JPH087114A - Device and method for extracting contour information - Google Patents

Abstract

PURPOSE:To simplify device configuration or processes in the case of extracting the contour vector of image data, to suppress the extension and cost up of the device and to accelerate processing speed. CONSTITUTION:The binary picture elements of raster scanning inputted from an input control circuit 4402 are latched by latches 4403 and 4404, and the corresponding binary picture elements of the preceding line are latched through a FIFO 4407 by latches 4405 and 4406. A CPU 4413 inputs the binary data of these 2X2 pieces of picture elements held by these latches through an input port 4408 and extracts the vector while paying attention on the grid positions of those picture elements.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a contour information extracting apparatus and method, and more particularly to a contour information extracting apparatus and method for extracting a contour line of a binary image stored in raster scanning order.

[0002]

2. Description of the Related Art Regarding the apparatus of this type, the applicant of the present application has already filed Japanese Patent Application No. 2-281958 (hereinafter referred to as "Proposal A") and Japanese Patent Application No. 3-272701 (hereinafter referred to as "Proposition B"). Called).

In Proposal A, a certain pixel is focused on, it is determined whether the focused pixel is a pixel at the contour position, and the focused pixel is determined to be a contour pixel according to the states of the focused pixel and a plurality of neighboring pixels. If it is done, from the relationship with neighboring pixels,
It discloses a method of extracting a horizontal vector and a vertical vector according to the states of a pixel of interest and its neighboring pixels.

According to this method, all the contour lines in the image can be extracted by only one raster scan, and the image memory for storing all the image data is not required, so that the memory capacity can be reduced. have. Further, the contour is extracted not by the center position of the pixel of the input image but by the vector connecting the inter-pixel lattices (that is, it is assumed that one pixel is a square and there are vectors on the four sides). This is a contour line extraction method having a significant width even for a thin line having a width of one pixel. Furthermore, it is introduced that extraction is possible in a 3 × 3 9-pixel region of pixels in the original image.

In the proposal B, the rule for extracting the contour points of the 3 × 3 9-pixel area proposed in the proposal A is set between two black pixels connected diagonally. It is proposed that one contour point be defined at each point when these two black pixels become the target pixel. It is introduced that by adopting this method, the extracting means of each part is configured to operate independently of the states of other surrounding pixels.

[0006]

[Problem to be Solved by the Invention] However, Proposal A
Further, Proposal B has the following problems.

(1) In order to enable extraction in a 3 × 3 9-pixel area, a 3 × 3 9 pixels centered on the pixel of interest
It was necessary to judge the state of black pixels or white pixels for all the pixel areas. At this time, since the processing content is determined for each pattern of each state, it is desirable that the number of pixel state patterns is as small as possible.

(2) Since there are four places where contour points can exist for one pixel of interest within a 3 × 3 9 image area,
The processing rules for each state were complicated, making it difficult to modularize.

(3) Register vector numbers in the order of extraction while retaining vector numbers for which connection relationships are undetermined,
When searching, the stack method was used to search backwards, so the amount of memory wasted and the search took time.

[0010]

[Means for Solving the Problems] and

According to the present invention, the contour information that simplifies the apparatus configuration or the processing process in the case of extracting the contour vector of the image data, suppresses the enlargement of the apparatus, the cost increase, and the processing speed. It is intended to provide an extraction device and method.

In order to solve this problem, for example, the contour information extraction device of the present invention has the following configuration. That is, raster scanning is performed while paying attention to a grid position between pixels of image data, and detection means for detecting a contour vector in the main scanning direction and a sub-scanning direction based on a state of a pixel group near the grid position, It is provided with a discriminating means for discriminating the connection state of the contour vectors and an extracting means for extracting the contour information of the image data based on the connection state of the contour vectors discriminated by the discriminating means.

Here, it is desirable that the neighboring pixel group is a total of 4 pixels of 2 × 2 centered on the inter-pixel grid. According to this, since the number of lattice positions is one, the number of cases that can occur is small, and the device configuration and processing are further simplified.

Further, a horizontal connection buffer which has at least one piece of information at most, in which the main-scanning connection relationship of the contour vector is undetermined, and one of the above-mentioned image data, at which the sub-scanning connection relationship of the contour vector is undetermined. A vertical connection table sufficient to hold the number of the inter-pixel grids existing in the raster; and a determination means for determining the connection state between the contour vectors using the buffer and the table. Is desirable. As a result, the contour information can be extracted with a small memory capacity.

Further, each column of the vertical connection table is
It is desirable to correspond to the position of each of the inter-pixel grids in the image data in the main scanning direction. This simplifies the process when searching for connection relationships.

