JPS61121183A - Discrimination for discontinuous segment graphic - Google Patents

Abstract

PURPOSE:To produce a histogram showing the features of each discontinuous segment pattern and discriminating automatically the type of each graphic, by applying the huff conversion to each of discontinuous segment graphics sorted into groups for each length of each element. CONSTITUTION:The coordinates on an X-Y plane of the start and end points ST and ED are detected for each element of discontinuous segment graphics alpha, beta and gamma. While the subject graphics are sorted into groups based on the length of each element. Then graphics alpha, beta and gamma are huff-converted onto a rho-theta plane, and the element of the largest length is extracted from each huff- converted graphic. Then each element which satisfies the prescribed conditions related to an angle is extracted again. Finally the elements excluded out of the huff-converted groups are neglected, and both start and end points of said graphics are calculated out of the points ST and ED of each element in each group. Then a histogram showing the number of features to the length of each element is produced for each group.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method for recognizing discontinuous line segment figures such as broken lines, dotted lines, and chain lines.

[Problems to be solved by conventional technology and invention]

Generally, a figure always includes not only continuous line segments but also discontinuous line segments.

For example, prefectural borders and contour lines are drawn in figures representing maps, but unlike other topographical features, discontinuous line segments such as dashed lines, dashed-dotted lines, and dashed-double-dotted lines are used for these.

Conventionally, when storing the above-mentioned discontinuous line segment figure as u@ in a database, the coordinates of each element are detected by digital type, and the type of the discontinuous line segment figure constituted by each element is input from the keyboard, for example, a broken line, I was entering information about.

However, there is a problem in that leaving the identification of discontinuous line segments to humans is extremely troublesome and causes a decrease in the efficiency of processing the graphics.

[Means for solving problems]

The present invention solves the above-mentioned problems and provides a method for automatically identifying discontinuous line segment figures. are classified based on the length dimension of each element, and each classified discontinuous line segment figure is determined by the distance ρ and angle θ from the origin of the XY plane to each element.
The Hough transform is performed on the ρ-θ plane defined by pρ and θ, and each reclassified discontinuous line segment figure is displayed as a histogram showing the number of pieces for the length dimension of each element. At the same time, the number of peaks in each of the histograms is detected, and the type of the discontinuous line segment figure that can be expressed on the X-Y plane is identified based on the number of peaks. Ru.

[Effect]

The method of the present invention performs a Hough transform to display the distance and angle from the origin to each discontinuous line segment figure on the side of each discontinuous line segment figure, which is grouped according to the length dimension of each element. Since a histogram showing the characteristics of a discontinuous line segment figure can be created, rice harvesting for each discontinuous line segment figure can be automatically performed from the feature histogram without human intervention.

〔Example〕

Hereinafter, the present invention will be explained by way of examples with reference to the attached drawings. @ Fig. 1 (A) is an example of a figure to which the method of the present invention is applied, α is a broken line, β is a dashed-dotted line, r indicates a chain double-dashed line.

First, the discontinuous line segment figure α shown in FIG. 1 (N).

Measurements are taken on the X-Y plane between the start point ST, which is the first point of each element of β and r, and the end point ED, which is the last point.

In the above formula, Ll is the length dimension of the first element, "
As is clear from the subscripts, s i s"@l and 7g l 'Fe L indicate the X and y coordinates in ST and ED, respectively.

Next, the target figure (first
Figure (A)) is classified into groups. An example of this is shown in Table 1.

In Table 1 above, group 1. It is roughly divided into three types: I [and ■]. The table for each group displays the coordinates of BT and ED (FIG. 2) and length t for each element number (FIG. 1(A)).

Group (2) has each dimension 11, 12, .

However, groups ■ and ■ are each further divided into groups a and b, and two types of length dimension ranges (ta-Lb,
Lb~1. )It is in.

However, by grouping only ■ and ■, it is only revealed that they are chain lines, and it is not possible to identify whether ■ and ■ are chain lines or chain double-dot lines.

Therefore, the figures α and β on the X-Y plane of the prisoner in Figure 1.

Transform γ onto the ρ-θ plane (Figure 1 (B))
.

Here, the −·fu conversion is as shown in FIG.

This means that a certain graphic element e expressed on the X-Y plane can be expressed on the ρ-θ plane defined by the distance ρ from the origin 0 and the angle θ.

