JPS62165277A - Document picture data filing system - Google Patents

Abstract

PURPOSE:To reduce a filing capacity for a document picture data by finding the minimum rectangle out of rectangles including all of the contents posted on the original of an input document picture data, and storing it in a form that a regulated size of space part is added in the neighborhood of the rectangle at a filing device. CONSTITUTION:The whole area 31 of a picture memory MEM1 are scanned in the (x) direction, and the accumulated value of black picture elements at each line is found, and it is written at a code memory CMEM in a table form as the accumulated value for (y) coordinate, and out of the accumulated values, a value which exceeds a prescribed threshold value epsilon1 is searched in order of the smallest (y) coordinate, and a search is terminated at a point y0 where is the first exceeding point. Next, the value is searched in the reverse order from the largest (y) coordinate, and a point y1 is found, and the part of a length l1 is found, and the part of a length ly, the (y) coordinate of which is between y0 and y1, is recognized as an effective part in (y) direction. Similarly, a point x0 and a point x1 are found, and the part of a length lx between the two points is recognized as the effective part in (x) direction. The bit of information, for example, is stored at the code memory CMEM as a pair of a left-upper coordinate (x0, y0), and a lateral length lx, and a longitudinal length ly. In this way, the filing capacity for the document picture can be reduced, and a retrieving speed for the document picture can be improved.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image file system, and particularly to a document image data file system suitable for document image data.

[Conventional technology]

An image file system that stores documents such as general documents, drawings, newspapers, academic papers, patent specifications, and bank terminal slips as image data on media such as optical disks is highly efficient and simple. There is a strong desire for a method for registering image data in

In conventional image file systems, for example, as described in Hitachi's optical disk file system operating manual (Hitachi Manual 6O-10-001-20), each operator controls the size of the image to be input using a switch or the like. Therefore, it was necessary to select and give instructions from among multiple types of candidates.

In other words, if two types of candidates, A3 size and A4 size, are prepared, for example, A4 size and 85 size (when inputting a document smaller than A4 size, use A4 size, and A3 size and 84 size). For example, when inputting a document that is thicker than A4 size (A3 size or smaller), select and specify A3 size. In this case, the input screen will be stored as an A4 and A3 size image in the file, so For originals that are smaller than the predetermined candidates, such as 85 and 84 sizes, margins are added around the edges and the page [A
4. It will be stored in a file in a format normalized to a size such as A3 size.

[Problem that the invention seeks to solve]

According to the above-mentioned conventional method, when an image is input using an original whose width is smaller than a predetermined width of an input image, the background around the original becomes a large blank area and is stored in a file together with the image data. Furthermore, even when an image is input using a document of a predetermined size, since the document itself has margins around its periphery, these margins are stored together with the data portion in one file. Therefore, there is a problem in that the blank space requires an unreasonable amount of file capacity.

An object of the present invention is to reduce the file size of document image data.

[Means for solving problems]

In order to achieve the above object, the present invention is characterized in that an input image is stored in a file device while excluding unsatisfactory blank areas. In order to exclude unstable blank areas, for example, digitized document image data is projected in the vertical and horizontal force directions to obtain the cumulative value of pixel density, and among those whose values exceed a predetermined threshold, In the vertical direction, the coordinates of the uppermost point and the lowermost point, and in the horizontal direction, the coordinates of the leftmost point and the leftmost point may be determined.

[Effect]

The above-mentioned top, bottom, left, and right dots, the characters that are the contents of the manuscript,
It represents the smallest rectangle that includes all figures, tables, etc., so
The file f is created by removing the margins outside the paper added by the image input device and the margins of the paper itself.
This can be stored. In addition, since each point above is determined by comparing the cumulative value of pixel density with a predetermined threshold value, it is independent of noise information in the margin area (this process can be achieved. It is also possible to add a margin smaller than the part and store it in a file, so the operability during searching does not deteriorate.

〔Example〕

An embodiment of the present invention will be described below.

FIG. 1 is a block diagram of the image file system. In the figure, 5CAN is an image input device fil, IMB
Ml is an image memo IJ that stores input images, IMF3M2
is an image memo IJ for storing converted images that have undergone image conversion processing such as cropping, CMEM is a memo 1 for storing code data such as table data, FILE is a file device that stores images, and PROC is the entire system. This is a processing device that performs image processing on image data and controls the image data, and operates according to a program that is stored internally and will be described later.

Next, the operation of storing document image data in a file in this system will be explained. FIG. 2 is a flowchart of the entire document image data input process, which is realized by the program of the processing device paoc.

In step 21, the original document is photoelectrically converted by the image input device 8CAN, digitized, and stored in the image memory IMEM1i.
Store this.

The range to be read is specified by selecting one of the candidates prepared in advance using a switch (not shown) or the like. Therefore, in general, background data outside the manuscript is also included in the IM.
In steps 22 to 27 taken into gMl, the image processing means, which will be described in detail later, detects the positional information of the effective portion from which the background and margins of the document image have been removed, and stores the code memo IJcMEME. In step 28, a partial image of the stored document image data is cut out from the main information E and the image memo IJIMEMI) is created as an image memo IJM B
M 2 t predetermined margin data are combined and stored.

