KR100343672B1 - Blocking method of document video - Google Patents

Abstract

PURPOSE: A blocking method of a document video is provided to increase a processing speed by decreasing a memory capacity into half and simplifying process. CONSTITUTION: A binary video is inputted. An RLSA(Run Length Smearing/Smoothing Algorithm) is executed horizontally based on a black pixel with respect to the inputted binary video. The executed result is stored in a memory. An RLSA is executed vertically based on a white pixel with respect to the inputted binary video, and a white pixel Run is extracted. If a length of the extracted white pixel is larger than a critical value, a value of the extracted white pixel is stored in an identical area of a horizontal RLSA result memory corresponded to an area of the white pixel Run.

Description

Blocking Method of Document Image

The present invention is to block the effective area in order to extract the effective area on the document, and in particular, half the amount of memory by blocking the effective area in order to make it easier to extract the effective area on the document such as text, pictures, photos, etc. The present invention relates to a method for blocking document images that can be improved by reducing the number of steps and simplifying the processing steps.

A conventional method of blocking a document image includes a first step of inputting a binary image as shown in FIG. 1, and a run length smearing / smoothing algorithm in a horizontal direction with respect to the binary image input in the first step. Performing a second step; storing a result of the second step; storing a result of the second step; performing a RLSA again in the vertical direction with respect to the binary image input in the first step; A fifth step of storing the results performed in step 4 and a sixth step of performing a logical AND on the results stored in the third and fifth steps, respectively, to obtain a blocked image.

Looking at the prior art made of each step as described above in detail as follows.

Conventionally, a method called RLSA (Run Length Smearing / smoothing Algorithm) is used. A set of consecutive black pixels is called a run, and after extracting those runs from an image, the interval between runs, that is, the length or number of pixels If is less than the threshold, the white pixels between two runs are replaced with black pixels.

In other words, when inputting a binary image as shown in FIG. 2A, the RLSA (Run Length Smearing / Smoothing Algorithm) is first performed in the horizontal direction with respect to the binary image. As a result, as shown in FIG. Get the result and store it in memory.

Next, RLSA (Run Length Smearing / Smoothing Algorithm) is performed on the input binary image in the vertical direction, and the result is obtained as shown in (C) of FIG. 2 and the result is stored in another memory.

When a value obtained by logically multiplying a value of a memory in which RLSA results are performed in different directions in a horizontal direction and a vertical direction is obtained, a blocked image is obtained as in (d) of FIG.

In addition, when looking at the method of the RLSA, first finds the run of the black pixel and the white pixel of the binary image to be input. For example, when the same run as in (a) of FIG. 3 is found, the next run of the black pixel is found. When the distance between the threshold and the threshold is smaller than the threshold, the white pixel between the two runs is changed to the black pixel, and if the threshold is larger than the threshold, the white pixel is left as it is. Same as in (b).

Here, the distance between two runs is calculated by the number of pixels.

However, in the conventional technique, at least twice as much memory as the input image is required by storing the result of performing the RLSA in the horizontal direction in a memory and the result of performing the RLSA in the vertical direction in another memory. As a result, a large amount of memory is required.

Accordingly, an object of the present invention to solve the conventional problems as described above is to provide a method for blocking document images that can improve the processing speed by reducing the amount of memory required by half and simplifying the processing steps. have.

Another object of the present invention is to perform a RLSA based on the black pixel in the horizontal direction and to store in the memory, and then, if the length of the white pixel obtained by performing the RLSA based on the white pixel in the vertical direction is larger than the threshold value, The present invention provides a method of blocking a document image so that the same result can be obtained by storing the same RLSA result corresponding to the area in the same area of the memory and reversing the horizontal / vertical order.

The present invention for achieving the above object is to perform a RLSA (Run Length Smearing / smoothing Algorithm) in the horizontal direction with respect to the black pixel on the first step of inputting a binary image, and the input image in the first step A second step, a third step of storing the result performed in the second step in a memory, and a white pixel run obtained by performing RLSA in a vertical direction with respect to the white pixel on the binary image input in the first step A fourth step of extracting (Run); and if the length of the white pixel extracted in the fourth step is larger than a threshold value, the value of the white pixel extracted in the same area of the horizontal RLSA result memory corresponding to the area of the white pixel run is determined. The fifth step is to store.

