step is completed, the image stored in the image memory 7 is shown in FIG. 9(d), and the image stored in the image memory 8 is shown in FIG. 9(e).

Processing step [phase]: The logical product of the image memory 7 and the image memory 8 is calculated, and the result is stored in the image memory 92. The image stored in the image memory 92 is as shown in FIG. 8(f).

Processing Step 7102 The image obtained by calculating the logical sum of the image memory 91 and the image memory 92 is inverted in black and white, and the result is stored in the image memory 10. The image stored in the image memory 10 becomes as shown in FIG. 9(g), and the integration of the regions, which is the purpose of this processing, is achieved.

Furthermore, in order to find the position of the above-mentioned integrated area, for example, scan the image in FIG. It is sufficient to trace the outline of a certain integrated area.

This completes the integration and extraction of regions in the document image.

As described above, according to the second embodiment, seven white runs in the vertical direction that are longer than the set value in the input document image and the parts that match the white runs in the horizontal direction that are longer than the set value are blacked out. By creating connected components of pixels, one or more regions in a document image can be integrated. Since the regions can be integrated in this way, the process of cutting out the regions can be easily performed.

(Effects of the Invention) According to the present invention, by excluding a blank area in an input document image that can include a preset rectangle and making it a connected component of black pixels, or by using a set value in an input document image, It is possible to integrate one or more areas in a document image by excluding the matching part of a long vertical white run and a horizontal white run longer than a set value and making it a connected component of black pixels. can. Therefore, even if there are columns, frames, etc., it is possible to appropriately determine the region, and the integration of regions according to the present invention facilitates processing when cutting out regions.

[Brief explanation of drawings]

FIG. 1 is a diagram showing a block configuration of a first embodiment of the present invention. FIGS. 2 and 3 are diagrams showing document images and their projections. FIGS. 4(a) and 4(b) are diagrams showing examples of regions to be cut out in a document image. FIGS. 5(a) to 5(C) are diagrams showing an example of the processing flow of the first embodiment. FIGS. 6(a) to 6(d) are diagrams showing examples of images generated during processing, and FIG. 6(e) is a diagram showing an example of images obtained as a result of processing. FIG. 7 is a block diagram showing the configuration of a second embodiment of the present invention. FIG. 8 is a diagram showing an example of the processing flow of the second embodiment. FIGS. 9(a) to 9(f) are diagrams showing examples of images generated during processing, and FIG. 9(g) is a diagram showing an example of images obtained as a result of processing. 1.1'...document image, 2.2'...projection, 3...
・Frame, 4... Columns, 5... Image input device, 6 to 10
, 91° 92... Image memory, 11... Control device,
12...Run extraction means, 12'...White run extraction means, I3゜13'...Image logical calculation output stage, 14...Contour tracking means, 15.15'...Image processing device, 16.
...Human power device, 17...Display device, 18...
- File device, 19... image output device. Patent applicant Fuji Xerox Co., Ltd. Figure 1 Figure 4 Figure 5 (b) Figure 5 (a) Figure (a) (b) (c) (d) (e) Figure 7 ( a) (b) (c) (d) (e) Figure 8 (a) Figure 8 (b) Figure (f)

Claims (2)

[Claims] (1) Image input means for inputting a document as a binary image; input image storage means for storing the input image; and processed image storage means for storing a new image obtained as a result of processing the image; Document image processing characterized by comprising a blank area detection means for detecting a blank area in a document image that can include a rectangle, and an image logical operation means for integrating areas in the document image based on the detected blank area. Device. (2) image input means for inputting a document as a binary image; input image storage means for storing the input document image; and processed image storage means for storing an image obtained as a result of processing the input image; a white run extraction means for extracting a white run in an input document image that is longer than a value; a vertical white run in the input document image that is longer than a first set value; and a horizontal white run that is longer than a second set value; By making all pixels black except for the matching part, and the matching part of the vertical white run longer than the third setting value and the horizontal white run longer than the fourth setting value, 1. A document image processing device, comprising: image logic operation means for integrating regions in an input document image.