JP2920195B2 - Shape conversion method and device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a method and an apparatus for transforming a shape by a computer used in the field of CAE and the like, and in particular, to modeling and mesh division which require classification and identification of shapes. More particularly, the present invention relates to a shape conversion method and a device suitable for the present invention.

[Conventional technology]

In the CAE, the boundary shape is represented by structural data that is connected to a plane-line-point, and several pieces of data such as the equation of a plane and a line and the coordinates of a point.

Conventionally, as a shape recognition method based on these data, the most commonly used methods are the quadtree method (two-dimensional) and the octree method (three-dimensional). Improved Quadtree Method for "Mark A. Ely and Mark E. Shepherd IEEE CG & A January 1983 P
39-46 (A Modified Quadtree Approach to Finite Ele
ment Mesh Generation Mark A. Yerry and Mark S. Sheph
ard IEEE CG & A January 1983).

Here, for the sake of simplicity, FIG.
This method will be described by taking a dimensional shape as an example.

First, this figure is placed in a reference square (reference square), and an integer coordinate system is set such that the length of one side of the square is 2 ⁿ . Then, the reference square is divided into four equal small squares, and the relationship between each area and the boundary line forming the figure is classified as follows.

(1) The square is inside the boundary (2) The square is outside the boundary (3) The square contains the boundary Here, only the square determined as (3) is again divided into four equal parts, Perform the above checks. Such an operation is continued to a level corresponding to an appropriate shape resolution.

The result of division until one side of the square becomes 1/8 of the reference square is shown in FIG. 17B, and the corresponding tree structure is shown in FIG. 17C. When the resolution of one rank is increased, the result becomes as shown in FIG. 17D.

Then, in shape recognition, the approximate characteristics of the shape are determined based on the tree structure in FIG. 17C.

The numbers entered in FIG. 17C indicate the relationship between the divided square area and the boundary line forming the target graphic, and correspond to the above three classifications.

[Problems to be solved by the invention]

In the above prior art, it is difficult to grasp the overall graphic characteristics of the shape itself, and there is a disadvantage that the recognition rate progressively decreases when the shape is distorted from a registered graphic (a graphic to be compared). Also, to create an approximate model,
No unified theory has been established regarding which coordinate axis is parallel to each side of the target shape, and it contains some ambiguity.

An object of the present invention is to provide a shape conversion device that generates an approximate model in which graphic characteristics of an arbitrary shape are stored. Another object of the present invention is to input an arbitrary shape,
An object of the present invention is to provide a shape conversion device that outputs an arbitrary shape divided into finite elements.

[Means for solving the problem]

The above object is achieved by providing a shape conversion device having means for converting an arbitrary shape into a shape consisting only of straight line segments, and having means for converting the straight line segments into line segments parallel to the coordinate axes.

The shape conversion device according to claim 1, wherein the means for converting the straight line segment into a line segment parallel to the coordinate axis includes an ambiguous operation unit for performing an operation using a membership function.

In addition, the above problem is that a boundary line or an ridge line of an arbitrary shape is approximated by a plurality of straight line segments, and the degree of approximation to the coordinate axis of each line segment is expressed by a variable from 0 to 1, and the variable is represented by an ambiguous rule. This is also achieved by a shape conversion method in which each line segment is corrected in parallel and assigned to any of the coordinate axes, and finally converges to one approximate model.

In addition, a shape conversion that approximates a boundary line or ridge line of an arbitrary shape with a plurality of straight line segments and converts the line segments so that adjacent line segments are on a straight line or perpendicular to each other to form a shape is performed. It is good also as a method.

Further, a boundary line or ridge line of an arbitrary shape is approximated by a plurality of straight line segments, an angle between each line segment and a coordinate axis is calculated, and the calculated angle and a membership function based on a preset ambiguous rule are used. A shape conversion method for creating an approximate model of the arbitrary shape by allocating a line segment in parallel to one of the coordinate axes may be used.

