JPH0532787B2 - - Google Patents

Description

[Detailed Description of the Invention] [Summary] For example, drawings of machines and structures drawn based on certain conventions such as the first angle projection method or the third angle projection method were originally drawn by humans. Since it is made to be used to judge the shape and structure of the object by looking at it, other side views (in this specification, for example, front view, Drawings that are drawn independently in different projection directions of the same object, such as side views and top views, are called surface views. In cases where the drawing becomes complicated and the structure of the object is difficult to understand, dimensions that should normally be shown on the drawing are often omitted. As described above, there has been a problem in that drawings with omitted parts cannot be reliably dealt with using conventional drawing recognition techniques.

In order to solve these conventional problems, the present invention uses image data related to input plan views to determine the dimensions of other plans based on the known dimensions of one plan view based on the correspondence between the plan views. This paper discloses a shape recognition technique for drawings that has a high information density and is rich in novelty by determining the dimensions of the shape lines in the drawings and further calculating dimensions that are not shown in the drawings.

[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, currently,
In terms of operation of this CAD system, inputting design information into a computer has become the biggest problem.

The present invention is a drawing recognition system that aims to improve the efficiency of inputting design information into a CAD system by reading a drawing drawn by a designer as an image and automatically extracting design information from the image data using pattern recognition technology. Regarding equipment. Particularly related to automatic drawing input devices for mechanical drawings, structural drawings, etc. drawn using the projection method, from the drawing of design drawings to inputting design information into a computer and passing it on to the actual manufacturing process using a CAD system. It is applied to a consistent system up to the creation of processing information (NC tape). The present technology can be used in a wide range of fields, such as mechanical drawings and structures, in which design drawings are drawn using projection methods.

[Background Art] Machine drawings and drawings of structures are generally drawn using projection methods such as the first angle method and the third angle method.

FIG. 8 is a diagram illustrating the relationship between each side view when drawing by the third angle method, and 51 is a front view;
2 is a left side view, 53 is a right side view, 54 is a plan view, 5
5 bottom view, 56 rear view, 57 X axis, 58 Y
The axis 59 represents the Z axis.

In other words, in the third angle method, the shape diagram in each direction by projection of the object is, for example, a front view 5, as shown in the same figure.
On the right side of 1, a right side view 53 of the object is drawn, and above the front view 51, a plan view 54 is drawn of the object as seen from above. (In the first angle method, contrary to the third angle method, the left side view is drawn on the right side of the front view, and the bottom view is drawn on the upper side of the front view.) A described in the front view 51 , B or C shown in the plan view 54 represent dimensions.

That is, this figure shows a cube whose width is A, height is B, and depth is C when viewed from the front.

[Problems to be Solved by the Invention] In the drawings shown in FIG. 8 mentioned above, the description of the shape and dimensions is as follows:
, the height dimension B, and the depth dimension C in the plan view.
However, as long as humans can see this, they can easily determine the dimensions of each part from the correspondence of each side view.
It is easy to imagine its shape.

However, when looking at this in units of plan views, there are some cases in which the dimensions cannot be immediately determined.

Conventional shape recognition devices can only recognize drawings drawn using projection methods such as the first angle method and the third trigonometric method, which are used to draw mechanical drawings and structural drawings. Drawings larger than a front view are processed in the same way as a one-dimensional view, and recognition that takes projection methods into consideration is not performed. For this reason, many mechanical drawings and structural drawings are very often drawn as three-view diagrams using third trigonometry, and most design drawings cannot be handled by conventional drawing recognition devices. Additionally, because the projection method of drawing is not taken into account during recognition processing, only a portion of the design information is obtained, which poses the problem of placing too much burden on subsequent CAD system work.

In view of these conventional problems, the present invention has been developed in a drawing recognition device that develops the shape of an object into a three-sided view, such as a mechanical drawing or a structural drawing. Understand drawing methods using projection methods such as first angle method and trigonometric method as determined notations, and determine the shape by setting dimensions while taking correspondence in each drawing for shapes that are not directly specified. The purpose of the present invention is to provide a shape recognition device that can recognize shapes.

[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 is a device that recognizes a drawing from image data related to a drawing created by developing the shape of a machine or structure using a projection method, which classifies line segments and characters in the image data, A means for recognizing characters, identifying the type of line segment and determining whether it is a shape line or an auxiliary line, and a means for determining the correspondence between the recognized character information and the shape line or the auxiliary line. , a means for dividing the shape line into each surface view to find the position on the drawing of the surface view and its apex, and determining the length of the shape line corresponding to the part for which dimension values are specified for the surface In addition to calculating the length of the shape line of the part for which no dimension value is specified from the known dimension value, and by using the correspondence between each surface drawing by the projection method, This is a shape recognition device that assigns a length value to a shape line using a known dimension value in a drawing or a value calculated based on the dimension value.

Hereinafter, the present invention will be described in detail based on examples.

[Embodiment] FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, in which 1 indicates a line segment/character/symbol classification section, 2
3 is a character recognition unit, 3 is a line segment identification unit, 4 is a correspondence between characters and shape lines/auxiliary lines, 5 is a shape line/auxiliary line identification unit, 6 is a plan view position recognition unit, 7 is a vertex extraction unit , 8
9 is a dimension setting unit, 9 is a dimension setting calculation unit, 10 is a dimension setting unit based on the correspondence of three orthographic views, 11 is a knowledge base, 12 is an interaction processing unit, 13 is a symmetry processing unit,
It represents. In the figure, a portion surrounded by a dotted line indicates a processing section corresponding to the technology of the present invention.