[0015]

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

In the extraction of the contour vector in the embodiment, as shown in FIG. 1, the processing is advanced by looking at the state of the pixels in the 4-pixel area, and the whole image is shifted by shifting by one pixel in the raster scanning direction. The processing is performed in such a manner that the processing is performed sequentially. Incidentally, "x" in 0101 of FIG. 1 means a point on the inter-pixel grid in the 4-pixel area, where a contour point can exist. This point will be referred to as "attention point" hereinafter.

In the case of an image of m × n pixels, the point on the inter-pixel grid as shown in FIG. 2 is defined as (0, 0) at the upper left and the upper right at the upper right. (M, 0), the lower left is (0, n), and the lower right is (m, n).

The classification corresponding to cases 00 to 15 in the embodiment of the proposal A is classified as shown in FIGS. 10 to 25 in this embodiment. 10 to 25,
The square represents a window of 2 × 2 4 pixel areas. The meanings of ○, △, and arrows in the window are the same as those defined in the embodiment of the proposal A, ○ is the starting point of the horizontal vector (vertical vector end point), and Δ is the starting point of the vertical vector (horizontal vector end point). ) Means. The arrow is the direction of the vector and is also the contour tracking direction (in the embodiment, tracking is performed so that a black pixel exists on the right side of the tracking direction).

When the vectors are extracted as described above, after all, even if the lengths of the contours of the binary image differ, the horizontal vector,
Alternately, vertical vector, horizontal vector, vertical ...

The processing for each pattern is performed according to the flowchart shown in FIG. 48, but here, the processing from steps S4802 to S4806 is performed by predetermining the processing from steps S4802 to S4806 as shown in FIGS. To do.

The processing for each case will be described below, but a table for accumulating processing results is prepared on the hard disk or RAM before these processings.

First, in order to register the extracted vectors, two tables as shown in FIGS. 3 and 4 are prepared.

When the contour point exists at the point of interest, it is determined whether the contour point is the starting point of the horizontal vector or the starting point of the vertical vector. When it is determined to be the starting point of the horizontal vector, the x coordinate of this contour point is registered in the column 0301 of the table indicated by the horizontal vector counter 0300 of FIG. 3, and the y coordinate is registered in the column 0302. In order to store the next vector at this time, the value of the counter 0300 is advanced by 1 (add 1). On the other hand, if it is judged as the start point of the vertical vector,
Register it in the table of FIG. Specifically, the column 040 designated by the vertical vector counter 0400.
The x, y coordinates are registered in 1, 0402, and the value of the counter 0400 prepared for the next data is incremented by one.

In the process of performing the above process, the process shown in FIG.
The table 0501 holds the vector numbers for which the relationship in the vertical connection is undetermined. This is to hold the number 0601 when the end point of the vertical vector 0601 is undecided as shown in FIG. 6 or when the vertical vector (start point) flowing into the horizontal vector 0701 is undecided as shown in FIG. This is for holding the number "0701". If the coordinates of the vector to be held are (x, y), the vector number is registered in the xth position of the table 0501. This table means that when the number of pixels in the main scanning direction of the image is m, it is sufficient to prepare a size that can hold 0 to m-th m + 1 vector numbers.

Further, a buffer 0502 for holding vector numbers whose relationship in the horizontal connection is undetermined is prepared. This is because the number 0801 is held when the end point of the horizontal vector 0801 is undecided as shown in FIG.
Alternatively, as shown in FIG. 9, it is for holding the number 0901 when the start point of the horizontal vector flowing into the vertical vector 0901 is undecided. This buffer may be of a size that can hold one vector number.

Next, each case of the vector extraction processing of the embodiment will be described. In this embodiment, since processing is performed in units of 2 × 2 areas, it is possible to divide into 16 ways (cases 0 to 15) in total.

<Case 0> FIG. 10 shows a state where all the pixels in the 4-pixel region are white pixels (Case 0). in this case,
Since the point of interest cannot be a contour point, no processing is performed and the area is shifted to the next raster.

<Case 1> In FIG. 11, only the upper left pixel of the 4-pixel region is a black pixel and the rest are white pixels (Case 1). In this case, the attention point of the area 1101 is ◯ 110.
2 is a contour point, which is the start point of the horizontal vector. FIG. 26
Shows the flowchart of the processing.

In step S2601, the coordinate (x, y) of the point of interest is set as the starting point of the horizontal vector, and x is registered in the column 0301 of the table in FIG.

Next, since the starting point of the vertical vector 1103 flowing into this contour point is in the area already scanned by raster scanning, the x-th v of the table 0501 in FIG.
It is registered in (x).

Therefore, in step S2602, FIG.
The routine I shown in the flowchart of FIG. 2 is called to clarify the connection relationship. First, in step S3601, v
The number registered in (x) is registered in the attention position column 0303 in FIG. This makes it possible to determine a vertical vector whose end point is the starting point position of the horizontal vector of interest.

Also for the vertical vector table,
Since the same processing is performed, in the next step S3602,
The value of the counter 0300 in FIG. 3 is registered in the column 404 registered in v (x) in the table in FIG.
As a result, the end point position of the vertical vector and the horizontal vector flowing out therefrom are determined.

