JPS6277691A - Contour line extracting device - Google Patents

Abstract

PURPOSE:To extract contour line data correctly from a contour point data group containing noises by scanning a closed loop binary image by a contour line extracting device and extracting nodes and branches of a graph from the contour point data group obtained in screen scanning order. CONSTITUTION:The contour point data sequence consisting of point numbers (x) coordinate values and (y) coordinate values stored in a contour point data memory 1 is read out by a graph generating circuit 2 to generate connection information on respective points, which is stored in a connection information memory 4. Further, information on a node which consists of branch information composed of a point number, a positive branch, and a negative branch and graph information consisting of a starting node, a next point, an end node, a positive adjacency, a negative adjacency, and length are generated and stored in a graph storage memory 3. A closed loop forming circuit 5 generate a closed loop having the contour of a body as contour data on the basis of the connection information and graph information and stores it in a contour data memory 6.

Description

DETAILED DESCRIPTION OF THE INVENTION [Summary] In a contour extraction device that extracts contour data along a closed-loop contour from a group of contour point data obtained in screen scanning order by scanning a closed-loop binary image.

The graph generation circuit extracts graph nodes and graph branches, and then traces two contours to correctly extract contour line data from a group of contour point data containing noise. Disclosed.

[Industrial application field]

The present invention provides a contour extraction device, in particular, a device for correctly extracting contour data by scanning a closed-loop binary image and extracting nodes of a graph and branches of a graph from a group of contour point data obtained in the order of screen scanning. The present invention relates to a contour line extraction device.

[Conventional technology]

The present inventor improved the invention previously disclosed in Japanese Patent Application No. 58-225622 to obtain a contour extraction device capable of processing at high speed, and filed a patent application as Japanese Patent Application No. 59-177678.

The latter patent application focuses on the fact that one contour extraction target is a binary image forming a closed loop.

detecting that there are at least two points during the raster scan that correspond to a line segment separating from the point consisting of the points, and checking whether these two points are connected to the line segment of the respective side route; When connected, the line segments are extended, and finally, by connecting the ends of the two line segments, 1
We are trying to determine one closed loop.

That is, when there is a contour binary image as shown in FIG. 5(A), contour point data groups as shown in FIG. 5(B) correspond to the raster scan: ■, ■, ■.

■...The data are obtained in the following order, but for the contour point data group, trace the line molecules 1 and T2 that are extended by one separation as shown in Figure 5 (C). , and finally, the ends of line molecules 1 and T2 are connected to determine one closed loop.

[Problem that the invention seeks to solve]

Conventionally, it has been proposed to perform the above-mentioned processing.

If noise is mixed into the two-set image as shown in FIG. 5(A), resulting in a binary image as shown in FIG. 6(A), line molecules as shown in FIG. 6(B) will be generated. 1 and T2 are generated, making it impossible to connect the end of line molecule 1 and the end of line molecule 2, and as a result, it becomes impossible to obtain one closed loop.

Also, when obtaining a contour in the shape of a so-called snowman as shown in FIG. 7, two illustrated line molecules I, T2. T3 will be generated, making it difficult to determine a closed loop.

Means for Solving Problem C] The present invention solves the above points, and as shown in FIG. ．． N3. N4... and branch B of the graph
l, B2. B3. B4. B5... is extracted, information is attached to each node and each school, and extracted as graph information, and a closed loop is created from the relationship between each node and each school, for example.

I try to decide as follows.

FIG. 1(A) shows a basic configuration diagram of the present invention. 1st
In Figure (8), 1 is a contour point data memory, 2 is a graph generation circuit, 3 is a graph storage memory, 4 is a connection information memory, and 5 is a closed loop forming circuit.

6 represents a contour data memory.

The contour point data memory 1 stores a group of contour point data corresponding to a binary image as shown in FIG. 6(A). The graph generation circuit 2 receives the two contour point data groups, and
Nodes Nl and N of the graph shown in Figure 1 (B) and Figure 1 (C)
2. ...and graph techniques Bl, B2...
and generate. The resulting graph information is stored in the graph storage memory 3. On the other hand, 1. While obtaining the above graph information,
It is possible to obtain connection information regarding the connection relationships among the points constituting the contour point data group.

The linked joy is stored in the linked information memory 4.

The closed loop forming circuit 5 determines a closed loop by determining, for example, that nodes N3 and N2 shown in FIG. 1(B) are connected by a branch B3 based on the node, branch, and connection information. .

[Effect]

The contour point data memory 1 contains 1 as shown in FIG. 2(A).
Points composing the data group obtained in the order of screen scanning ■, ■, ・
.

In the graph generation circuit 2, first, nine points are connected to point ■, nine points ■ are connected to point ■, etc. (by examining the connection relationship of each point, As shown in FIG. 2(B), connection information is generated and stored in the connection information memory 4.The process of obtaining this connection information is as follows.

Do as follows.

In other words, 1, for example, when point ■ is found, a point adjacent to point ■ whose y coordinate value is equal and whose X coordinate value is larger (
For example, assume that point ■) is extracted and upper line molecules 1 and T2 start from point ■ and point ■. Regarding point 9 ■, the y-coordinate value and the x-coordinate value are within the predetermined threshold,
Also, a point having a smaller y-coordinate value and a smaller x-coordinate value (for example, a point ■) is extracted. Regarding point ■, a point (for example, point ■) having a smaller y-coordinate value and a larger x-coordinate value is similarly extracted.

By obtaining the above connection information, the points ■, ■, ■, ■ corresponding to the nodes of the graph are obtained, and the nodes Nl, N2, . N3. Get N4. Corresponding to these sections, "point numbers", "seiwaza", "
Information such as "food core" is given. On the other hand, line segments corresponding to the branches of the graph are obtained, and branch B1.
B2゜B3. Get B4. Corresponding to these branches, the "first node", "second node", "final node", and "regular neighbor" shown in Figure 2 (D) are shown.
, "fish scale", and "length" are given. The information in FIGS. 2(C) and 2(D) is stored in the graph storage memory 3.