As shown in FIG. 4, the apparatus for performing each of the above steps includes a scanner 70 that receives an image from a document, and a digital signal processor that performs signal processing by receiving an image transmitted from the scanner 70. 90, a scanner interface 80 for transmitting digital data from the scanner 70 to the digital signal processor 90, and an image signal of the digital signal processor 90 through a host interface 60. A host computer 10 which processes a process of blocking an effective area to extract an effective area on a document, and a local memory 100 that stores necessary data in the digital signal processing unit 90 and makes use of the stored data. And a control logic unit 20 and an arbitration logic unit 30 for adjusting the operation of the digital signal processing unit 90 according to the control operation of the host computer 10; The global memory 50 stores an input image, and the DRAM controller 40 controls the operation of the global memory 50.

Referring to the operation and effect of the present invention configured as described in detail as follows.

When the images of the document are input in the form of binary data through the scanner 70, the scanner interface 80 transmits the images to the digital signal processor 90.

Then, the digital signal processor 90 stores the input binary image in the local memory 100 to perform character recognition.

Next, the digital signal processing unit 90 processes the data stored in the local memory 100 into data that allows the effective area to be blocked so as to extract the effective area on the document, and then through the host interface 60. Deliver to (10).

Therefore, the host computer 10 adjusts the image input through the control logic unit 20 and the arbitration logic unit 30, and simultaneously stores the input image in the global memory 50 and blocks the data. Let's take a look at this.

The host computer 10 first performs a Run Length Smearing / Smoothing Algorithm (RLSA) in a horizontal direction on the binarized image.

That is, for the input binary image, as shown in (a) of FIG. 3, the runs are found as black pixels, and then the distance between the black pixels and the black pixels (ie, the white pixels) and the threshold value are compared. If the distance is smaller than the threshold, the white pixel between the two runs is replaced with black pixels. If the distance between the two runs is larger than the threshold, the white pixel between the two runs is left as it is. Change to the same pixel value.

Here, black pixels refer to pixels having information constituting letters and the like, and white pixels refer to background pixels.

Next, RLSA is performed in the vertical direction with respect to the white pixel among the binary images.

That is, a run of white pixels is extracted in the vertical direction based on the white pixels, and at this time, the length of the extracted pixels is compared with a threshold value. storing the result, and save the white pixel value obtained by performing the horizontal RLSA the same region of memory, and places the stored value if the whitening bovine length greater than a threshold value as the horizontal direction RLSA memory as a result.

In other words, if the length of the extracted white pixel is larger than the threshold value, the horizontal pixel RLSA results change the black pixel of the memory to the white pixel, so that the same block image as in the conventional method can be obtained.

In addition, the same applies to the case of performing RLSA by changing the horizontal / vertical order as described above.

That is, if the RLSA of the center of the black pixel is first performed in the vertical direction, the same result may be obtained by performing the RLSA of the center of the white pixel in the horizontal direction.

As described in detail above, the present invention has the effect of reducing the amount of memory required by half and simplifying the processing steps to improve the processing speed.

1 is a process diagram for a method of blocking a conventional document image.

FIG. 2 is an exemplary diagram showing a block method by Run Length Smearing / Smoothing Algorithm (RLSA) in FIG.

(A) is an example of input video,

(B) is the horizontal RLSA result,

(C) is the vertical RLSA result,

(D) is a blocked image diagram.

FIG. 3 is an explanatory diagram showing the process of RLSA in FIG.

(A) shows the extracted run,

(B) shows an example where the pixel values are reversed.

4 is a circuit diagram of a document image segmentation apparatus.

5 is a process diagram for a method of blocking a document image of the present invention.

******* Explanation of symbols on the main parts of the drawings *******

10: host computer 20: control logic section

30: arbitration logic section 40: DRAM controller

50: global memory 60: host interface

70: scanner 80: scanner interface

90: digital signal processing unit 100: local memory

Claims (1)

A first step of inputting a binary image, a second step of performing RLSA (Run Length Smearing / Smoothing Algortithm) in a horizontal direction with respect to the black pixel on the binary image input in the first step, and the second step A third step of storing the result of the operation in the memory; and a fourth step of extracting the white pixel run by performing RLSA on a vertical direction with respect to the binary image input in the first step; And if the length of the white pixel extracted in the fourth step is larger than the threshold value, storing the extracted white pixel value in the same area of the horizontal RLSA result memory corresponding to the area of the white pixel run. How to block document image.