JPS61286985A - Outline tracing system for picture - Google Patents

Abstract

PURPOSE:To lower the capacity of a memory and to trace a picture outline part at high speed by treating a binary picture as the aggregation of the sequence of points information in a prescribed direction and extracting a picture outline. CONSTITUTION:The binary picture information is detected as the segment of aggregation in an X direction or the like by the aggregation detecting means 5 of an outline tracing device 1, and its initial and final point coordinates are stored in a memory means 8. The prescribed operation of the initial end coordinate of a noticed segment and the coordinate of the comparison segment of the next line is carried out by a line direction adjacent relation judging means 6, the existence of the comparison segment and the position relation of the noticed segment and the comparison segment when the comparison segements is present or the like are judged. As for the coordinate of a final end of the noticed segment, similarly, based on the judging result of the means 6, a boundary point between the segments is deter mined by the first boundary point memory and notice point changing means 9 and stored in the memory means 8. Similarly to a column direction boundary, in accordance with a judging output that all the segments from a noticed point adoption judging means 11 are considered to be the noticed segments, the boundary coordinate is outputted from the means 8 and the picture outline is traced at high speed with a reduced memory capacity.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image contour tracking method for tracing the contour of a binary image.

[Conventional technology]

Conventional image contour tracking methods treat a binary image as a collection of pixels that make up the image, so first, a boundary search is performed for each pixel that makes up the binary image to find the boundary. It has been common practice to extract points and then perform contour tracking by searching for connectivity for each of the extracted boundary points.

[Problem that the invention seeks to solve]

However, in the conventional contour tracking method described above, 2
Search for boundary points for each pixel constituting the value image, extract boundary points such as the upper point sequence, the lower point sequence, the left point sequence, and the right point sequence, and store these, and then calculate the connectivity of each boundary point. Since boundary point sequence data is created by tracking the boundary points, there are problems in that it requires a long processing time and requires a large amount of memory capacity.

Therefore, this invention was made by focusing on the problems of the conventional example, and extracts the outline of an image input as a binary image by treating the binary image as a collection of point sequence information in a predetermined direction. However, it is an object of the present invention to provide an image contour tracking method that can simultaneously process contour information in which boundary point data is arranged in order in the circumferential direction at high speed and reduce memory capacity.

[Means for solving problems]

In order to achieve the above object, the present invention provides an aggregate detection means for detecting a binary image as an aggregate of point sequence information in the column direction, and focusing on a predetermined aggregate and positioning its starting point coordinates and ending point coordinates. a row direction adjacency relationship determining means for determining the presence or absence of a first aggregate adjacent to the first aggregate with a predetermined relationship in the row direction based on the relationship; and when the determination result of the row direction adjacency relationship determining means is the first aggregate; a column direction adjacency relationship determining means for determining the presence or absence of a second aggregate that is adjacent to the target aggregate with a predetermined relationship in the column direction; a storage means for storing boundary point information; and a row direction adjacency relationship determining means ( and boundary point storage/point-of-interest changing means for storing boundary point coordinates in the storage means based on the determination results of the column-direction adjacency relationship determination means and making the opposite end point a new point of interest; and the binary image. and a point of interest determination means for determining whether or not the starting point coordinates and the end point coordinates of each aggregate are adopted as the point of interest, based on the boundary point information stored in the storage means and the connection information of the points of interest. , is characterized by tracking the contour of a binary image.

[Effect]