Further, the fuzzy rule is that the line segment is assigned at least in the direction of the coordinate axis where the angle formed by the line segment is the smallest, and that the angle between two adjacent line segments is smaller than a certain angle. The shape conversion method according to claim 5, further comprising: assigning an angle to a different direction in the same direction as much as the angle formed is larger than a certain angle.

In addition, after independently creating an approximate model of a boundary surface or a boundary line of a given shape and an approximate model of a hole included in the boundary surface or the boundary line, regardless of the hole in the given actual shape, A shape conversion method that forms a smooth grid and determines the relative position of the hole to the boundary shape in the approximate model based on which grid point the vertex that forms the hole in the real shape corresponds to Is also good.

Furthermore, the shape conversion method may be characterized in that the length of a line segment forming the approximate model is converted to the minimum integral multiple of the unit length while maintaining the topological characteristics of the shape of the approximate model. .

Further, an arbitrary shape to be analyzed is input, and an approximate model of an arbitrary shape composed of only straight line segments parallel to the coordinate axes of the orthogonal coordinate system is generated from the arbitrary shape, and each of the recent models is generated from the approximate model. A mapping model having a grid shape in which a straight line segment is gridded so as to be an integral multiple of the unit length is generated, and an analysis mesh of an arbitrary shape is automatically generated from the mapping model by performing a mapping operation. It may be.

[Action]

The arbitrary shape is converted into a shape consisting of only straight line segments,
Further, since the straight line segment is converted to be parallel to any one of the coordinate axes, the arbitrary shape is converted to a figure composed of only straight lines parallel to the coordinate axes.

An ambiguous operation unit that performs an operation using the membership function selects which coordinate axis each line segment should be transformed in parallel to in accordance with a predetermined ambiguous rule.

In creating an approximate model of an arbitrary shape, first, all ridge lines (in the case of a three-dimensional shape) or boundary lines (in the case of a two-dimensional shape) of the arbitrary shape are approximated by straight line segments, and each line segment and a coordinate axis (x, An angle formed between the y-axis and the y-axis is calculated. Next, at the time of assigning which coordinate axis is parallel to each line segment, at least the following two basic vague rules are used.

Each line segment is assigned in the direction of a coordinate axis with the smallest angle that can be made.

Two adjacent sides are assigned to different coordinate axis directions as the angle between them becomes smaller than the fixed angle, and are assigned to the same direction as the angle between the adjacent sides becomes larger than the fixed angle.

Then, a membership function in fuzzy theory is used to express the ambiguity of this rule. First, the degree of approximation of each line segment to the coordinate axis is represented by 0 to 1. In this case, the degree of approximation approaches 1 as the angle formed with the coordinate axis approaches 0 degrees, and the degree of approximation approaches 0 as the angle formed approaches 90 degrees.

Further, with respect to two adjacent sides, the degree of the same direction is represented by -1 to 1 based on the angle formed by the two sides. In this case, the degree of the same direction approaches 1 as the angle between the two sides approaches 180 degrees, changes to -1 by 90 degrees, and becomes constant at -1 below 90 degrees.

Next, for two adjacent sides, the degree of influence on each side is calculated based on the degree of approximation to the coordinate axis and the degree of the same direction. In this case, when the degree of the same direction is positive, for example, the degree of approximation of the affected side in the X direction acts to increase the degree of approximation of the affected side in the X direction and to decrease the degree of approximation of the Y and Z directions. When the degree of the same direction is negative, for example, the degree of approximation of the affected side in the Y direction lowers the degree of approximation of the affected side in the Y direction and increases the degree of approximation of the X and Y directions. Then, such an influence degree is represented by a quantity, and the X, Y, Z
The degree of approximation to each coordinate axis is modified.

When the correction of all sides is completed, the degree of influence is similarly calculated for two adjacent sides based on the corrected degree of approximation, and thereby the degree of approximation to the coordinate axes is corrected again. Such an operation is repeatedly performed, and for all sides,
If the degree of approximation in one direction among the degrees of approximation to each coordinate axis is sufficiently close to 1, the state is set to a convergence state and the direction assignment of each side is determined.

