JPS63104189A - Drawing read system - Google Patents

Abstract

PURPOSE:To speed up processing by expressing a figure that an operator wants to specify by basic graphic elements such as straight lines, circles, triangles and rectangles and recognizing the figure. CONSTITUTION:A binarization means 12 binarizes the figure read out of a drawing 11 into binary picture data. A binary picture memory means 13 stores the data and hands it to a basic graphic element recognition means 14. First, it converts the binary picture data into short vector data, obtains vector data from the former vector data and recognizes basic graphic elements. The elements that are recognized as straight lines, circles, arcs, triangles and rectangles, and parameters such as coordinates and sizes are stored in a drawing basic graphic element group storage means 15. A specific figure recognition means 14 compares a set of the basic graphic elements that the drawing basic graphic element group storage means 15 reads out of the figure with a set of basic graphic elements of a specific figure in a dictionary, which is outputted from a dictionary storage means 16, calculates a degree of similarity and the difference, evaluates the difference, and recognizes the specific figure.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a drawing reading method for detecting a predetermined figure in a drawing and recognizing its shape.

[Conventional technology]

Figure 7 shows, for example, Information-technology Promotion Association Technology Center Report 1F.
This is a configuration diagram of a decision tree showing the conventional drawing reading method shown in 59 technique-043. In the figure, 1 is a feature extraction box that extracts the feature from binary image data, and 2 is a These are decision boxes for classifying figures, and have a hierarchical structure, and the feature quantity from the feature extraction box 1 is received by the decision box 2 at the top level. 3
is a munching box that performs pattern matching by holding one group of classified specific graphic symbols, and is provided corresponding to each branch of the judgment box 2 in the lowest hierarchy.

Next, the operation will be explained. A predetermined figure read from the drawing is sent to the feature extraction box 1 as binary image data. The feature extraction box 1 thins the received binary image data to create vector data, then extracts feature quantities such as the number of closed curves and the number of open end points, and passes this to the determination box 2 at the top layer. In the judgment box 2, the predetermined figure is sequentially branched for each layer based on its characteristic amount, and the matching box 3 is selected from the judgment box 2 in the lowest layer.
Pass it to 01.

The matching box 3 to which the predetermined figure has been passed in this manner performs pattern matching between this and the symbols of the group it holds, thereby specifying and recognizing the symbol of the predetermined figure.

[Problem that the invention seeks to solve]

Conventional drawing reading methods are configured as described above, so when you want to add a specific figure type that you want to recognize, you need to know the type and number of feature quantity determinations, as well as the number of symbol groups for pattern matching. must be increased,
There were problems such as the need to reconsider the specific method each time, and the recognition target area had to be extracted one by one.

This invention was made in order to solve the above-mentioned problems, and it provides a drawing reading method that does not require reconsideration of specific means when adding a specific figure type and speeds up the recognition of specific figures. The purpose is to obtain.

[Means for solving problems]

The drawing reading method according to the present invention has a dictionary that stores a plurality of specific figures to be recognized as combinations of basic figure elements, reads the drawing to be read as binary image data,
The figure read from the drawing from the binary image data is recognized as a basic figure element, and then the same type of basic element as the element with the largest size among the basic figure elements of the figure in the drawing is selected as the basic figure of the specific figure in the dictionary. The largest element among the graphical elements is selected as a candidate, and each of these candidates is compared with a set of graphical elements to evaluate the mutual placement positions of basic graphical elements as errors.

[For production]

The drawing reading method in this invention is one straight line and one circular arc.

You can create a dictionary by collecting specific shapes that you want to recognize using basic graphical elements such as circles, triangles, and squares, and also compare the specified shapes read from the drawing with the specific shapes in the dictionary to determine the placement position among the basic graphical elements. A specific figure is recognized by evaluating the error in the figure, and when adding a figure type to be specified, it is sufficient to simply express the figure with basic figure elements, and it is not necessary to reconsider specific means. I don't.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings. 1st
In the figure, 11 is the drawing you want to read, 12 is this drawing 1
Binarization means 13 creates binary image data of the read figure; 13 is binary image storage means for storing the binary image data from the binarization means 12; 14 is a straight line from the binary image data; 2 arcs.