This invention detects a binary image as a collection of point sequence information using an collection detection means, and uses a row direction adjacency relationship determination means to detect a binary image as a collection of point sequence information.
Focusing on a predetermined aggregate among each aggregate, it is determined whether there is a first aggregate adjacent thereto with a predetermined relationship in the row direction based on the positional relationship of its starting point coordinates or end point coordinates. Then, as a result of the determination by the row direction adjacency relationship determination means, if there is no first aggregate, the boundary point storage/objective point changing means converts the point sequence information on the opposite side of the focused aggregate to the boundary point information. The data are sequentially stored in the storage means, and the row direction adjacency relation determination means executes determination by using the other end as a new point of interest. Furthermore, when the row direction adjacency relationship determining means determines that there is a first aggregate, the column direction adjacency determining means determines whether there is a second aggregate that is adjacent to the focused aggregate with a predetermined relationship in the column direction. . As a result of the determination, if there is no second aggregate, the boundary point storage/point of interest change means sequentially stores a series of points extending outward from the point of interest of the first aggregate in the storage means as boundary point information; The end point is used as a new point of interest for determination by the row direction adjacency relationship determining means. Further, if the column direction adjacency relationship determining means determines that there is a second aggregate, the boundary point storage/point of interest changing means stores point sequence information between the first aggregate's point of interest and the second aggregate. are sequentially stored in the storage means as boundary point information. In this way, by sequentially storing the connection relationships between the start points and end points of adjacent aggregates in the storage means, the boundary point information in the circumferential direction of the image is stored in the storage means as outline information.

In addition, the point of interest adoption determination means determines whether the starting point and end point of each aggregate constituting the binary image have been adopted as the point of interest, and if there is a starting point or ending point that has not been adopted, the adjacent point in the row direction is The determination by the relationship determination means is repeated, and when the start and end points of all the aggregates are adopted as points of interest, the contour tracking is terminated.

〔Example〕

An embodiment of the present invention will be described below with reference to FIG.

In FIG. 1, reference numeral 1 denotes a contour tracking device into which image information is input, and a binary value represented by two values of white and black, which do not include halftones, is inputted from an image input device 2 such as an imaging device. At the same time, contour information resulting from contour tracking is output to the image output device 3.

The contour tracking device 1 includes an aggregate detection means 5 that detects image information from the image input device 2 as an aggregate (hereinafter referred to as a segment) S in a predetermined direction (X direction) as shown in FIG.
Row direction adjacency relationship determination means 6, column direction adjacency relationship determination means 7
, storage means 8 for storing boundary point information (boundary point coordinates), first and second boundary point storage/point of interest change means 9.10, and point of interest adoption determination means 1).

When image information 13 represented by dots on the intersections of the vertical and horizontal grids 12 shown in FIG. A collection of dots is detected as one segment S as shown by the chain line, and the coordinates of their starting point st and ending point en are stored in a predetermined storage area of the storage means 8.

In the initial state, the row direction adjacency determining means 6 searches for a segment S existing at the minimum Y coordinate position shown in FIG. 6, sets the starting point st of this segment S as the first boundary point SN, and Every time a new point of interest ◎ is set,
It is determined whether or not there is an adjacent segment with a predetermined relationship in the row direction regarding the point of interest SN, ■.

The predetermined relationship in this case is the example shown in FIG. 4, where the target segment S! Starting point st of (i-1, 2...M)
and the coordinates of the end point en as (Xis, Yt) and (
XiE+Yi ), and the first comparison segment C3lA or C3I1 . starting point st
and the coordinates of the end point en (X js, Y = + 1
) and (Xj,.

Yi+1) or (Xjs, Yt 1) and (Xjt
, Yi-1), when the starting point st of the target segment S5 is the target point ◎, the first comparison segment C
Coordinates of start point st and end point en of 3lA<Xas, Y
i +1) and (Xji, Y51), and determine whether they satisfy either of the following equations (1) and (2).

Xl-1≦Xj3≦X1+1 ・・・・・・・・・・・・
・(1) Xl-1≦xji≦Xts+1 ・・・・・・
・・・・・・・・・(2) At this time, the above (1) 2 or (2)
) If there is no first comparison segment CS, A that satisfies the equation, it can be determined that the state shown in FIG. 5(d) exists. If the first comparison segment C81A exists, it is determined that the state is in any one of (a) to (C1) in FIG. 5, and then the starting point coordinates (Xjs, Yi +1) satisfies the following equation (3) and the end point coordinates (Xj't, Yt + 1) satisfy the following equation (4).

Xjs≦Xts 1 ・・・・・・・・・・・・(
3) X, t≧Xts 1 ・・・・・・・・・・・・
- (4) At this time, if there is a first comparison segment C3lA that satisfies both equations (3) and (4) above, it is possible to determine that the state is as shown in FIG. 5(a) or (C1). If the condition is satisfied and the condition is not satisfied, it can be determined that the condition is as shown in FIG. 5(b).