If the direction assignment of each side is determined, the length of each side on the approximate model is determined for each loop (graphic). At this time, the following method is used as a basic method for determining the line length. As shown in FIG. 15, if the two-dimensional figure is composed of only line segments parallel to the coordinate axes, the directions of the line segments are classified into four directions by following the loop. Then, the total of the line segment lengths of the direction components of the real shape line segments corresponding to the respective line segments is calculated for each direction, and the line segment lengths of the line segments corresponding to direction 1 and direction 2 in the real shape are calculated. The average value of the sum is set to the sum of the lengths of the line segments having the directions 1 and 2 on the approximate model,
The same operation is performed for directions 3 and 4. Then, when the total value of the line segments in the same direction on the approximate model is determined, this total value is assigned to each straight line segment assigned in parallel to the coordinate axis according to the ratio of the line length in the actual shape. , And the length of each side is determined on the approximate model. In this way, an approximation model composed of only line segments parallel to the coordinate axes, in which the geometric characteristics and the phase characteristics are preserved as much as possible, is created.

After an approximate model is constructed for each constituent unit (boundary shape and holes included in the shape), an appropriate unit length is determined, and all sides are modified so as to be an integral multiple of this unit length. Based on the unit length, a grid can be formed independently of the boundary shape and the hole shape. Then, if a grid is formed on the boundary shape of the approximate model, this grid is mapped to an actual shape that does not consider holes by using the curve coordinate conversion method.

Here, the curve coordinate conversion method is, as shown in FIG.
A mathematical method for forming a uniform grid in an arbitrary shape based on an orthogonal grid.

If a grid is formed inside the boundary of the actual shape, a grid point closest to the feature point of the hole is obtained, and a correspondence is taken on the grid formed inside the boundary of the approximate model, and the boundary of the approximate model of the hole is obtained. Optimization of the relative position with respect to the approximate model of the shape is achieved,
An overall approximation model including the holes is constructed.

If an overall approximation model is created, as shown in FIG. 16, on the assumption that the phase state of the approximation model is maintained,
A recognition model converted so that each side takes a minimum integer value is configured. The length of each side of the corresponding approximate model is given as an attribute to each side of the recognition model.

 The recognition of the recognition model is performed in the following procedure.

Classification by size of recognition model (NX, NY) Classification by shape of recognition model Comparison based on line length of corresponding line of each side Registration corresponding to the original arbitrary shape by the above three steps of recognition procedure A figure is selected.

〔Example〕

FIG. 1 is a block diagram showing the overall configuration of an embodiment of the present invention. A graphic input unit 1 is provided which is connected to an arbitrary shape setting unit 8 and allows a user to directly input a shape from components such as a keyboard 2, a tablet 3, and a mouse 4. The graphic input unit 1 further includes a display control unit. CR with part 6
Connected to T display 5. A shape reading unit 7 is further connected to an input side of the arbitrary shape setting unit 8,
On the output side of the arbitrary shape setting unit 8, a shape for calculating the degree of approximation of each side to the coordinate axis and the degree of the same direction of two adjacent sides based on the information obtained by the curve conversion unit 9 and the geometric operation unit 10. Information generator 11
Is connected. The shape information generation unit 11 is further connected to an approximate model generation unit 15, and the geometric operation unit 10 is connected to the shape information generation unit 11 and the approximate model generation unit 15. The approximate model generation unit 15 is further connected to an overall approximate model generation unit 18, and the overall approximate model generation unit 18 is connected to a recognition model generation unit 19. The recognition model generation unit 19 is connected to the recognition result display unit 22, and the recognition result display unit 22 is further connected to the CRT display 5. The shape information generation unit 11 further includes a curve conversion unit 9 for linearly approximating a ridge line (boundary line) of an arbitrary shape and a direction assignment in a coordinate axis direction of each side determined by performing a fuzzy operation. Is connected to a matching check unit 14 for checking whether or not the condition is satisfied.

