JPS63316171A - System for extracting broken line and chain line - Google Patents

Abstract

PURPOSE:To attain an accurate extraction of element and to shorten the processing time by extracting a picture element composed of only arcs from a group whose direction of isolated picture elements is confirmed among groups whose direction of isolated picture elements are confirmed and not confirmed and taking the elements whose height is within a threshold level as broken lines/chain lines. CONSTITUTION:A picture data inputted from a picture input section 1 and whose floating vector is within a threshold level is grouped by a group extraction section 2. A direction deciding circuit 3 decides it as a group whose tilt is not clear even if any distance is more than the threshold level, when the tilt of a line segment K is decided to be the tilt of the group while distances l2-ln-1 of the center of rectangular vectors 91-94 are all within a threshold value with respect to the line segment K tying centers P1, P2 of grouped head and end rectangular vectors 7, 8. Broken line/chain line extraction circuits 4, 5 extract groups composed of arcs only and decide it respectively as chain/ broken lines and characters when the heights of the groups whose tilt is in existence and not clear are within and above the threshold level respectively.

Description

[Detailed Description of the Invention] [Summary] Broken lines and chain lines are often used in design drawings of machines, structures, etc. In many cases, this plays an important role in recognizing the structure of the object.

For example, the internal shape that cannot be seen from the surface may be displayed with a broken line, or the center line or the center position of a hole to be drilled may be displayed with a dashed line or a dashed double dotted line. On the other hand, when figures are automatically recognized, the process of identifying these broken lines and chain lines is quite difficult because the figure elements that constitute them are discontinuous. The present invention groups isolated vectors from image data related to figures under certain conditions, examines the shape of each vector in the group, finds candidates for broken lines or chain lines, and then This paper discloses a technology for extracting and recognizing broken lines and chain lines based on the connection relationships between candidates using information such as end points and directions.

[Industrial Application Fields] In recent years, the introduction of CAD systems in design work in various fields has been effective in reducing the number of man-hours required for the development of various products. Under such circumstances, the task of inputting design information into a computer is currently becoming the biggest problem in the operation of this CAD system.

The present invention is a drawing recognition system that aims to improve the efficiency of manually inputting graphic information into a CAD system by reading drawings drawn by designers as images, automatically extracting graphic information from the image data using pattern recognition technology. Regarding equipment. The present invention is a basic technology applied to the drawing recognition device, and relates to a method for extracting broken lines and chain lines drawn together with characters, symbols, and solid line segments in various design drawings, distinguishing them from other elements. The present technology can be used in a wide range of fields that handle design drawings with seven dashed lines, such as mechanical drawings and structural drawings.

[Prior Art] Most of the conventional broken line extraction processing methods are "sequential search methods." This is a method in which when a figure that looks like a broken line is found, it is assumed that the next broken line is an extension of that figure, and the area where the next broken line exists is searched. In this method, when starting broken line tracing, broken line elements are classified, one of the pre-classified elements is taken out, and other elements existing adjacent to both end points are searched. ,
This is done by connecting elements one after another.

The above-mentioned search for adjacent elements is performed by setting a square range touching the end points of the element, and selecting the element with the best continuity from other elements included within the range.

The size of this square range may be made variable as the tracing of the broken line progresses.

As a parameter for determining continuity with the next element obtained in the search area, for example, distance.

It uses width, angle, and intersection. Distance is the spacing between elements and is a central parameter of conventional methods.

Width is the vertical distance of an element, and is a parameter that indicates smoothness along with angle and intersection.

In such a conventional method, the continuity of a broken line element is determined by determining the angle between the element and other elements using a fixed threshold value.

Therefore, if the broken lines are mutually distorted or intersect in a complicated manner, such as hand-drawn broken lines, appropriate extraction processing cannot be performed, and once the search direction is wrong, there is no going back. There are so many lines that it is impossible to extract broken lines, or
Even if extraction is possible, there is a problem in that the performance is often unsatisfactory in terms of accuracy and IA processing time.

In an attempt to solve these conventional problems,
Japanese Patent Application No. 60-249322 proposes a technique for extracting broken lines and dashed lines by using the angles and distances of individual short line segments constituting the broken lines.

In other words, this technology uses a graphic processing device that is equipped with a function to approximate and record a line graphic image as a broken line, and includes a broken line candidate selection means that narrows down target data based on topological information, and an area search method that determines the optimal combination partner. The optimal candidates are combined by a means for dynamically setting dashed line conditions that takes into account image noise (line distortion, breakage, etc.), and means for verifying the linearity of the dashed line, etc. , the broken line is extracted from the line figure approximated by the broken line, and it is possible to shorten the processing time and improve the accuracy of extraction.

[Problems to be solved by the invention] Among the conventional techniques for extracting broken lines, etc. mentioned above, in the case of the latter,
Although it was superior in terms of processing time and accuracy,
Since the angles and distances of the individual short line segments that make up the broken line are used for extraction, if you are trying to absorb the handwritten distortion of the broken line or distortion during vectorization, or if you are trying to find the arc of the broken line, etc., it will be difficult to extract it. This posed the problem of having to increase the angle threshold, increasing the possibility of extracting other elements (letters, symbols) by mistake.

In view of these conventional problems, it is an object of the present invention to provide a figure recognition device that can stably extract broken lines/61 lines without being affected by handwritten distortion or vectorization distortion. .