Returning to FIG. 26, when the processing proceeds to step S2603, the horizontal vector 110 flowing out from this contour point is output.
Search for the end point of 4. The end point of this horizontal vector is a contour point that has already been scanned (in the embodiment, it is assumed that the binary image is scanned in the order of left → right, and when the scanning is completed, the scanning is advanced downward and the same scanning is performed). Horizontal connection undecided buffer 0
The number is already stored in 502. Therefore,
In step S2603, the routine II shown by the flowchart in FIG. 37 is called to clarify the connection relationship.

First, in step S3701, the number registered in the horizontal connection relation undetermined vector holding buffer 0502 (see FIG. 5) is registered in the column 0303 in FIG. In next step S3702, column 04 in FIG.
The value of the counter 0300 is registered in the item of the number registered in the holding buffer 0502 in 04.

Thus, the process proceeds to the final step S2604 of FIG. 26, 1 is added to the counter 0300, and the process of case 1 is completed. Then, the area is shifted to the next raster.

<Case 2> In FIG. 12, only the upper right pixel of the 4-pixel region is a black pixel, and the remaining pixels are white pixels (Case 2). In this case, the attention point Δ120 of the area 1201
2 is a contour point, which is the start point of the vertical vector. FIG. 27
Shows the flowchart of the processing.

In step S2701, the coordinate (x, y) of the point of interest is set as the starting point of the vertical vector, and x is registered in the column 0401 and the value y is registered in the column 0402 of the table.

Next, since the starting point of the horizontal vector 1203 flowing into this contour point is in an unscanned area, step S
At 2702, the routine shown by the flowchart of VII in FIG. 42 is called, and the value of the counter 0400 is registered in the buffer 0502 here.

In step S2703, the end point of the vertical vector 1204 flowing in from this contour point is searched. Since the end point of this vertical vector is the contour point that has already been scanned, it is registered in the x-th v (x) of the vertical connection undetermined table 0501. Therefore, step S270
In step 3, the routine IV shown by the flowchart in FIG. 39 is called to clarify the connection relationship.

First, in step S3901, the number registered in v (x) is registered in the field 0403. In the next step S3902, the value of the counter 0400 is registered in the item of the number registered in v (x) in the column 0304.

Returning to FIG. 27, the final step S270.
At 4, the counter 0400 is incremented by 1, and the processing of case 2 is completed. Then, the area is shifted to the next raster.

<Case 3> FIG. 13 shows a state where the upper half pixels are black pixels (Case 3). In this case, the point of interest does not become the contour point (because it is a point in the middle of the horizontal vector and both the start point and the end point are indefinite and need not be considered), so no processing is performed and the area is expanded to the next raster. Shift.

<Case 4> In FIG. 14, only the lower left pixel of the 4-pixel area is a black pixel and a white pixel (case 4).
Is. In this case, the attention point Δ1402 in the area 1401 is a contour point and is the start point of the vertical vector. FIG. 28 shows a flowchart of the processing.

In step S2801, the coordinate (x, y) of the point of interest is set as the starting point of the vertical vector, and x is set in the position column 0401 indicated by the counter 0400 in the table of FIG.
The value of y is registered in the column 0402.

In step S2802, the starting point of the horizontal vector 1403 flowing into this contour point is searched. Since the starting point of this horizontal vector is the contour point that has already been scanned,
That number is held in the horizontal connection undecided buffer 0502. Therefore, in step S2802, the routine III shown in the flowchart of FIG. 38 is called to clarify the connection relationship.

First, in step S3801, the buffer 05
The number registered in 02 is the counter 0 at that time.
It is registered in the column 0403 of the position indicated by 400. In the next step S3802, the buffer 0502 in the column 0304
To the number registered in the counter 0400
Register the value of.

Next, since the end point of the vertical vector 1404 flowing out from this contour point is in the unscanned area, the routine V shown in the flow chart of FIG. 43 in step S2803.
III is called, where the value of the counter 0400 is registered in the x-th term v (x) of the table 0501.

Thus, the final step S28 in FIG.
At 04, 1 is added to the counter 0400, and the processing of case 4 is completed. Then, the area is shifted to the next raster.

<Case 5> FIG. 15 shows a state where the left half pixel is a black pixel (Case 5). In this case, the point of interest does not become a contour point, so no processing is performed and the area is shifted to the next raster.

<Case 6> FIG. 16 shows a state where the upper right and lower left pixels are black pixels (Case 6). In this case, the area 16
The target point of 01 has two contour points, both of which are considered to be the starting points of the vertical vector. The point 1602 is processed by the same routine as case 14 described later. Point 16
No. 03 is processed by the same routine as case 7 described later. Therefore, the processing of case 6 is shown in the flowchart of FIG. After that, the area is shifted to the next raster.