In the final step 29, the data of the image memo IJIMEM2 is sent to the file device FIIJ together with the width and height information of the image data in the image memo IJIMEM2.
Store 1 item.

Next, a method for detecting an effective area will be explained using FIG. 2.

In the figure, 31 is a frame corresponding to the range of the image memory lJMEM1, which corresponds to the image reading range of the image input device 8CANi selected by the operator.

32 is the size of the surface of the read manuscript, and 35 is the size of the effective area corresponding to the written contents of the read manuscript, which corresponds to the smallest rectangle that includes all the data contents such as characters, figures, etc. First, the entire image memory MBM1 area 31 is
The cumulative value of black pixels in each row f is determined by scanning in the direction f (step 22).

In the figure, this is represented by graph 33, but in reality, y
Code memo IJCM in table format as cumulative value for coordinates
Write to EM. Next, from among these accumulated values, those whose y-coordinate exceeds a predetermined threshold ε1 are searched for in descending order, and the point y that exceeds it for the first time is found. to terminate the search (step 23
). Next, the point y1 is similarly searched in reverse order starting from the one with the largest y coordinate (step 24). The y coordinate is y. There is a certain length between E and yl! Let the part be the effective part in the y direction.

Exactly in the same way (Thus, find the point X. and xl (Step 25)
~27), and the portion of length j8 between them is defined as the effective portion in the X direction. The valid portion of the image data obtained as a result of the above matches frame No. 35. This information t'i, for example, is stored in the code memory CMgMt as a set of upper left coordinates (xo, yo), horizontal length j8, and vertical length j.

The above processing is realized by the program stored in the processing device PROC+, and its specific flowchart is shown in the fourth section.
Figure 1 is shown. In order to simplify the explanation, this flowchart describes an example that targets a binary image. Steps 401 to 410 in the figure are the above-mentioned y. The process of finding
411 to 420 correspond to the process of calculating yl. Also f
(y) represents the cumulative value of pixel density, that is, the peripheral density distribution. However, in this program, the process 40 to obtain (maf(y))
4°414 and end point y. Process for calculating +Y1 405°41
5 are performed in the same loop, and the calculation of f(y) that is unnecessary for the search is omitted. The effective part obtained above corresponds to one rectangle that circumscribes the content E of the document, etc., but
In step 28, a margin of a predetermined width is added to make the area effective. This is concretely about 5 kneading.If the width of the upper margin is less than a1, the lower part is b1, the left side is C9, and the right side is d, the upper left point of the effective part is (xo-C+Yo-a), and the vertical length is (
−j'.

+a+b), and the horizontal length is defined as (j, -)-c-)-d). The effective portion found here corresponds to frame 261.

Note that this flow omits steps 403 and 413 and replaces steps 404 and 414 with f (y)=f (y)+I(x,
y) (I(x, y) is the pixel density of point (X, y)), it can be easily extended to the processing of grayscale images.

Figure $5 shows the file format of the document image data created above. In the figure, 501 is the horizontal length of the document image data, and corresponds to jx+c+d.

502 also has a vertical length and corresponds to j-1-a-1-j. 503-505 are yo-a<y(,y.

This corresponds to the portion xo−c<x x, +d, and each corresponds to one row of image data. 506 to 514 are yo<y
<y,, x oc<x<x1+d (corresponds to this.Special Eko 506.509.512 is x.

-c < x < xo part, 507, 510, 51
3 is X. <x<xl part, 508, 511, 514i
tx, corresponds to the part where x<x1+d. 515°516 is Y, <y<yl +b, Xo-c<x<
The part x1+d (corresponds to this).

Note that, depending on the application, the file capacity can be minimized by setting a to d in FIG. 3 to 0 and making the frame 36 coincide with the frame 35.

〔Effect of the invention〕

According to the present invention, since information regarding the position and size of the valid portion of input document image data can be detected, only the valid portion can be stored in a file without the operator instructing this. Therefore, there is an effect of reducing the document image file capacity and improving the document image search speed accordingly. Further, since it is possible to add a margin of a predetermined size around the valid portion and store it in a file, the operability of the search does not deteriorate.

[Brief explanation of drawings]

FIG. 1 is a block configuration diagram of an image file system according to an embodiment of the present invention, FIG. 2 is a flowchart of document image data file storage processing, and FIG. 3 is a schematic diagram of image memory contents in effective area detection processing E. , FIG. 4 is a flowchart of effective area detection processing, and FIG. 5 is an example of the format of document image data stored in a file. 8CAN... Image input device, IMEMl. IMgM2... Image memo IJ, PROC...
...Processing device, FILE...File device,
CMEM...Kot 1 bacteria + 2 mouths

Claims (1)

[Claims] 1. In a document image file consisting of an image input device, image memory, processing device, and file device, find the smallest rectangle that includes all the contents written in the manuscript of the input document image data, and calculate the rectangle around this rectangle. A file method for document image data characterized by storing it in a file device with a margin of a certain size added.