[Means for Solving the Problems] According to the present invention, the above objects are achieved by the means described in the claims. That is, the present invention includes means for grouping rectangular vectors within a certain threshold into the same group, and means for detecting an open endpoint vector without curvature before an open endpoint vector without curvature. Generate a line segment connecting the centers of the last rectangular vectors, and if the distance of a perpendicular line from the center of each rectangular vector to the line segment is within the total threshold, use the slope of the line segment as the slope of the group, Orientation determining means for determining that the inclination is unknown because the arrangement of rectangular vectors is complicated in other cases; means for determining a group with a tilted trunk as a broken line/dashed line if the height of the group is within a certain threshold; This is a dashed line/dashed line extraction method that includes a means for combining the extracted lines divided by a line segment crossed by a line segment as the same dashed line/dashed line if each starting point and ending point are within a certain threshold and in the same direction.

[Embodiment] FIG. 1 is a block diagram showing an embodiment of the present invention.
Reference numeral 1 represents an image input section, 2 a grouping extraction circuit, 3 an orientation determination circuit, 4 a broken line/chain line determination circuit, 5 a broken line/chain line extraction circuit, and 6 a combination processing circuit.

FIG. 2 is a flowchart showing the processing of this embodiment.

The operation of this embodiment will be described below based on these drawings.

In FIG. 1, the grouping extraction circuit examines floating vectors (isolated collections of pixels) for image data input manually from the image input unit 1, and groups those vectors that are within a threshold value. become Then, as shown in FIG. 3, a line segment (K) connecting the center (Pl) of the first rectangular vector (vector of a rectangular area including one floating vector) 7 and the center (Pl) of the final rectangular vector 8 Set.

Furthermore, the distance between the line segment (K) and the center of each rectangular vector 91 to 94 is determined. (This distance is found by dropping a perpendicular line to the line segment K.) In Fig. 3, the distance between the center of the rectangular vector 91 to be found and the line segment (K) is ρ2, ) is shown as J2, etc.

If the distance i2 to 1n-1) between the line segment (K) and each rectangular vector is within a threshold value, the direction determination circuit shown in FIG. (K
) is determined to be the slope of that group.

On the other hand, if there is even one rectangular vector having a distance equal to or greater than the threshold value, the group is determined to have an unknown slope.

If it is determined that there is an inclination, the broken line/chain line determination circuit 4 in FIG. 1 selects one of the broken line/chain line candidates shown in FIG.
Select a group consisting of vectors that are all arc 1, open at both ends, and have a slope, as in the example shown in (a) ((b) shows an example of a vector whose slope is unknown). , if the height of the group is within the threshold value (β) as shown in FIG. 5(a) or (b), it is determined that the line is a broken line or chain line. Further, as shown in FIG. 2C, if the height is equal to or greater than the threshold value (β), it is determined that it is not a line but a character or the like.

If the direction determination circuit 3 in FIG. 1 determines that the inclination is unknown, the dashed line/dashed line extraction circuit 5 checks whether there is at least one arc 1, and if there is, If the open end point light has a vector of arc 1 having an open end point, it is indicated by a broken line/dashed line.

As shown in the example shown in FIG. 6, the combination processing circuit 6 determines whether the start point and end point of another broken line/chain line exist within the threshold and are the same. If the lines are in the same direction, they are treated as the same broken line or chain line and combined to form a single broken line or chain line.

[Effects of the Invention] As explained above, according to the present invention, for a collection of isolated pixels (floating vector), image data in which adjacent pixels within a threshold are grouped,
Those with an inclination (forming a row in a certain direction in jIL net),
For groups with an inclination, we take out those consisting only of arc 1, and the height of the group is
Those within the threshold are extracted as broken lines or chain lines. This has the advantage that even broken lines and chain lines that are distorted by handwriting or that are distorted during line thinning can be extracted accurately. -Also, broken lines and chain lines
Even if they are extracted separately by a line segment that crosses them, if their starting points and ending points are in the same direction within the threshold, they can be extracted accurately by connecting them as the same broken line or chain line.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing an embodiment of the present invention, Fig. 2 is a flowchart showing the processing of this embodiment, Fig. 3 is a diagram explaining the extraction of group orientation, and Fig. 4 is a diagram showing dashed/dashed lines. Figure 5 is a diagram showing an example of candidates, Figure 5 is a diagram showing a method for determining whether a broken line/chain line is determined, and Figure 6 is a diagram showing an example of a candidate.
The figure is a diagram illustrating a combination process. 1... Image human power section, 2... Grouping extraction circuit, 3
...Direction determination circuit, 4...Dotted line/chain line determination circuit, 5
. . . Broken line/dashed line extraction circuit, 6 . . . Combination processing circuit, 7
．． 8... Rectangular vector m/Tochi 1 Jiao's Ichin Seiki's Fang 1 shows 7th 4th Figure 7a plot/Seki showing how to run the chain plot 52 3rd meeting processing 11... Translation and recitation 7 Figure 6

Claims (1)

[Scope of Claims] Regarding input image data, the image data includes means for classifying rectangular vectors within a certain threshold value into the same group, and means for extracting an open end point vector without curvature before an open end point vector without curvature. A drawing recognition device that generates a line segment connecting the center of the first grouped rectangular vector and the center of the last rectangular vector, and calculates the distance between the center of each rectangular vector in the group and the line segment. If all the distances obtained by the above means are within the threshold, the slope of the line segment is determined to be the slope of the group, and if there is a rectangular vector with a distance exceeding the threshold, the slope of the group is determined. means for determining that the inclined group is unknown; means for determining that the inclined group is a broken line or a chain line when its width direction dimension is within a predetermined threshold; When the start point or end point of the extracted items divided by a crossing line segment is within a predetermined threshold and their directions are the same, they are combined as a dashed or chain line on the same line. A broken line/
Dashed line extraction method.