<Case 7> In FIG. 17, only the lower right pixel of the 4-pixel area is a white pixel and a black pixel (case 7).
Is. In this case, the point of interest Δ1702 in the area 1701 is the contour point, which is the start point of the vertical vector. FIG. 30 shows a flowchart of the processing.

In step S3001, the counter 04 is set with the coordinates (x, y) of the point of interest as the starting point of the vertical vector.
The x is registered in the column 0401 of the position indicated by 00, and the value of y is registered in the column 0402.

Horizontal vector 170 flowing into this contour point
Since the start point of 3 is in the unscanned area, step S300
At 2 the routine VII is called.

Next, since the end point of the vertical vector 1704 flowing out from this contour point is in the unscanned area, routine VIII is called in step S3003.

At the last step S3004, the counter 04
1 is added to 00, and the processing of case 7 is completed. And
Move the area to the next raster.

<Case 8> In FIG. 18, only the lower right pixel of the four pixel area is a black pixel and a white pixel (case 8).
Is. In this case, the attention point “◯” 180 in the area 1801
2 is a contour point, which is the start point of the horizontal vector. Figure 31
Shows the flowchart of the processing.

In step S3101, the counter 030
X in the column 0301 of the position indicated by 0 and y in the column 0302
Register the value of.

Vertical vector 180 flowing into this contour point
Since the start point of 4 is in the unscanned area, step S310
40, the routine V shown in the flowchart of FIG. 40 is called, and the value of the counter 0300 is stored in the table 0501.
It is registered in the x-th term v (x) of.

Next, since the end point of the horizontal vector 1803 flowing out from this contour point is in the unscanned area, the routine V shown in the flow chart of FIG. 41 in step S3103.
I is called, and the value of the counter 0300 is registered in the buffer 0502 here.

In the last step S3104, the counter 03
1 is added to 00, and the processing of case 8 is completed. And
Move the area to the next raster.

<Case 9> FIG. 19 shows a state where the upper left and lower right pixels are black pixels (Case 9). In this case, 1901
There are two contour points at the point of interest, and both are considered to be the starting points of the horizontal vector. The point 1902 is processed by the same routine as the case 13 described later. Point 1903
With respect to the above, the same routine as in case 11 described later is performed. Therefore, the processing of case 9 is shown in the flowchart of FIG. After that, the area is shifted to the next raster.

<Case 10> FIG. 20 shows a state where the right half pixel is a black pixel (Case 10). In this case, the point of interest does not become a contour point, so no processing is performed and the area is shifted to the next raster.

<Case 11> In FIG. 21, only the lower left pixel of the 4-pixel area is a white pixel and a black pixel (Case 1).
1). In this case, the attention point ∘ 210 of the area 2101
2 is a contour point, which is the start point of the horizontal vector. FIG.
Shows the flowchart of the processing.

In step S3301, the coordinate (x, y) of the point of interest is set as the starting point of the horizontal vector, and x is registered in the column 0301 and the value y is registered in the column 0302.

Vertical vector 210 flowing into this contour point
Since the starting point of 3 is in the unscanned area, step SS3
At 302, routine V is called.

In step S3303, since the end point of this horizontal vector 2104 is the contour point that has already been scanned,
Routine II is called.

In the last step S3304, the counter 03
1 is added to 00, and the processing of case 1 is completed. And
Move the area to the next raster.

<Case 12> FIG. 22 shows a state where the lower half pixel is a black pixel (Case 12). In this case, the point of interest does not become a contour point, so no processing is performed and the area is shifted to the next raster.

<Case 13> In FIG. 23, only the upper right pixel of the 4-pixel region is a white pixel and a black pixel (case 1
3). In this case, the attention point “◯” 2 in the area 2301
A contour point 302 is the start point of the horizontal vector. FIG. 34 shows a flowchart of the processing.

In step S3401, the coordinate (x, y) of the point of interest is used as the starting point of the horizontal vector to register x in the column 0301 and the value y in the column 0302 of the table.

Vertical vector 230 flowing into this contour point
Since the start point of 3 is the contour point that has already been scanned, routine I is called in step S3402.

In step S3403, since the end point of this horizontal vector 2304 is in the unscanned area, routine VI is called.

In the last step S3404, the counter 03
1 is added to 00, and the processing of case 1 is completed. And
Move the area to the next raster.

<Case 14> FIG. 24 shows a state where only the upper left pixel of the 4-pixel region is a white pixel and a black pixel (Case 1
4). In this case, the attention point “Δ” 2 in the area 2401
Reference numeral 402 is a contour point, which is the start point of the vertical vector. FIG. 35 shows a flowchart of the processing.

In step S3501, with the coordinates (x, y) of the point of interest as the starting point of the vertical vector, x is registered in the column 0401 and the value y is registered in the column 0402 of the table.