Basic geometric elements such as circles, triangles, and squares and their sizes.

Basic figure element recognition means for obtaining parameters such as position and direction; 15, drawing basic figure element group storage means for storing the basic figure elements and their parameters from the basic figure element recognition means 14; 16, when the specific figure to be recognized is a straight line 2; Dictionary storage means 17 stores a dictionary composed of basic graphical elements such as arcs, circles, triangles, squares, etc. A dictionary storage means 17 stores a set of basic graphical elements of drawing figures from the drawing basic graphical element group storage means 15. Specific figure recognition means 18 recognizes a specific figure by comparing it with a set of basic figure elements of a specific figure in the dictionary of the dictionary storage means 16 and evaluating the error between mutual basic figure elements; 18 is a specific figure recognition means 17; This is a recognition result storage means for storing recognition results.

Next, the operation will be explained. Here, FIG. 2 reads n
A plan view showing an example of drawing 11, and FIGS. 3 and 4 are explanatory diagrams showing the recognition process of its basic graphical elements.

Figures read from drawing 11 (for example, broken line 21 in Fig. 2)
The portion surrounded by ) is converted into binary image data by the binarization means 12. FIG. 3A shows this binary image data, and each black point 22 is given an X coordinate and an X coordinate as a position. This binary image data is stored in the binary image storage means 13.
and is passed to the basic graphic element recognition means 14.

The basic graphic element recognition means 14 first converts the binary image data into short vector data, and then obtains vector data from the short vector data to recognize basic graphic elements. FIG. 3B shows this short vector data, and each short vector 23 has coordinates of a starting point and an ending point, and is obtained from a minute portion of a mass of black points 22 of the binary image data.

Moreover, FIG. 3C shows the connection relationship of each of these short vectors 23,
Vector data obtained from vector angle changes, etc.
Each side of the figure 21 read from the drawing 11 is recognized as a long vector 24 having the coordinates of a starting point and an ending point.

Here, if the figure read from drawing 11 is a circle, it is recognized as a circle 26 defined by the center coordinates and radius shown in FIG. 4B from the short vector data 25 as shown in FIG. 4. Straight line 9 circles 2 arcs recognized in this way,
Basic graphical elements such as triangles and squares and parameters such as their coordinates and sizes are stored in the drawing basic graphical element group storage means 15.

The specific figure recognition means 17 compares the set of basic figure elements of the figure read from this drawing basic figure element group storage means 15 with the set of basic figure elements of the specific figure in the dictionary storage means 16, The degree of similarity is calculated as an error, and this error is evaluated to recognize a specific figure.

FIG. 5 is an explanatory diagram showing this recognition process, and FIG. 6 is a flowchart showing the flow of the process. FIG. 5A shows a figure read from a portion surrounded by a broken line 21 in FIG. 11, which is converted into basic figure elements into six line segments D1 to D60 on the right side.

In step ST1, the largest element among the basic graphical elements in the dictionary, for example, zl, is selected, and in step ST2, the largest element among the basic graphical elements in the drawing (for example, D
I) is selected, and dictionary constituent elements having an element similar to this element as the largest element are restored as candidates. The left side of Figure 5 is one of these candidates, 21~z
It is converted into basic graphical elements into six line segments of 6.

ST4 to ST3 are error evaluation procedures for one candidate. The two pieces selected in step ST4 are associated with each other, and as shown in FIG. 5B, the basic figure element D
1 is moved to the position of the basic figure element z1, and in step ST5, the remaining basic figure elements D2 to D6 are calculated as shown in FIG. 5C. Move according to the following.