Furthermore, when the end point en of the target segment Si is the target point ◎, the coordinates (Ls, Yi-1) and (Xjt,
Yt 1) is read, and it is determined whether these satisfy the above-mentioned formula (1) or (2), and if there is no first comparison segment CS that satisfies the above, then It can be determined that the state is (1)
The first comparison segment C3+1 that satisfies the formula or (2) formula
) exists, it is determined that the states shown in FIG. It is determined whether the following equation (6) is satisfied and the end point coordinates <xjt, yt i) are satisfied.

Xjs≦Xts 1 ・・・・・・・・・・・・(
5) X j E≧Xi, -1・・・・・・・・・・・・
(6) At this time, if there is a first comparison segment C3lIl that satisfies both equations (5) and (6) above, it can be determined that the state shown in FIG. 5(e) or (phantom) exists. , if not satisfied, it can be determined that the state shown in FIG. 5(f) is reached.

The column direction adjacency relationship determination means 7 determines whether the target segment Si has a predetermined value in the column direction (X direction) when the determination result of the row direction determination means 6 is that a first comparison segment C81A or C3+++ having a predetermined adjacency relationship exists. It is determined whether there is a second comparison segment C8n or C32m that is adjacent in relation to each other.

The predetermined adjacency relationship in this case is that when the starting point st of the segment of interest Si is the point of interest ◎, the second comparison segment c
It is determined whether or not the end point coordinates (Xmt, Yt) of sza satisfy the following equation (7).

Xjs 1≦XkE≦Xts-2・・・・・・・・・
... (7) At this time, if there is no second comparison segment C32A that satisfies the above formula (7), then as shown in FIG.
It is determined that the state is a), and the second comparison segment C3z
When a exists, it can be determined that the state shown in FIG. 5(C) exists.

Furthermore, when the end point en of the segment of interest S1 is the point of interest, the starting point coordinates (X
ms, Yi) satisfies the following equation (7).

X+t+2≦xms≦Xj, +1 ・・・・・・・・・
...(8) At this time, the second
Comparison segment C8! When ll does not exist, the fifth
(It is determined that the state is el, and the second comparison segment C
When 8□ exists, it can be determined that the state is in a phantom state (see Figure 5).

The first boundary point storage/point of interest change means 9 stores the point of interest when the judgment result of the row direction adjacency relation judgment means 6 is in the state shown in FIG. 5(d) or (or FIG. 5(h)). The coordinates on the intersection point 14 of the grid 12 up to the end point en (or start point st) on the opposite side from ◎ are sequentially stored in the storage means 8 as boundary points △, and the end point en (or start point st) of the segment Si of interest is drawn as a chain line. When the new point of interest shown in the figure is set to ◎ and the determination result of the column direction adjacency relationship determining means 7 is in the state shown in FIG. 5(a) (or FIG. 5(e)), the first comparison segment C
3lA focus point ◎ to the left (or first comparison segment C
The coordinates on the intersection point 14 of the lattice 12 of the point of interest ◎ of 3+s (on the right side from the point of interest ◎) are sequentially stored in the storage means 8 as a boundary point △, and the starting point st of the first comparison segment C3lA (or the end point en of the first comparison segment) CSzi ) is set as the new point of interest ◎ shown by the chain line.

The second boundary point storage/point of interest change means 10 performs the first comparison when the judgment result of the column direction adjacency relation judgment means 7 is in the state shown in FIG. segment C3
Point of interest of lA (or C3+s) ◎ and second comparison segment) End point en of C3zA (or second comparison segment C3
The coordinates on the intersection 14 of the grid 12 between the starting point st) of tm are sequentially stored in the storage means 8 as boundary points Δ, and the ending point en of the second comparison segment CS t^ (or the starting point st of the second comparison segment) C3z++ ) is set as the new point of interest ◎ shown by the chain line.