The horizontal vector 240 flowing into this contour point
Since the starting point of 3 is the contour point that has already been scanned, routine III is called in step S3502.

The vertical vector 24 flowing out from this contour point
Since the start point of 04 is the contour point that has already been scanned, routine IV is called in step S3503.

At the final step S3504, the counter 04
1 is added to 00, and the processing of case 1 is completed. And
Move the area to the next raster.

<Case 15> FIG. 25 shows a state where all the pixels in the 4-pixel region are black pixels (Case 15). In this case, the point of interest does not become a contour point, so no processing is performed and the area is shifted to the next raster.

By the procedure described above, one process is performed for each case. Then, steps S4801 to S4807 are repeated until it is determined in step S4807 that the processing for the entire input binary image has been completed.

FIG. 44 is a block diagram showing an example of the hardware configuration for extracting the contour of this embodiment. The programs for all the flowcharts described above are shown in FIG.
Stored in the memory 4414 in
It is executed by the CPU 4413.

In FIG. 44, 4402 is the signal line 4.
An input control unit for exchanging images input via 401, and binary image data is sequentially input from the signal line 4401 in a raster scanning format. A latch 4403 holds the input data from 4402 while sequentially updating the input data pixel by pixel with a pixel synchronization clock (not shown). At the next pixel synchronization clock, the latch 4403 inputs the next pixel data from 4402. At this time, the image data already held is held in the latch 4404 in synchronization with the image clock.

Reference numeral 4407 is a FIFO (first-in first-out memory) that holds pixel data for one raster. The FIFO 4407 sequentially takes in the output of the latch 4404 in synchronization with the pixel synchronization clock and outputs the data one raster before to 4405.

The latches 4405 and 4406 function similarly to the latches 4403 and 4404.

In this way, the latches 4403 and 440
The four pixels stored in 4,4405,4406 are shown in FIG.
This means that the pixel data of the 4-pixel area shown in FIG.

Reference numeral 4408 denotes an input port of the CPU 4413, which inputs the data of the latches 4403, 4404, 4405, 4406. Reference numeral 4410 is a main scanning counter indicating the inter-pixel grid position in the main scanning direction, which is set by a sub scanning synchronizing signal (not shown) and counts up by the pixel synchronizing signal. Reference numeral 4411 denotes a sub-scanning counter which indicates an inter-pixel grid position in the sub-scanning direction, which is reset by a page sync signal (not shown) and counted up by the sub-scan sync signal. Reference numeral 4409 denotes an input / output port for input / output control, a signal for instructing the input / output control circuit 4402 to execute and hold pixel data input, and 4402.
A signal or the like for notifying the CPU 4413 of pixel data update is held. Reference numeral 4415 denotes an input / output device of a hard disk 4416 (used to store the finally extracted contour information). I / O control port 4409,
Main scan counter 4410, sub scan counter 4411, input port 4408, memory 4414, disk I / O
The 4415 is connected to the CPU 4413 via the bus 4412.

In this way, the CPU 4413 updates the pixel data via the input / output control port 4409, knows the window position of the 4-pixel area via the main scanning counter 4410 and the sub-scanning counter 4411, and then the input port 44
It is possible to know the pixel state in the window through 08.

When the CPU 4413 completes one processing,
The pixel data stored in the four latches is instructed to be updated via the input / output control port 4409, and at the same time, the pixel data update completion signal is reset. I / O control port 44
When the update instruction is received, the signal 09 instructs the pixel data update instruction signal to be cleared, and the image data latched by the latch at the subsequent stage to be updated. When this update is completed, the input / output control port 4409 is notified of the completion of the update. Output a signal.

The CPU 4413 monitors the input of the update completion signal from the input / output control port 4409 after issuing the update instruction. When this update completion signal is input, the processing relating to the pixel data newly stored in the four latches is executed, and the same is repeated thereafter. The input / output control port 4409 is used when the input / output control port 4409 finishes processing the last four pixel area of the image area.
An end signal is output to 409.

As described above, according to the present embodiment, by using a window of a total of 4 pixel areas of 2 pixels vertically and horizontally as a scanning unit of raster scanning, there is only one place where an outline point can exist in the window. Limited This reduced the pattern of pixel states in the window and reduced the number of processing modules. In addition, since the processing rule for each state can be simplified, it is possible to obtain the effect of reducing the hardware configuration scale.

[Explanation of the Second Embodiment] In the above embodiment, in the processing of Case 1, Case 2, Case 4, Case 7, Case 8, Case 11, Case 13, and Case 14, the second step and the third step are performed. The order of the second step (the step of calling routines I to VIII) can also be reversed.

However, the processing order of Case 6 and Case 9 cannot be changed.