The closed loop forming circuit 5 shown in FIG. 1(A) determines the existence of a closed loop based on the contents of the memory 1.3.4, and stores the contour line data corresponding to the closed loop in FIG.
E) Obtain as shown and store in the contour data memory 6 shown in FIG. 1(A).

〔Example〕

FIG. 3(A) shows a flowchart of graph generation processing in the graph generation circuit shown in FIG.

In the illustration in FIG. 3(A), the illustration in FIG. 3(B) is
In the processing up to y=yi, the "endpoint of the technique in the branch list" has been obtained, and in the first step y=yi-1, the "point to be cut out from the contour point data" ( In general, it indicates how to process multiple items).

To briefly comment on it, in the process (a) shown in FIG. Check the change in the y-coordinate value of the next point (let's call it Nn (P+1)) based on that point (let's say P) as a reference point.Next, use the point N11 (P+1) as a reference. Next, examine changes in coordinate values for the cut out point (temporarily 1kL (P+2))...
I'm going to go.

In process (b), it is checked whether the point is connected to an edge on y=yi. - In process (c), the connection procedure is performed under the relationship shown in Figure 4.

For example, in the example shown in FIG. 3(A), in terms of the relationship between one technique and one point, the X coordinate value of the right edge (α) of the illustrated technique and the right edge of the illustrated point ( It represents the process of comparing the X coordinate value of β).

The illustrated process (e) and process (f) generate a starting point, and the process (g) and process (h) generate an end point.

FIG. 4 is an explanatory diagram summarizing and schematically showing the connection procedure. "Before connection" shown in FIG. 4 indicates the positional relationship between the branch found at y=yi and the point cut out at y=yi-1, and "after connection" shown in FIG. The result of the connection process performed in process (c) is schematically shown.

Note that the point cut out in the above is not limited to one dot, but can be a number of dots that have the same y-coordinate value and consecutive X-coordinate values that are grouped together into one point. This is called the above point.

Also, in the figure, "○" in "Are the branches already connected?" indicates that the end point of the branch on y=yi can be connected to a point on y=yi or y=yi-1. This corresponds to cases where the situation is already known.

“×” corresponds to the case where this is not the case.

In "Direction of branch end point", "left" corresponds to the case where the branch extends to the left due to the relationship from the previous process, and "right" indicates that the branch extends to the right. This corresponds to cases where there are.

In "Are the points already connected?", what is indicated by "○" indicates that the point on +y-yi-1 (the point to be cut out and processed) is already connected to another branch before y-yi. It corresponds to the connected state.

“×” corresponds to the case where this is not the case.

``(Right x of the point) 5 (Right x of the branch) What is represented by ``>'' in J.

(i) The X coordinate value (α) of the rightmost edge of the branch on y-yi.

(ii) X coordinate value (β) of the rightmost point on y=yi-1
and.

Extract each of (iii).

β〉α It corresponds to the relationship between

Of course, what is represented by "〈".

βkuα It corresponds to the relationship between

``○'' in ``Has the branch already been terminated?'' corresponds to the case where the connection relationship of the branches has been determined, as indicated by hatching in the ``Before connection'' column in the first diagram. . “×” corresponds to the case where this is not the case.

[Effects of the Invention] As explained above, according to the present invention, two nodes and branches are extracted using graph theory, attribute information about these nodes and branches is checked, and a closed loop is determined. I have to. Therefore, even if noise is mixed in the original contour data of 9, it is possible to correctly determine the closed loop.

Note that when using the contour line extraction device according to the present invention, 2
For example, it becomes easy to draw a figure as shown in FIG. 8, which corresponds to a perspective view of a rectangular parallelepiped.

[Brief explanation of drawings]

FIGS. 1(A), 1(B), and 1(C) represent a principle configuration diagram and a processing concept explanatory diagram of the present invention. FIG. 2 shows an example of processing in the case of the present invention, FIG. 3 is a processing flowchart in the graph generation circuit, and FIG. Mode, Fig. 5 (A) (B
)(C), FIGS. 6(A), (B), and 7 are explanatory diagrams for explaining the prerequisite problems of the present invention, and FIG. 8 shows an example of drawing a rectangular parallelepiped. In the figure, 1 is a contour point data memory, 2 is a graph generation circuit,
Reference numeral 3 represents a graph storage memory, 4 a connection information memory, 5 a closed loop forming circuit, and 6 a contour data memory. $ 2 Figure
II21 Explanation of Figure 6 Figure 7 Example of stone power Figure O

Claims (1)

[Scope of Claims] In a contour extraction device that extracts contour data along the contour of an object from a group of contour point data obtained in the order of screen scanning by scanning a closed-loop binary image constituting the contour of an object, the contour point Contour point data memory (1
), a graph generation circuit (2) that sequentially extracts nodes and branches of a graph from the sequence of contour point data read out from the contour point data memory (1), and a graph generation circuit (2) that sequentially extracts nodes and branches of the graph.
) and a connection representing a connection relationship between the graph storage memory (3) that stores the graph information extracted by the graph generation circuit (2) and the contour point data obtained while extracting the graph information by the graph generation circuit (2). A connected information memory (4) for storing information, the contents of the graph storage memory (3), the contents of the contour point data memory (1), and the contents of the connected information memory (4) are used to create the object. It is characterized by comprising a closed loop forming circuit (5) for forming a closed loop constituting the contour of the contour, and a contour data memory (6) for storing the contour data obtained by the closed loop forming circuit (5). Contour extraction device.