JPH0452976A - Picture processing system - Google Patents

Abstract

PURPOSE:To attain the speedup and the promotion of the efficiency of picture processing by storing picture data in a high-speed memory after dividing it into plural parts so as to be nearly similar figures, and executing the picture processing by the same processing routine. CONSTITUTION:The picture data read out from a medium by a sensor 1 and analog-digital converted is stored in a picture memory 2. Then, the picture data in this picture memory 2 is divided into plural parts so as to be the nearly similar figures, and a readout direction is determined so that all these divided picture data are arranged in the nearly same shape, and the divided picture data are read out successively, and are stored in an internal memory 3-1 capable of being accessed at high speed, and they are picture-processed successively by the same processing routine. Thus, the speedup and the promotion of the efficiency of the picture processing can be attained.

Description

[Detailed Description of the Invention] [Summary] Regarding an image processing method that divides and processes image data read from a medium, the purpose is to divide the image data and perform image processing at high speed and with high efficiency using the same processing routine. Then, the image data read from the medium is divided into multiple parts so that they have approximately similar shapes, and the reading direction is determined so that all of the divided image data are arranged in approximately the same shape, and the data is read out and stored in high-speed memory. The configuration is such that the images are stored sequentially and image processing is performed using the same processing routine.

[Industrial application field]

The present invention relates to an image processing method that divides and processes image data read from a medium.

[Prior art and problems to be solved by the invention] Conventionally, paper sheet processing equipment uses an optical sensor, a thickness sensor, etc. to read the pattern shading, external shape, etc. of the medium to be inspected (paper sheets), and then The read values are converted into gradation signals and stored. Then, desired image processing is performed on these gradation signals to measure, for example, the external shape and large area of the medium.

At this time, the image data of the medium read by the sensor is stored in an image memory, the image data is read from this image memory as necessary, and stored in the high-speed memory inside the central processing unit, and this high-speed memory is accessed. I was trying to perform image processing.

For this reason, the capacity of high-speed memory is generally small, and it is not possible to store and process all of the image data on the medium at once, making it necessary to access the slow external image memory many times. There was a problem that processing could not be performed. Another possibility is to divide the image data into multiple blocks and store each divided block one by one in the internal high-speed memory of the central processing unit for image processing, but separate processing routines are required for each of these multiple blocks. However, there is a problem in that efficiency is reduced.

The present invention aims to perform image processing at high speed and with high efficiency by dividing image data into multiple parts having substantially similar shapes, storing them in a high-speed memory, and performing image processing using the same processing routine. .

[Means to solve problems]

FIG. 1 shows a block diagram of the principle of the present invention.

In FIG. 1, a sensor 1 reads image data from a medium.

The image memory 2 is a memory that stores image data read by the sensor 1 and converted from analog to digital by an A/D converter.

The internal memory 3-1 is a memory that can be accessed at high speed.

[Effect]

As shown in FIG. 1, the present invention provides image data (or further multiple pixels m
(image data in which one pixel has a value obtained by adding up each value of
The reading direction is determined so that all of these divided image data are arranged in approximately the same shape (for example, so that the corner of the image data of the read medium is in the center),
The images are sequentially read out and stored in an internal memory 3-1 that can be accessed at high speed, and image processing is performed sequentially using the same processing routine.

Therefore, image data can be processed quickly and efficiently by dividing the image data into multiple parts with almost similar shapes, storing them in high-speed memory so that they are arranged in the same way, and sequentially performing image processing using the same processing routine. It becomes possible to do so.

〔Example〕

Next, the configuration and operation of one embodiment of the present invention will be explained in detail using FIGS. 1 to 7.

In FIG. 1, a sensor 1 is an optical sensor, a thickness sensor, or the like, and reads image data corresponding to the shading and outline of a pattern from a medium (for example, a sheet of paper).

The read image data is converted from analog to digital by an A/D converter (not shown) to generate digital image data.

The image memory 2 stores image data read by the sensor 1 and converted into analog-to-digital data by the A/D converter (or a value obtained by adding up each value of multiple pixels mxn of this image data to one pixel). image data)
It is a memory that stores.

The internal memory 3-1 is a memory that can be accessed at high speed, such as within the central processing unit 3.

Next, the operation of the configuration shown in FIG. 1 will be explained in accordance with the order shown in the flowchart of FIG. 2 and with reference to FIG.