[Third Embodiment] A connection determining method as introduced in the first embodiment, that is, a buffer for holding vector numbers whose relationship of horizontal connection is undecided is set to one.
In addition, instead of stacking the table holding the vector numbers for which the vertical connection relationship is undetermined, the connection determination method by matching the x-coordinate of the starting point of the vector and the item number of the table is proposed. It can also be implemented in combination with the method of extracting contour points in the 3 × 3 9-pixel region introduced in the embodiment.

In this case, the vector number for which the connection relation is undecided
Tables and buffers for storing
Prepare two sets each. And in the form described below
carry out. For example, a horizontal vector within a 3 × 3 9 pixel area
It is assumed that the starting point of the rule is extracted. Then, with the first embodiment
Similarly, enter the coordinates of the start point in columns 0301 and 0302 of FIG.
Enter. The value of the counter 0300 at that time is p.
The x coordinate of the vector is x_pIt was This horizontal vector
If it is not determined because the end point of the
This is stored in the outflow vector holding buffer 4504 for flat connection undecided.
Hold the vector number p. The end point has already been operated
Inflow area, inflow vector for horizontal connection undecided
The number held in the holding buffer 4503 (if
Since there is a connection relation with the vertical vector of
Enter the bar q in the column of number p of 0304, and enter 0 in FIG.
Enter the number p in the number q field of 403. Next
And the vertical vector with the horizontal vector of this number p as the end point.
If the starting point of the cuttle is in the unoperated area, it cannot be determined.
In case of vertical connection undecided inflow vector holding table 4501
X _pSecondly, the number p is held. Number p water
The start point of the vertical vector that ends at the flat vector has already been manipulated.
If it is in the prefabricated area, outflow vector for vertical connection undecided
Torx holding table 4502 x_pHeld in th
Number (probably q ') vertical vector and connection
Therefore, the number q'is the column of the number p of 0303.
Enter q'in the box and enter the number in the 0q number q'column.
-P.

Similarly, it is assumed that the starting point of the vertical vector is extracted within the 3 × 3 9-pixel area. Then, similarly to the first embodiment, the coordinates of the starting point are entered in the columns 0401 and 0402 of FIG. It is assumed that the value of the counter 0400 at that time is r and the x coordinate of this vector is x _r . If the end point of this vertical vector cannot be determined because it is in the untouched area,
Vertical connection undecided outflow vector holding table 4502 x
_{The r-} th vector number r is held. If the end point is in the already operated area, there is a connection relation with the horizontal vector of the number (provisionally s) held at the x _r- th position in the vertical connection undetermined inflow vector holding table 4501. The number r of 0404
And the number r in the number s column 0303 in FIG. Next, if the start point of the horizontal vector having the vertical vector of the number r as the end point is in the unoperated area, it is not determined, and the horizontal connection undetermined inflow vector holding buffer 4503 is made to hold the number r. If the starting point of the horizontal vector having the vertical vector of the number r as the end point is in the already operated area, the horizontal vector of the number (provisionally s') held in the horizontal connection undetermined outflow vector holding buffer 4504. Therefore, the number s ′ is entered in the column of number r of 0403, and the number s ′ is entered in the column of number r ′ of 0304.

[Fourth Embodiment] Using the method of extracting contour points in the 4-pixel area introduced in the first embodiment, a table for holding vector numbers whose relationship of vertical connection is undecided, It is also possible to use a stack method of registering in the table in the order in which they are extracted and searching backwards when searching.

[Fifth Embodiment] The contour point may be registered in one table without distinguishing between the start point of the horizontal vector and the start point of the vertical vector. In this case, FIG. 46 is used instead of FIGS.
Case 1, Case 2, Case 4, Case 7, Case 8, Case 11, Case 13, Case 14 of the first embodiment.
In the processing of, the counters 0300 and 0400 are all in the counter 4600, columns 0301 and 0401 are all columns 4601, and columns 0302 and 0402 are all columns 4602.
Further, all the columns 0303 and 0403 are replaced with the column 4603, and all the columns 0304 and 0404 are replaced with the column 4606.

However, in case 9, a sub-buffer is prepared and the processing of case 9'shown in FIG. 47 is performed.

In step S4701, the coordinates (x,
y) is registered in the tables 4601 and 4602, and the routine I is called in step S4702. Then, in step S4703, the counter 460 is temporarily stored in the sub-buffer.
After the value of 0 is held and the value of the counter 4600 is incremented by 1, the case 11 is executed, the buffer 0502 is made to hold the number held by the sub-buffer, and the processing of case 9 ′ is ended.

As described above, according to the present embodiment, the scanning unit of raster scanning is set to a window of 4 pixel areas of 2 pixels vertically and horizontally, so that there can be only one place where an outline point can exist in the window. I tried to be. This allows
The pattern of pixel states in the window was reduced and the number of processing modules was reduced. Also, the processing rules for each state have been simplified. Therefore, it is possible to obtain the effect of reducing the hardware configuration scale.