Based on the shape information, the approximate model generation unit 15 is connected to a fuzzy calculation unit 13 and the matching check unit 14 for correcting the degree of approximation of each side to the coordinate axis according to the fuzzy rule, and the fuzzy calculation unit 13 An ambiguous rule setting unit 12 for setting various rules for performing shape conversion is connected. The approximation model generation unit 15 that constructs an approximation model based on the geometric data obtained from the geometric operation unit 10 and the phase data obtained from the fuzzy operation unit 13 is based on a square lattice attached to the approximation model. A mapping operation unit 16 that generates a lattice in an actual shape using the curve coordinate transformation method is connected, and the mapping operation unit 16 includes a relative position calculation unit 17 that detects a relative position of the hole with respect to the boundary shape using the lattice. The overall approximation model generation unit 18 is connected via the above. In addition, the basic model database 21 is connected to the recognition model generation unit 19 that converts the approximate model into a recognition model, based on the recognition model.
Is provided with a recognition operation unit 20 for performing comparison with a figure registered in the base figure. The recognition operation unit 20 is connected to a basic figure database 21 in which basic figures are registered. The recognition calculation unit 20 also includes a recognition result display unit that displays the calculation result of the recognition calculation unit 20.
Connected to 22. The geometric operation unit calculates the angle between each straight line and the coordinate axis and the adjacent side, and calculates the length of each side of the approximate model based on the actual shape. The overall approximate model generation unit 18 configures an overall approximate model by combining the boundary shape and the approximate model of the hole based on the information provided from the approximate model generation unit 15 and the relative position calculation unit 17.
The CRT display 5 is connected to an arbitrary shape setting unit 8.

The above-described fuzzy rule setting unit 12 and fuzzy calculation unit 13 form a fuzzy calculation unit 30A, and the fuzzy calculation unit 30A
, A geometric operation unit 10, a shape information generation unit 11, and a matching confirmation unit
The approximation model generation unit 14 and the approximation model generation unit 15 constitute a unit 30 that converts a straight line segment into a line segment parallel to the coordinate axes.

Next, the operation of the above embodiment will be described. When the two-dimensional shape a shown in FIG. 2 is input from the graphic input unit 1 or the shape reading unit 7 to the arbitrary shape setting unit 8, the curve conversion unit 9 linearly approximates the curved portion of the shape. b is generated. Each line segment forming the shape b is assigned in a direction parallel to the x-axis or the y-axis, and is converted into an approximate model such as the shape c in FIG.

A method of configuring this approximate model will be described. First,
Regarding the generation of the phase information of the approximate model (information on which of the x and y axes each line segment is to be assigned to in parallel), the following four basic rules are used as ambiguous rules.

Rule 1: Each line is assigned parallel to the direction of the coordinate axis with the smallest possible angle.

Rule 2: With respect to two adjacent sides, the angle formed is assigned to a direction of different coordinate axes as much as smaller than a predetermined fixed angle.As the angle formed is larger than the predetermined angle, the directions are preferably allocated to the same coordinate axis. Can be

Rule 3: Line groups with a small change rate of inclination are assigned to one direction as much as possible.

Rule 4: Parallel lines are preferably assigned in the same direction.

Rule 1 is realized by initial setting of the degree of approximation to the coordinate axis obtained based on the angle between each line segment and the coordinate axis. x
As shown in FIGS. 3A and 3B, the degree of approximation to the x-axis and y-axis is represented by the angle (θx, θ
y), and the vertical axis represents the degree of approximation Px, P in the x-axis direction or the y-axis direction.
It is represented by a membership function taking y (0 ≦ Px ≦ 1, 0 ≦ Py ≦ 1). The degree of approximation is defined to be closer to 1 as the angle between the line segment and any one of the coordinate axes is closer to 0 degree, and closer to 0 as the angle is closer to 90 °.

Rule 2 is implemented by modifying the degree of approximation of each side to the coordinate axis based on the degree of approximation and the degree of the same direction, which is the relationship between two adjacent sides. The same direction of the two sides adjacent P _R
Is membership function as shown an angle theta _R of two sides adjacent the horizontal axis, the Figure 3C the same direction of _{P R (-1 ≦ P R ≦} 1) 1) taken on the vertical axis Indicated by In this case, the degree of the same direction approaches 1 as the angle between the two sides approaches 180 degrees, and changes from -1 to 180 as the angle approaches 180 degrees to 90 degrees. It is constant at 1.