In FIG. 2, ■ divides the image memory in which paper sheets exist into four parts. As shown in FIG. 3(a), the image data of paper sheets stored in the image memory 2 is divided into the first block, second block, third block, and fourth block as shown by dotted lines in the figure. Divide into 4 blocks.

O is the image memory 2 so that the corner of each block is in the center.
Set the read direction (scan direction). this is,
Regarding the first to fourth blocks divided into four as shown using the dotted lines in Figure 3 (A) in (2), Figure 3 (B)
The reading direction is set so that the corners of the first to fourth blocks are centered when stored in the internal memory 31, as shown in FIG. To be more specific, in FIG. 3(b), the first block is from S (starting point) to E.
(set the readout direction from left to right, top to bottom) in the direction of the arrow shown in the figure.Similarly, the second block, the third block,
For the fourth block as well, the reading direction is set as shown in the diagram S-E.

(2) is transferred to the internal memory 3-1. This is the 1st block, 2nd block, 3rd block, and 1st block divided into 4 from the image memory in the read direction set in
The read image data is transferred to the internal memory 3-1, and written so that the corners are centered as shown in either (a) to (d) of FIG. 3(c).

[Phase] determines whether or not four phases have been completed. If YES, process [phase] and [phase]; if NO,
Repeat 0 for the next block.

[Phase] scans in the horizontal direction (from left to right) pixel by pixel, and if a medium is found, the area after that is filled with 1's.

[Phase] is scanned in the vertical direction (from top to bottom) pixel by pixel, and if a medium is present, the area after that is filled with 1. -1 to 4
For example, the image data that is divided and stored (the first one shown in al)
Scan the block horizontally (from left to right) and if a medium exists, fill it with 1s, and scan vertically (from top to bottom) and if a medium exists, fill it with 1s. , Similarly, in Figure 3 (C)), fo
The second block, third block, and fourth block shown in l, (dl) are scanned in the horizontal and vertical directions indicated by the arrows in the figure, and if a medium is present, the subsequent blocks are scanned.
Fill it with

(2) performs an AND operation on a pixel-by-pixel basis. This is ■, [phase]
After performing the above processing, an AND operation is performed on the two pixel by pixel to obtain the outline of the medium.

[Phase] is calculated by calculating the value of large area - number of pixels processed x 8 one pixel. Here, "8", which is the number of pixels x 8, is read from the medium by the sensor 1, converted to binary image data by the A/D converter, and then calculated by calculating the sum of the binary values for 2 x 4 pixels. This is because the value of this sum is converted into one pixel, and the number of processes in image processing is reduced to ⅛. The process of determining this large area is shown in Figure 6 (a), for example.
For the original image of the first flock, (1) find the number of pixels with 1 in the image data (FIG. 6) obtained by AND operation of 0 x 8;

(2) Calculate the sum of the values of the pixels in FIG. 6 (narrow width) corresponding to all pixels within the range of 1 in FIG. 6 (d).

(3) Value found in (1) above - Value found in (2) above =
Find it as a large area. This means that the large area of the first block divided into four has been determined. Similarly, the other second to fourth blocks are sequentially determined.

As described above, the image data read out sequentially from the image memory 2 so that the corners of the first block, second block, third block, and fourth block divided into four blocks are in the center can be accessed at high speed in the central processing unit 3. internal memory 3-1
write the same processing routine (same module)
By performing each of the processes shown in Figure 2 [Phases] and following, and sequentially determining the large area of the medium, it is possible to divide a large amount of image data and load it into a small capacity high-speed memory for high-speed image processing. It becomes possible.

FIG. 3 shows an explanatory diagram of the operation of the present invention.

FIG. 3(A) shows an example in which the image data of the medium stored in the image memory 2 is divided into four parts. , 4th part '07.

FIG. 3(b) shows the reading direction of the image data of the medium. This is shown in the figure from S to E for the first to fourth blocks divided into four, so that the corner of the medium is in the center.
The image data is read out in the direction of .

Figure 3 (c) shows the first to fourth blocks of Figure 3 (b).
The image data of each block is read in the directions from S to E in the figure, and stored in the internal memory 3-1 with the corners of the medium in the center. ) is stored in the internal memory 3-1 so that the corner of
The same module) can be used to perform image processing, for example, image processing for calculating the large area of paper sheets as described above, at extremely high speed. Here, (a) the first block,
2nd block, Ic) 3rd block, (d+4th block)
The data is read in the direction from S to E shown in the fourth block and stored in the internal memory 3-1.