Further, the processing time can be shortened by improving the method for holding the vector numbers whose connection relations are undetermined as described above. Further, in the embodiment of the proposal A, the table for this holding is shown in FIG.
Since the table shown in FIG. 34 is required, the improvement requires the table shown in FIG. 5 and the buffer, so that the memory capacity can be significantly reduced.

The present invention may be applied to a system composed of a plurality of devices or an apparatus composed of a single device. Further, it goes without saying that the present invention can be applied to the case where it is achieved by supplying a program to a system or an apparatus.

In the embodiment, only the outline vector is extracted from the binary image, but the binary image is input from the image reader or from a medium such as a floppy disk storing the binary image. May be obtained from an image obtained through communication (for example, a facsimile reception image), and any means may be used.

Since the extracted vector is composed of only vertical and horizontal vectors, it may be used for variable-magnification print output or facsimile image transmission after undergoing processing such as smoothing processing of the vector. .

[0105]

As described above, according to the present invention, the apparatus configuration or the processing process in the case of extracting the contour vector of the image data is simplified, the size of the apparatus is prevented from increasing, and the processing speed is reduced. It is possible to raise.

[0106]

[Brief description of drawings]

FIG. 1 is a diagram showing a relationship between a scanning direction of an image, a point of interest, and pixels in the vicinity thereof.

FIG. 2 is a diagram for explaining how to take coordinates for expressing the position of a point of interest.

FIG. 3 is a diagram showing an example of a horizontal vector table.

FIG. 4 is a diagram showing an example of a vertical vector table.

FIG. 5 is a diagram showing a table holding vector numbers for which a vertical connection relationship is undecided and a buffer for holding vector numbers for which a horizontal connection relationship is undecided.

FIG. 6 is a diagram showing an example in which the end point of a flowing vertical vector is undetermined.

FIG. 7 is a diagram showing an example in which a starting point of an inflowing vertical vector is undetermined.

FIG. 8 is a diagram showing an example in which the end point of the outflowing horizontal vector is undetermined.

FIG. 9 is a diagram showing an example in which a starting point of an inflowing horizontal vector is undetermined.

FIG. 10 is a diagram showing a state of case 0 in the contour vector extraction processing in the embodiment.

FIG. 11 is a diagram showing a state of case 1 in the contour vector extraction processing in the embodiment.

FIG. 12 is a diagram showing a state of case 2 in the contour vector extraction processing in the example.

FIG. 13 is a diagram showing a state of case 3 in the contour vector extraction processing in the example.

FIG. 14 is a diagram showing a state of case 4 in the contour vector extraction processing in the example.

FIG. 15 is a diagram showing a state of case 5 in the contour vector extraction processing in the example.

FIG. 16 is a diagram showing a state of case 6 in the contour vector extraction processing in the example.

FIG. 17 is a diagram showing a state of case 7 in the contour vector extraction processing in the example.

FIG. 18 is a diagram showing a state of case 8 in the contour vector extraction process in the example.

FIG. 19 is a diagram showing a state of case 9 in the contour vector extraction processing in the example.

FIG. 20 is a diagram showing a state of case 10 in the contour vector extraction processing in the example.

FIG. 21 is a diagram showing a state of case 11 in the contour vector extraction processing in the example.

FIG. 22 is a diagram showing a state of case 12 in the contour vector extraction processing in the example.

FIG. 23 is a diagram showing a state of case 13 in the contour vector extraction processing in the example.

FIG. 24 is a diagram showing a state of case 14 in the contour vector extraction processing in the example.

FIG. 25 is a diagram showing a state of case 15 in the contour vector extraction process in the example.

FIG. 26 is a flowchart showing the processing contents of case 1;

FIG. 27 is a flowchart showing the processing contents of case 2;

FIG. 28 is a flowchart showing the processing contents of case 4;

FIG. 29 is a flowchart showing the processing contents of case 6;

FIG. 30 is a flowchart showing the processing contents of case 7;

FIG. 31 is a flowchart showing the processing contents of case 8;

FIG. 32 is a flowchart showing the processing contents of case 9;

FIG. 33 is a flowchart showing the processing contents of case 11.

FIG. 34 is a flowchart showing the processing contents of case 13;

FIG. 35 is a flowchart showing the processing contents of case 14;

FIG. 36 is a flowchart showing the processing contents of routine I.

FIG. 37 is a flowchart showing the processing contents of routine II.

FIG. 38 is a flowchart showing the processing contents of routine III.

FIG. 39 is a flowchart showing the processing contents of routine IV.

FIG. 40 is a flowchart showing the processing contents of routine V.

FIG. 41 is a flowchart showing the processing contents of routine VI.

FIG. 42 is a flowchart showing the processing contents of routine VII.

FIG. 43 is a flowchart showing the processing contents of routine VIII.

FIG. 44 is a block configuration diagram of an apparatus according to an embodiment.