The modification of the approximation will be described with reference to two line segments shown in FIG. 4 as an example. First, according to FIG. 3A, the x-axis approximation Px and the y-axis approximation Py of the line segment are 0.8 and 0.2, respectively, and the X and Y-axis approximations of the line segment are 0.4 and 0.6, respectively. since There is 108 degrees, the Figure 3C, the same direction of P _R of the line segment is -0.6, which means that strength two sides are assigned to different directions of the coordinate axes is 0.6. Therefore, based on these values, first, the degree of influence on a line segment is calculated. Note that the degree of influence is defined as indicating the degree of correcting the degree of approximation. The calculation of the degree of influence is realized by performing the following four operations.

(I) The degree of influence from the x-axis of the line segment to the x-axis of the line segment Since the two sides of the Qxx line segment have the same negative direction as described above, the x-likeness of the line segment is the x-likeness of the line segment. Is denied. As shown in FIG.5A, the degree of approximation of the line segment in the x-axis direction is 0.
8. Since the degree of approximation of the line segment in the non-x-axis direction is 0.6, the degree of influence Qxx from the x-axis of the line segment to the x-axis of the line segment is calculated by the following equation (1).

Qxx = (degree of approximation of line segment in x-axis direction) x (degree of approximation of line segment in non-x-axis direction) x (degree of same direction) ... (1) = 0.8 x 0.6 x (-0.6) = -0.288 (ii) The degree of influence from the y-axis of the line segment to the x-axis of the line segment Qyx Since the two sides formed by the line segment have the same degree of negative direction, the y-likeness of the line segment affirms the x-likeness of the line segment. As shown in FIG. 5B, the approximation degree of the line segment in the y-axis direction is 0.2, and the line segment x
Since the degree of axial approximation is 0.4, the degree of influence Qyx from the y-axis of the line segment to the x-axis of the line segment is calculated by the following equation (2).

Qyx = (degree of approximation of line segment in y-axis direction) × (degree of approximation of line segment in x-axis direction) × (same direction degree × (−1)) (2) = 0.2 × 0.4 × (−0.6) × ( -1) = 0.048 (iii) Influence from the x-axis of the line segment to the y-axis of the line segment Since the two sides composed of the Qxy line segments have the same negative direction, the x-likeness of the line segment is y I affirm the likeness. As shown in FIG. 5C, the degree of approximation of the line segment in the x-axis direction is 0.8, and the degree of approximation of the line segment in the y-axis direction is 0.6. , Calculated by the following equation (3).

Qxy = (degree of approximation of line segment in the x-axis direction) × (degree of approximation of line segment in the y-axis direction) × (same direction degree × (−1)) (3) = 0.8 × 0.6 (−0.6) × (− 1) = 0.288 (iv) The degree of influence from the y-axis of the line segment to the y-axis of the line segment Qyy Since the two sides formed by the line segments have the same degree of the same direction, the y-ness of the line segment is the y-ness of the line segment. Deny. As shown in FIG. 5D, the approximation degree of the line segment in the y-axis direction is 0.2, and the approximation degree of the line segment in the y-axis direction is 0.4. It is calculated by the following equation (4).

Qyy = (degree of approximation of the line segment in the y-axis direction) × (degree of approximation of the line segment in the non-y-axis direction) × (the same degree of the same direction) (4) = 0.2 × 0.4 × (−0.6) = − 0.048 (i) According to the calculation of (iv), the degree of influence on the x-axis direction approximation of the line segment is Qxx + Qyx = −0.288 + 0.048 = −0.24 The degree of influence on the y-axis direction approximation of the line segment is Qxy + Qyy = 0.288−0.048. = 0.24. The correction of the degree of approximation is executed by adding (influence degree × calculation constant) to the degree of approximation calculated in FIGS. 3A and 3B. For example, when the calculation constant is 0.1, the approximation of the line segment in the x-axis direction is reduced from 0.4 to 0.4 + (− 0.24) × 0.1 = 0.376, and the approximation of the y-axis is 0.6 to 0.6+ (0.24) × 0.1. = 0.624, and the direction assignment of the line segment is inclined in the y-axis direction.
Similarly, by calculating the degree of influence of a line segment on a line segment, a result is obtained in which the direction assignment of the line segment is inclined in the x-axis direction.