The point of interest adoption determining means 1) determines whether or not the start point st and end point en of each segment constituting the binary image are adopted as the point of interest ◎, and all the start points st and end points en are adopted as the point of interest ◎. If so, the contour information stored in the storage means 8 is output to the image output device 3, and if there is a start point st or end point en that has not been adopted, the start point st or end point en is set as the point of interest ◎. .

Next, the operation of the above embodiment will be explained with reference to the flow chart showing the processing procedure in FIG. Now, suppose that image information @13 shown in FIG. 3 is input from the image input device 2 to the contour tracking device 1, the contour tracking device 1 generates a Segment S+ (i=1
．． 2...20), and its starting point s
t and the coordinates of the end point en are stored in the storage means 8, and based on these the processing shown in FIG. 7 is started.

That is, first, the segment S constituting the image information 13
! Search for the segment existing at the minimum Y coordinate Y3 in
This segment is set as the first segment S1 of interest (step 2).

Then, the starting point coordinates (X,
□, Y, ) and end point coordinates (XI4.Ys), and set the starting point coordinates (X+z, Y3) to the first boundary point SN.
It is stored in the storage means 8 as a point of interest, and is set as a point of interest ◎ (step ■).

Next, it is determined whether or not there is a segment in the column whose Y coordinate is "1" larger than the segment of interest St for which the first comparison segment CS, A satisfies the above formula (1) or (2). (Step ■).

In this case, the first equation that satisfies both equations (1) and (2) is
Since there are segments S and l that are the comparison segment C3lA, proceed to step ■ and select this segment S3.
Determine whether or not satisfies the above equations (3) and (4),
However, since segment ) Sz satisfies both equations (3) and (4), proceed to step (2) and check whether there is a segment that becomes the second comparison segment C3th that satisfies equation (7). Determine.

At this time, since there is no segment that becomes the second comparison segment C3Z&, it is determined that the state shown in FIG. The coordinates (X1)+Y4) of the point Δ are stored in a predetermined storage area of the storage means 8, and the starting point st of the segment S3 is set as a new point ◎ of interest (step 0), and then the process returns to step ◎.

At this time, since there is a segment S4 that satisfies equation (2) below segment S3, which is the segment of interest,
Shifting to step ■, this segment S4 is
Since formula and formula (4) are satisfied, move to step ■,
Since there is a segment S2 which becomes the second comparison segment C80 that satisfies the equation (5), it is determined that the state shown in FIG. ., ys), and (X9.YS) are stored in a predetermined storage area of the storage means 8 in that order.

Next, proceed to step ■ and compare the second comparison segment C.
End point coordinates of segment S2 (Xll, Y4
) is set as the target segment, and then moves to step (3) to execute end point target processing.

Here, since there is no first comparison segment that satisfies equation (1) or (2), we move to step [phase] and calculate the left boundary point coordinates (from the point of interest ◎ (starting point) of segment) Sz. X?, Ya), (Xa, Ya) and (X
s, Ya) are temporarily stored in a predetermined storage area of the storage means 8 in that order, and then the starting point st of the segment S is set as a new point of interest ◎.
It is determined whether the set point of interest ◎ is equal to the first boundary point SN set first (step ■).

Here, since the new point of interest ◎ is different from the first boundary point SN, the process returns to step ■ and executes the start point focusing process from the point of interest ◎ (starting point st) of the segment S2.

At this time, the segment (S4) is below the segment of interest S2.
exists, and its starting point st and ending point en meet the above (3)
Since the formula and (4) are satisfied, the process moves to step ■ via step ■. Furthermore, since the second comparison segment C82A does not exist, it is determined that the state shown in FIG. 5(a) is reached, and the process moves to step @. Then, temporarily store the boundary point coordinates (X4.YS) to the left of the point of interest ◎ of the segment S4 in a predetermined storage area of the storage means 8 (step @), and then set the starting point st of the segment S4 as the new point of interest ◎. Step @), return to step ■.

In this case as well, since there is a segment S, which is the first comparison segment, below the target segment S4, the process moves from step ■ to step [phase] via steps ■ and ■.
A boundary point △ on the left side of the point of interest ◎ is stored in the storage means 8.