Next, the present invention will be explained in detail using FIGS. 4 to 7.

FIG. 4 shows the image data (original) on the image memory. This image data is read from the medium by the sensor 1,
After converting the data into binary image data using an A/D converter, the total sum of 2[values] is calculated for 2×4−8 pixels, and the data is converted into one pixel having the value of this sum. In this way, eight pieces of binary image data are combined into a weighted 1
By converting to one pixel, the number of processing steps in image processing is reduced to 178.

FIG. 5 shows a four-part diagram of the image data (original).

The first to fourth blocks are obtained by dividing the fourth drawing image data (original) into four at the positions of X=8 and Y=16. Here, it is divided into four blocks, and one block is read out from the image memory 2 in the direction from S to E in FIG. 3-1 and sequentially performs image processing block by block using the same module as shown in the flowchart of FIG. 2, it becomes possible to perform image processing at extremely high speed. This will be explained below.

FIG. 6 shows an explanatory diagram for extracting a large area of the first block according to the present invention. This is for writing to the accessible internal memory 3-1 and obtaining a large area.

FIG. 6(a) shows the first block original image.

This shows what happens after the first block in FIG. 4 in the image memory 2 is read out in the direction from S to E in the first block in FIG. 3 (b) and stored in the internal memory 3-1.

Figure 6 (b) shows a diagram in which the search is performed from the direction of the arrow, and if 8 is found for the first time, the subsequent parts are filled with 1. The figure below shows that when 8 is found for the first time, it is filled with 1. Here, 8 is the value when there is no hole in the 2x4 pixel area of the binary image data read by the sensor from the medium. It is.

Figure 6 (c) shows a diagram in which the search is performed from the direction of the arrow, and if 8 is found for the first time, the subsequent parts are filled with 1. A diagram is shown in which the search is performed, and when 8 is found for the first time, the subsequent parts are filled with 1.

FIG. 6(d) shows a contour extracted by the AND product of (b) and () , perform an AND operation,
This is an extracted outline.

The outside of the portion 1 is the outline of the medium.

The large area inside the contour line extracted in FIG. 6(d) is calculated using the following formula (1).

Large area = (Figure 6 (d) Total number of dots inside the contour line x 8
)-(Sum of actual data of the original image in FIG. 6(a) corresponding to the inside of the contour line in FIG. 6(d)) (1) Here, the calculation is performed assuming that the large area=15. This large area portion is the shaded portion in the figure.

FIG. 7 shows the contour extraction results of four blocks according to the present invention.

Figure 7 (a) shows the first block, second block, and
The contour extraction results for the third and fourth blocks are shown.
The contours of the blocks and the fourth block were sequentially determined as explained using FIG.

Figure 7 (b) is a summary of Figure 7 (a).

This is a result of dividing the first block, second block, third block, and fourth block in FIG. 7(a) into four, finding the contours of each block, and then summarizing them. As a result, the outline of the fourth cabinet image data (original) is obtained as shown in FIG. 7 (b). There is a hole in the shaded area, and its large area is calculated as the sum of "22" obtained for each of the first to fourth blocks using equation (1) described above.

Processing (for example, processing of large areas of media) can be performed at high speed and with high efficiency.

[Brief explanation of drawings]

FIG. 1 is a principle block diagram of the present invention, FIG. 2 is a flowchart explaining the operation of the present invention, and FIG. 3 is a diagram explaining the operation of the present invention.
Figure 4 shows the image data (original) on the image memory, the 5th rEJ
6 is a four-part diagram of the image data (original), FIG. 6 is an explanatory diagram of large area extraction of the first block according to the present invention, and FIG. 7 is a diagram showing the contour extraction results of four blocks according to the present invention. In the figure, 1 represents a sensor, 2 represents an image memory, and 3-1 represents an internal memory that can be accessed at high speed. 〔Effect of the invention〕

Claims (1)

[Claims] The image data read from the medium is divided into multiple parts so that they have approximately similar shapes, and the reading direction is determined so that all of these divided image data are arranged in approximately the same shape, and the data is read out and stored sequentially in high-speed memory. An image processing method characterized in that the image processing method is configured such that the image processing is performed by the same processing routine.