That is, the length of the extension line of element e having ST and ED as end points and the perpendicular line OH from the origin that intersects with it is defined as the distance ρ of element e.

Further, the angle formed between the X axis and the above ρ is defined as the angle θ of the element e. In addition, the relationship between X, 7 and ρ, θ is ρ: X Co
Sθ+y-route θ.

The result of Hough transformation on the ρ-θ plane defined by ρ and θ defined in this way is shown in FIG. 1(B). The dashed circles α, β, and r are α in Figure 1 (b), respectively.
, β+rK are supported.

Originally, as is clear from Figure 1, mathematically, one type of broken line α should have the same values of ρ and θ for each element 1, 2, 5, 11, etc. .

However, in the actual broken line α, as shown in FIG. 3, each element 1, 2, 5, 11, . . . has a different direction.

Therefore, for element 2, the distance is ρ2. The angle is θ2, but for element 5, the distance is ρ5. The angle is θ5, and the values of ρ and θ differ depending on the element even within the same broken line α.

Therefore, as shown in FIG. 1(B), the result of the hub transformation is each discontinuous line segment figure α represented by ρ and θ
, β, and r are not a single point but a set of points surrounded by a circle.

In this way, all elements of the maximum length dimension are extracted from each of the hub-transformed figures α, β, and r.

For example, in Group I (Table 1), if M is the element with the maximum length dimension 'max, then ρ=ρ
, θ=θ is detected.

rn m Then, each element satisfying 1ρ-ρ1≦Δρ and 10m-θ11≦Δθ is re-extracted and newly divided into each group as shown in Table 1.

This operation is performed with the purpose of excluding elements that deviate from the above Δρ and Δθ from the viewpoint that the element force of the maximum length dimension in each group is the most reliable element in each group.

Finally, ignoring the elements excluded from the Hough-transformed groups, from ST and FD of each element in each group, start point Sa and end point of α, start point Sβ and end point of lβ Eβ.

The start point S and end point E of γ are calculated respectively.

Next, a histogram representing the number of elements relative to the length dimension is created for each group (FIG. 4).

In the above histogram, when the only peak exceeding a certain MiTH4 is el, the broken line α (Fig. 4 (A))
, if the peaks are the same el and 62, the single-point chain helix β (fourth
Figure (B)), there are two peaks, e and 2, @+>6
In case 2, the two-dot chain line (FIG. 4(C)) is shown.

A dashed-dotted line consists of two types of segments, short and long, but the number of both is almost equal, while a dashed-double line has more short segments than long segments, so the peak e1 62
This causes a difference in the distribution of (Figure 4 (B), (
C) ).

In this way, the NM of each discontinuous line segment figure α, β, γ is identified and expressed as a folding vector, as shown in FIG.

In Figure 5, α is S (the broken line from X to E4 (the fourth
Figure (A)), a dashed-dotted line from β beam to E (Figure 4 (
B)), r is the chain double-dashed line from S to E (Figure 4 (C
)).

〔Effect of the invention〕

According to the present invention, the distance and angle from the origin to each discontinuous line segment figure are displayed for each discontinuous reduced figure grouped according to the length dimension of each element. By applying this method, a histogram showing the characteristics of each discontinuous line segment figure can be created, so that the type of each discontinuous reduced figure can be automatically identified from the feature histogram without human intervention.

[Brief explanation of drawings]

FIGS. 1A and 1B are diagrams showing the features of the system of the present invention, and FIG. 2 is an explanatory diagram of the Hough transform used in the system of the present invention. Figure 3 is a relationship diagram between discontinuous line figures and ρ and θ, Figure 4 (C) is a histogram corresponding to each discontinuous line figure,
FIG. 5 is a diagram showing the identification results using the above-mentioned histogram. 1.2.3... Each element of a discontinuous line segment figure, α... broken line, β... one-dot chain line, r... two-dot chain line, el +
82...Histodalam peak.

Claims (1)

[Claims] Discontinuous line segment figures that can be expressed on the X-Y plane are classified based on the length dimension of each element, and each classified discontinuous line segment figure is calculated by the distance ρ from the origin of the X-Y plane to each element. and angle θ
By performing Hough transform on the ρ-θ plane defined by and detecting the number of peaks in each of the histograms, and identifying the type of the discontinuous line segment figure represented on the XY plane based on the number of peaks.