These data received by the processing unit of the present invention include character strings (instruction dimension values) obtained by the processing of the character and shape line/auxiliary line correspondence unit 4, data of the corresponding shape lines/auxiliary lines, and shape line/auxiliary line data. The line recognition unit 5 and the plan position recognition unit 6 collect data grouped into net units with connected shape lines and the plan position of each group (front view, plan view, etc.).
This is the data.

First, as shown in FIG. 2, the vertex extraction unit 7 extracts vertex positions by projecting the end point coordinates of the shape line onto the X and Y axes for each plan view. In this case, if the end point coordinates of each vertex are close, it is normalized to one vertex.

Next, the dimension setting section 8 first creates a two-dimensional array table as shown in FIG. 3 having the numbers of vertex positions on the X and Y axes. And characters and shape lines/
From the shape line/auxiliary line that corresponds to the dimension value (distance) obtained by the processing of the auxiliary line correspondence unit 4, the dimension value is set in the corresponding vertex section. here,
Only the distance of the vertex section with dimension line instructions is set. The distance setting of this vertex section is determined from values specified by text information, such as horizontal dimensions, vertical dimensions, diagonal dimensions as shown in FIG. 4, and angular dimensions as shown in FIG. 5. In the example of FIG. 5, the scale ratio is considered. Other horizontal/vertical dimensions, circles, arc dimensions, etc. are similarly calculated to determine the distance between the vertex sections and set in a two-dimensional table as shown in FIG. 3 mentioned above. Next, in the process of the undetermined dimension setting calculation unit 9,
The two tables Y are received, and undetermined dimension values are calculated and set from the set dimension values using the following equation.

l _(n, 〓 ₎ =l _(n,n+1) +l _(n+1, 〓 ₎ −(1) where α=m+2,m+3,,,N−1 1≦m＜N−2 l ₍ 〓 _{, o)} =l _(o-1,o) +l ₍ 〓 _,o-1) -(2) However, β=n-2, n-3,,,1 3<n<N Figure 6 a to h are book This is a diagram showing the processing flow of the embodiment. Through the processing up to this point, a table is obtained in which dimension values determined within the range for each side view are set, as shown in FIG. 6c. That is, for example, dimension values are set on the shape lines indicated by the O marks in FIG. However, the dimensions of the shape line marked with an x shown in FIG. 7 cannot yet be set here.

The undetermined dimension calculation unit 9 in FIG. 1 uses the knowledge base 11 from the dimension setting table obtained for each surface
While referring to the correspondence of the surface views stored in , the vertices of the corresponding surface views are normalized as shown in Figure 6 d, and the X,
Create a table for setting three dimensions for the Y and Z axes. Then, the same processing as that of the undetermined dimension setting calculation unit 9 is performed on the three tables (Tx, Ty, Tz). If there are still undetermined dimensions in the process up to this point, please
The symmetry processing unit shown in the figure eliminates undefined dimensions by performing symmetry processing, an example of which is shown in FIG. 6g. In addition, there are cases where the distance between shape line segments cannot be set due to a mistake in specifying dimension values in the first drawn drawing, so if undetermined dimensions remain, an inquiry is made to the dialog processing unit 12 to specify accurate dimension values. to determine the shape.

[Effects of the Invention] As explained above, according to the present invention, when recognizing design drawings written according to the projection method, such as mechanical drawings and structural drawings, it is possible to write dimensional value instructions on all surface drawings. By recognizing the design drawings that do not exist, it becomes possible to determine the length and size of the shape line of the surface drawings where dimension values are not written, by establishing the correspondence between each surface drawing using the technology of the present invention. Drawing information that could not be obtained with conventional recognition devices can be obtained. This makes it possible to extract most of the drawing information written in the design drawings, which has the advantage of greatly reducing the work of adding missing information in subsequent CAD systems.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, FIG. 2 is a diagram explaining the process of finding vertices,
Figure 3 is a diagram showing a table of information related to vertices, Figure 4 is a diagram explaining how to calculate the dimensions between vertices from specified values, and Figure 5 is the distance between vertices given an angle. Figure 6 explaining the case of finding
The figure explains the process flow, Figure 7 explains whether dimensions can be set for each side view, and Figure 8 explains the relationship between each side view when drawing using the third angle method. It is a diagram. 1... Line segment/character/symbol classification section, 2... Character recognition section, 3... Line segment identification section, 4... Character and shape line/auxiliary line correspondence section, 5... Shape line/auxiliary line identification section, 6... Surface Diagram position recognition unit, 7... Vertex extraction unit, 8... Dimension setting unit, 9... Undefined dimension setting calculation unit, 10... Dimension setting unit based on three-view diagram correspondence, 11... Knowledge base, 1
2... Dialogue processing section, 13... Symmetry processing section.

Claims (1)

[Claims] 1. A device that recognizes a drawing from image data of a drawing created by developing the shape of a machine or structure using a projection method, which classifies line segments and characters in the image data. A means for recognizing characters, identifying the type of line segment and determining whether it is a shape line or an auxiliary line, and a correspondence relationship between the recognized character information and the shape line or the auxiliary line. means for determining the position on the drawing of the surface view by dividing the shape line into each surface view, and a means for determining the apex thereof; Using a method that assigns a length value and calculates the length of the shape line of the part for which no dimension value is specified from known dimension values, and the correspondence between each plan view based on the projection method,
A shape recognition device comprising means for assigning a length value to a shape line using a dimension value known in another view or a value calculated based on the dimension value.