The above calculation is performed for all pairs of two adjacent sides in the graphic (shape b in FIG. 2) obtained by directly approximating the curved portion of the target graphic, and the degree of approximation is corrected as a whole. Next, the degree of influence is calculated based on the corrected degree of approximation,
Correct the approximation again. By repeating such operations,
In general, the degree of approximation of each side (line segment) converges to 1 in one direction (eg, x-axis direction approximation) and 0 in the other direction (eg, y-axis direction approximation). Converges to Then, by adopting the direction assignment in the convergence state, it is possible to generate the phase information of the approximate model as seen from the conversion from the shape b to the shape c in FIG.

For a two-dimensional figure, (x, y) × (x, y)
The degree of approximation is corrected by calculating the items, but for a three-dimensional figure, the degree of approximation is corrected by calculating nine items of (x, y, z) × (x, y, z).

In addition, although there are four basic vague rules as described above, in addition to these, as shown in the example of FIG. There are auxiliary rules, such as the assignment of three directions that must not exist in one three-dimensional surface to be assigned. By appropriately setting these rules, an approximate model is efficiently generated.

Next, a method of determining whether or not a shape is topologically established by the direction assignment obtained by the above method will be described with reference to FIGS. 7A and 7B as an example. A shape is obtained by linearly approximating the curved part of the target figure f having an arbitrary shape, the direction assignment of each line segment for generating an approximate model from the shape, and the direction of the line segment when the shape is traced counterclockwise. ^{a, x +, y +, x} -, y - expressed in. x ⁺ and y ⁺ are the x axis,
parallel to the y-axis direction, the direction in which the numerical value increases, x ^-, y ^-, respectively, showing a line segment x-axis, the numerical value parallel to the y-axis direction is assigned to a decreasing direction. Shapes g and i in FIG. 7A show examples of the direction of a line segment assigned to shape f. The difference between shape g and shape i is that, in shape g, the line segment at the upper left assigns x ⁻ . On the other hand, in the shape i, the corresponding line segment is assigned y ⁺ . When each figure is approximated by a line segment parallel to the assigned x-axis and y-axis, shape g becomes shape h and shape i becomes shape j. When tracing a figure composed of only line segments parallel to the x-axis and the y-axis in the counterclockwise direction, the corner around each line segment is one of the eight types shown in FIG. 7B,
Mark each corner with the corner symbol written in each corner of Figure 7B. The number written at each corner of the shape h and the shape j is this corner number.

If the assigned line segment is topologically aligned, if the assigned line segment is traced counterclockwise, the sum of the corner numbers will be 10 and the line segment will be traced clockwise. In this case, there is a property that the sum of the corner numbers becomes -10. No.
As shown in FIG. 7A, based on this property, it is determined whether or not the topological matching of the figure composed of the assigned line segment directions has been achieved.

As a countermeasure in a case where topological matching cannot be obtained, a method of changing the current direction assignment from a side having a high degree of ambiguity with reference to a past calculation result and searching for an assignment pattern that can achieve matching is used. There is.

Next, generation of geometric information of an approximate model (determination of the length of each side) will be described.