In this case, since there is no boundary point △ on the left side of the point of interest ◎, the process directly proceeds to step @ and stores the starting point st of the segment Ss as a boundary point in the storage means 8.
Set this as a new point of interest ◎ and then return to step ■.

Similarly, sequential segments S6. S? , S9゜SI
In step ■ after setting the start point st of l + S I ff to the point of interest ◎ and setting the start point st of segment SI3 to the point of interest ◎, the segment S + s that becomes the first comparison segment CS, A is calculated by formula (1). and (2) are not satisfied, it is determined that the state is dl in FIG. ,Yz
) and (X?, Y++) are stored in the storage means 8 in that order, and then the end point en of the segment Sl+ is set to the point of interest ◎, and the process moves to the subsequent end point focusing process in step [phase].

When this end point focusing process is executed, segment 5131
Since sll, So + 37 have the relationship shown in (f) in Fig. 5, the processes of steps [phase], ■, ■, @ are repeated, and the segment S1) + 59 *
S? The end point en of is stored in the storage means 8 as a boundary point △, and this is set as a new point of interest ◎, and the segment S.

After the end point en is set as the point of interest ◎, the state is as shown in FIG. Coordinates of the boundary point on the right side and before the starting point st of segment Sll (Xs, Yy), (X9, Yt
) are stored in a predetermined storage area of the storage means 8 in that order, the start point st of the segment S8 is set as the point of interest ⊚, and the start point focus processing from step ◎ is executed.

Therefore, sequentially segments 5oon sll and SI
The starting point st of 4 is set as the point of interest ◎. Then, in step ■ after setting the starting point st of segment SI4 as the point of interest ◎, segment SI4 and segment)
Since S's has the relationship shown in FIG.
Coordinates of the boundary point on the left from the point of interest ◎ (X, ., 'to' +
z), (X9, Y'+□), (Xs
, Yrt).

(X?, Yl□), (X!, Ylri,
(Xs, Yrt), (X4.Yrt), CXs
, '/+z) and (X2°Yrt) are stored in the predetermined storage area of the storage means 8 in that order, and the starting points st of the segments S and S are set as points of interest ◎ (step [phase]).

After that, the starting point st of the segment Sl&+ 317 is set as the point of interest ◎, and the segment S1. The boundary point coordinates (X
4, Yl4) is stored in the storage means 8, and then the start point st of the segment 319 is set as the point of interest ◎, and the coordinates of the boundary point on the right side thereof are stored, and then the end point en of the segment 319 is set as the point of interest ◎. .

In this way, the outer side that constitutes the image 13 is
The boundary points of the I contour are sequentially stored in a predetermined storage area of the storage means 8, and the end point e of the first focused segment S is
If n is set as the point of interest ◎, in this case, Fig. 5 fh)
Since the state is, step [phase] to step [phase]
, the coordinates of the boundary point on the left side of the point of interest ◎ are sequentially stored in a predetermined storage area of the storage means 8, and then the coordinates of the boundary point on the left side of the point of interest
Set the starting point st of 1 to the point of interest ◎ (step [phase])
. Next, the process moves to step ■, and it is determined whether or not the point of interest ◎ is the first boundary point SN. Since they match, the process moves to step [phase], where the point of interest ◎ is determined as the first boundary point SN. It is determined whether or not there are unused starting points st and ending points en as the points of interest ◎. In this case, since there are no unused starting points st and ending points en, proceed to step [phase]. Then, the boundary point information stored in the storage means 8 is outputted as contour information to the image output device 3, and after displaying this, the process ends.

Also, if there is an unused start point st or end point en,
Moving to step [phase], it is determined whether the unused end point is the starting point st, and if it is the starting point st,
Return to step ①, and when the end point en is reached, return to step [phase] and add these to the unused start point st and end point en.
Repeat until no longer exists.

As a result of the above processing, each boundary point Δ constituting the contour of the input image l is sequentially stored in the circumferential direction in the storage means 8, and the contour information of the input image '1 is stored according to the stored contents. It will have been formed.