Next, through a loop of steps ST7 and STg, the basic graphical elements 22 to z5 of the dictionary are made to correspond to the basic graphical elements D2 to D5 of the drawing, and those with the minimum error are made into pairs, and the errors are cleared in step ST6. The total error is sequentially added to the calculated error area.

Here, if the basic graphic element is a line segment, this error is calculated by defining it as the square of the distance between the end point of the line segment on the dictionary side and the end point of the line segment on the drawing side, for example, In the case of a circle, the square of the distance between the center positions and the square of the radius difference; in the case of a triangle, the sum of the squares of the distances between each of the three vertices; in the case of a square, the sum of the squares of the distances between each of the four vertices It is calculated by defining it as the sum of squares of the distances.

Next, the total error calculated in this way is compared with the previously calculated total error in step ST9, and only if it is small, the value is updated and the process returns to step ST3.

Here, in the storage area for this total error, as an initial value,
For example, the maximum value that can be stored is set. In step ST3, where the process is returned, one of the remaining candidates is selected from among the restored candidates, and the same process is repeated thereafter.

In step ST10, the final remaining minimum total error is compared with a tolerance value. As a result, if Itjn is smaller than the tolerance value, it is determined in step ST11 that the combination of basic graphic elements that obtained the minimum total error is the specific shape to be recognized, and post-processing is performed; In step 5T12, it is determined that there is no specific figure to be recognized, and post-processing is performed.

In the above embodiment, a case has been described in which the error evaluation is performed in accordance with the coordinate system on the dictionary side, but the error evaluation may be performed in accordance with the coordinate system on the drawing side, and the same effect as in the above embodiment is achieved.

〔Effect of the invention〕

As described above, according to the present invention, since the figure to be specified is represented by basic figure elements such as two straight circles, triangles, and squares, and the specific figure is recognized, when the figure type to be specified is added. However, there is no need to reconsider the specific means for recognition, and furthermore, it is possible to use a drawing reading method that speeds up the recognition of specific shapes because the candidates are narrowed down to the maximum number of basic graphic elements. It has the effect of

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing a drawing reading method according to an embodiment of the present invention, FIG. 2 is a plan view showing an example of a drawing to be input,
FIGS. 3 and 4 are explanatory diagrams showing the recognition process of basic graphic elements in this invention, FIG. 5 is an explanatory diagram showing the recognition process of specific shapes in this invention, and FIG. 6 is a flowchart showing the flow of the process. , FIG. 7 is a configuration diagram of a judgment water showing a conventional drawing reading method. 11 is a drawing, 12 is a binarization means, 13 is a binary image storage means, 14 is a basic figure element recognition means, 15 is a drawing basic figure element group storage means, 16 is a dictionary storage means, 17 is a specific figure recognition means, D1 to D6 and z1 to z6 are basic graphical elements. In addition, in the figures, the same reference numerals indicate the same or equivalent parts. Patent applicant Mitsubishi Electric Co., Ltd. Agent Patent attorney EB and tvt [J”'
”1 (2 others) Figure 6

Claims (1)

[Claims] In a drawing reading method for detecting a predetermined figure in a drawing and recognizing its shape, a binary image storage means for storing binary image data of a figure read from the drawing;
a drawing basic figure element group storage means for storing basic figure elements obtained from value image data and parameters such as the size, position, direction, etc. of the basic figure elements; The dictionary storage means stores a plurality of dictionaries in which basic figure elements are registered as dictionary constituent elements, and the basic figure is of the same type as the largest basic figure element among the basic figure elements of the figure read from the drawing. Select the dictionary component that has Element as the largest basic figure element as a candidate for the specific figure you want to recognize, and for each candidate, select the combination of the basic figure elements of the figure read from the drawing among the basic figure elements of the specific figure in the dictionary. A drawing reading method characterized in that the errors between these two basic figure elements are calculated and summed in descending order of magnitude, and the candidate with the smallest sum is determined to be a specific figure to be recognized.