As shown in the shape m of FIG. 8, if the loop (boundary line forming the figure) is traced in one direction, the directions of the sides of the approximate model are classified into four directions (1) to (4) in the figure, and are classified into directions. The sum of the lengths of the sides divided by the direction is equal to the sum of the lengths of the sides classified by the direction. With respect to an actual shape such as shape n in FIG. 8, the average value of the total value of the sides in the direction and the total value of the sides in the direction is calculated, and this is calculated as the sum of the lengths of the sides classified into the approximate model and the direction. Set as a value. The same applies to the direction and direction. Thereby, since the total value of the lengths of the sides in each direction in the approximate model is determined, for each direction, the total length is proportionally divided based on the ratio of the lengths of the sides in the actual shape, The length is set as the length of each side of the approximation model, and the approximation model for each loop is completed as shown in shape 0 in FIG.

Here, an application example of this approximate model will be described. First, the curve coordinate conversion method will be described. What is the curve coordinate transformation method?
As shown in FIG. 9, when a grid shape γ composed of only an arbitrary shape p and a straight line parallel to the coordinate axis corresponding to the arbitrary shape p is set, a uniform grid is generated in the arbitrary shape by performing a mapping operation. This is a method for obtaining the shape q.

Therefore, when an arbitrary shape is set, an approximate model is created using the present invention, and the shape is corrected so that each side of the approximate model is an integral multiple of the unit length, and based on the unit length, If a grid is set as a grid shape and a curve coordinate conversion method is applied thereto, automatic division into finite elements of an arbitrary shape is performed.

FIG. 10 shows an example of automatic division of two-dimensional and three-dimensional figures.

In the case of two dimensions, an approximate model u2 of an arbitrary shape composed of only straight line segments parallel to the coordinate axes of the rectangular coordinates is generated from the arbitrary shape u1, and each straight line segment of u2 is an integral multiple of the unit length. By generating a mapping model having a grid shape in which a grid is stretched so as to perform a mapping operation for mapping the grid shape to an arbitrary shape u1, from the mapping model, a uniform grid is generated in u1. The generated analysis mesh shape u3 is automatically generated.

Similarly, in the case of three dimensions, from the arbitrary shape v1, an approximation model v2 of the arbitrary shape corresponding to the above u2, which is composed of only straight line segments parallel to the coordinate axes of the rectangular coordinate system, is generated, and v2 is a unit length. By generating a mapping model that is a grid shape divided so as to be a set of integer multiples of a cube consisting of, by performing a mapping operation to map the grid shape to an arbitrary shape v1
From the mapping model, a uniform grid was generated in v1,
An analysis mesh shape v3 corresponding to u3 is automatically generated. In FIG. 10 v3, only the grid in one cross section of the shape is shown to show the state of the internal grid.

Next, creation of an approximate model of a shape including a hole will be described. First, as shown in FIG. 11A, regarding the approximation model, the shape is modified so that each side becomes an integral multiple of the unit length independently of the boundary shape and the hole shape, and a grid is formed based on this length. . Next, as shown in FIG.11B, a grid is generated on the boundary shape of the real shape using the curve coordinate transformation method based on this grid, and the grid to which the feature point of the hole is closest is obtained. The correspondence on the grid attached to the boundary shape of the model is taken. Then, as shown in FIG. 11C, the approximate model of the hole-shaped grid is moved on the grid stretched on the boundary shape of the approximate model, and the total value of the difference between the corresponding points is minimized. The position is found, the relative position of the hole shape to the boundary shape is determined, and an overall approximation model as shown in FIG. 11D is created.

With this method, it is possible to automatically create an arbitrary lattice shape having holes, and thereby, finite element division relating to an arbitrary shape having holes as shown in FIG. 12 is also automated.

Next, a method of constructing a recognition model based on an approximate model and a recognition method using the same will be described.

If an overall approximation model is created, as shown in FIG. 13A, a recognition model modified so that each side takes a minimum integer value is constructed on the premise that the phase state of the approximation model is maintained. The length of the side of the corresponding approximate model is given as an attribute to each side of the recognition model.

In addition, in order to represent the relative position of the hole, as shown in FIG. 13B, a point having the smallest Y coordinate belonging to the leftmost side is detected independently of the boundary shape of the approximate model and the hole shape independently, and the boundary shape is determined. The distance in the actual shape between the corresponding point and the corresponding point of each hole shape is set as an attribute.