Here, steps ■, ■, ■, ■ and steps [phase],
Processing of step (2) corresponds to the row direction adjacency relationship determination means 6, step (2) and step [phase] corresponds to the column direction adjacency relationship determination unit 7, and steps (2), (2) and step 0. [phase]
The processing corresponds to the second boundary point storage/objective point changing means, and the processing of steps ① to 0 and steps 0.0, 6 to [phase] corresponds to the first boundary point storage/objective point changing means 9. , the processing of steps ① to [phase] corresponds to the point of interest adoption determination means 1).

Note that in the above embodiment, the image 13 having a hollow portion
In the case of Since the information is stored twice, accurate contour tracking information of the image can be obtained by editing to detect and delete the twice-stored boundary points from the contour information.

In addition, in the above embodiment, processing was explained in which each pixel constituting an image is treated as an aggregate (segment) in the column direction (X direction), but the process is not limited to this (, row direction (Y direction) ) can also be treated as a collection.

〔Effect of the invention〕

As explained above, according to the present invention, the input image is handled as a collection of segments in a predetermined direction, and the connection relationship with the start and end points of adjacent segments is determined to search for boundary points. , it is possible to perform contour tracing of an image at high speed, and it is also possible to significantly reduce the storage capacity of the storage memory.Furthermore, since the boundary points are memorized by determining the connection relationship between the start and end points of adjacent segments, , the storage order is in the circumferential direction of the input image, and there is an advantage that there is no need to convert the boundary point information into circumferential contour tracking information.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a block diagram showing an example of a contour tracking device, FIG. 3 is an explanatory diagram showing an example of an input image, and FIG. 4 is a block diagram showing an example of an input image. An explanatory diagram of the relationship, FIGS. 5(a) to 5(h) are explanatory diagrams providing a detailed explanation of the present invention, and FIG. 6 is an explanatory diagram showing an image in which the input image is detected as a collection of point sequence information. 7 are flowcharts showing the processing procedure of the image processing apparatus. In the figure, l is a contour tracking device, 2 is an image input device, 3 is an image output device, 5 is an aggregate detection means, 6 is a row direction adjacency relationship determination means, 7 is a column direction adjacency relationship determination means, and 8 is a storage means ,
9 is a first boundary point storage/point of interest change means; 10 is a second boundary point memory/point of interest change means; 1) is a point of interest adoption determination means;
12 is a grid, 13 is image information, 14 is an intersection, ■ is a point of interest, and Δ is a boundary point.

Claims (2)

[Claims] (1) An aggregate detection means that detects a binary image as an aggregate of point sequence information in the column direction; a row direction adjacency relationship determining means for determining the presence or absence of a first aggregate that is adjacent to each other with a predetermined relationship; a column direction adjacency relationship determining means for determining the presence or absence of a second aggregate adjacent to each other with a predetermined relationship, a storage means for storing boundary point information, and determination results of the row direction adjacency relationship determining means and the column direction adjacency relationship determining means. boundary point storage/point of interest changing means for storing boundary point coordinates in the storage means based on the above, and setting the opposite end point as a new point of interest; and starting point coordinates and end point coordinates of each set of the binary images. a point of interest determination means for determining whether or not a point of interest has been adopted as a point of interest, and tracing the contour of the binary image using the boundary point information stored in the storage means and the connection information of the points of interest. An image contour tracking method featuring: (2) The boundary point storage/objective point changing means is configured so that the determination result of the row direction adjacency relationship determination means or the column direction adjacency relationship determination means is the first
Or, when the second aggregate is not present, the point sequence information in a predetermined direction from the point of interest of the focused aggregate or the first aggregate is stored in the storage means as boundary point information, and the end point on the opposite side is stored as boundary point information. When the judgment result of the first boundary point storage/point-of-interest change means to be a new point of interest and the column direction adjacency relation judgment means is that there is a second aggregate, the point of interest of the first aggregate and the second The point sequence information between the aggregates is stored in the storage means as boundary point information, and the second boundary point storage/objective point changing means stores an end point on the opposite side of the second aggregate as a point of interest. An image contour tracking method according to claim (1).