FIG. 45 is a diagram showing a table holding vector numbers whose vertical connection relations are undecided and a buffer holding vector numbers whose horizontal connection relations are undecided, which are used in the third embodiment.

FIG. 46 is a vector registration table used in the third embodiment.

FIG. 47 is a flowchart showing the processing contents of case 9 ′.

FIG. 48 is a flowchart showing the overall flow of processing.

FIG. 49 is a flowchart showing the processing contents of routine I ′.

FIG. 50 is a flowchart showing the processing contents of routine II ′.

FIG. 51 is a flowchart showing the processing contents of routine III ′.

FIG. 52 is a flowchart showing the processing contents of routine IV ′.

FIG. 53 is a flowchart showing the processing contents of routine V ′.

FIG. 54 is a flowchart showing the processing contents of routine VI ′.

FIG. 55 is a flowchart showing the processing contents of routine VII ′.

FIG. 56 is a flowchart showing the processing contents of routine VIII ′.

[Explanation of symbols]

 4402 Input control unit 4403 to 4406 Latch 4407 FIFO 4408 Input port 4409 Acquisition control port 4410 Main scanning counter 4411 Sub scanning counter 4413 CPU 4414 Memory 4415 Disk I / O 4416 file

Claims (10)

[Claims] 1. A detection means for performing raster scanning while paying attention to a grid position between pixels of image data, and detecting a contour vector in a main scanning direction and a sub scanning direction based on a state of a pixel group near the grid position. A discriminating unit that discriminates the connection state between the contour vectors, and an extracting unit that extracts the contour information of the image data based on the connection state of the contour vectors discriminated by the discriminating unit. A characteristic contour information extraction device. 2. The contour information extracting apparatus according to claim 1, wherein the neighboring pixel group is a total of four 2 × 2 pixels centered on the inter-pixel grid. 3. A horizontal connection buffer, which has at least one piece of information for which the main-scanning connection relationship of contour vectors is undetermined, and one of the above-mentioned image data, for which the sub-scanning connection relationship of contour vectors is undetermined. A vertical connection table sufficient to hold the number of the inter-pixel grids existing in the raster; and a determination means for determining the connection state between the contour vectors using the buffer and the table. The contour information extracting device according to claim 1, wherein: 4. The contour information according to claim 3, wherein each column of the vertical connection table corresponds to a position in the main scanning direction of each of the inter-pixel grids in the image data. Extractor. 5. The image data is raster-scanned while paying attention to each pixel, and a contour vector in the main scanning direction and a sub-scanning direction is formed on an inter-pixel grid surrounding the pixel of interest based on a state of neighboring pixels of each pixel. Detecting means, a horizontal connection buffer sufficient to hold at least one inter-pixel grid information whose connection relationship between the connection source and the connection destination in the main scanning direction is undetermined, and a connection source in the sub-scanning direction. At most 1 of each connection destination
A vertical connection table sufficient to hold inter-pixel grid information corresponding to the number of inter-pixel grids existing in the raster is provided, and the connection state between the contour vectors using the horizontal connection buffer and the vertical connection table. A contour information extraction device comprising: 6. Detection in which raster scanning is performed while paying attention to a grid position between pixels of image data, and contour vectors in a main scanning direction and a sub scanning direction are detected based on a state of a pixel group near the inter-pixel grid. A step of determining the connection state between the contour vectors, and an extraction step of extracting the contour information of the image data based on the connection state of the contour vectors determined by the determination step. A method for extracting contour information. 7. The contour information extracting method according to claim 6, wherein the neighboring pixels are a group of four 2 × 2 pixels centered on the inter-pixel grid. 8. A horizontal connection buffer, which has at least one piece of information for which the main scanning connection relationship of the contour vector is undetermined, and one of the above-mentioned image data, for which the sub-scan connection relationship of the contour vector is undetermined. A vertical connection table sufficient to hold the number of the inter-pixel grids existing in the raster, and a determination step of determining the connection state between the contour vectors using the buffer and the table. The contour information extracting method according to claim 6, wherein. 9. The contour information extracting method according to claim 8, wherein each column of the vertical connection table corresponds to a position of each of the inter-pixel grids in the image data in the main scanning direction. 10. An image data is raster-scanned while paying attention to each pixel, and based on a state of neighboring pixels of each pixel, a contour vector in a main scanning direction and a sub-scanning direction is formed in an inter-pixel grid surrounding the pixel of interest. Detection process, a horizontal connection buffer sufficient to hold inter-pixel grid information for which the connection relationship for each of the connection source and connection destination in the main scanning direction is undetermined, and the connection source in the sub-scanning direction. At most 1 of each connection destination
A vertical connection table that is sufficient to hold inter-pixel grid information corresponding to the number of inter-pixel grids existing in the raster is used, and the connection state between the contour vectors using the horizontal connection buffer and the vertical connection table. And a step of discriminating the contour information extraction method.