Then, based on the recognition model, recognition is performed according to the following three procedures.

(I) As shown in FIG. 13C, the size of the recognition model (NX, N
Classified by Y). NX and NY are the sum of the attribute values given to each side of the recognition model in the x-axis direction and the y-axis direction, respectively.

(Ii) Classification by shape of recognition model (iii) Comparison based on attributes (length of corresponding line segment, relative position of hole) given to the model, and an application example of this recognition method will be described below.

Consider a case where a two-dimensional shape So, to is given as shown in FIG. First, regarding the shape So, various shape deformations can be considered as shown in FIG. 14 S ₁ and S ₂ by changing the viewpoint position, but these are all replaced with the same recognition model shown in FIG. 14 S ₃ . Can be The shape to is also replaced by the same recognition model, but as shown in FIGS. 14 s and t, the shape So is clearly distinguished from the shape to by the attribute of the recognition model. As described above, by using the present invention, graphic recognition that is not easily affected by distortion from the original shape is performed.

In the above description, a two-dimensional figure has been described.
In the case of a two-dimensional figure, the edges of the figure are first approximated to straight line segments, and then a similar technique is applied.

FIG. 18 is a flowchart summarizing the procedure from the input of an arbitrary shape to the recognition of a figure as described above.

Conventionally, it was impossible to create an approximate model by a uniform mathematical method. However, according to the present embodiment, by using a membership function, various conversion rules related to shape conversion can be defined. Mathematical expression has become possible, and a universal shape transformation method that reflects judgments involving human ambiguity has been established. In determining the relative position of the hole shape with respect to the boundary shape, the generation of a uniform grid using the curve coordinate conversion method reduced the adverse effects due to the distortion of the boundary shape. Furthermore, by performing graphic recognition using a recognition model converted from an approximate model, even if a certain recognition shape is deformed from the original shape, a recognition result that is less affected by the shape is obtained. By dividing the procedure into three steps, the recognition work was made more efficient.

〔The invention's effect〕

According to the present invention, an approximate model consisting of a straight line parallel to a coordinate axis is automatically generated for an arbitrary shape. Therefore, a grid coordinate conversion method can be applied to this approximate model, and a curve coordinate conversion method can be applied. Has the effect of enabling automatic finite element division.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a schematic configuration of a shape conversion apparatus according to an embodiment to which the present invention is applied, FIG. 2 is a plan view showing an example of a conversion procedure to an approximate model, and FIGS. 3A to 3C are members. FIG. 4 is an explanatory diagram showing an example of an angle between a straight line segment and a coordinate axis, FIGS. 5A to 5D are graphs showing an example of the degree of approximation, and FIG. FIGS. 7A to 7B are plan views showing an example of the fuzzy processing, FIGS. 7A to 7B are explanatory diagrams showing an example of a method for confirming the phase matching of the approximate model, and FIG. 8 is a plan view showing a method for determining the length of each line segment of the approximate model. FIG. 9, FIG. 9 is a plan view showing an example of the curve coordinate transformation method, FIGS. 10 and 12 are plan views showing an example in which a finite element is divided from a figure, and FIGS. 11A to 11D show shapes with holes. FIG. 13A to FIG. 13C are plan views showing an approximate model creation procedure, and FIG.
The figure is a plan view showing the figure recognition procedure using the approximate model.
FIG. 15 is a plan view for explaining the directionality of the line segments constituting the approximate model, FIG. 16 is a plan view for explaining the change from the approximate model to the recognition model, and FIGS. FIG. 18 is a flowchart showing an example of a graphic recognition method, and FIG. 18 is a flowchart showing an example of a procedure for performing graphic recognition by applying the present invention. 30 ... means for converting a straight line segment into a line segment parallel to the coordinate axes,
30A ... Ambiguous calculation unit.

Continuation of the front page (58) Field surveyed (Int.Cl. ⁶ , DB name) G06F 17/50 G06T 17/20 JICST file (JOIS)

Claims (9)

(57) [Claims] 1. A means for inputting an arbitrary shape, and determining which of the coordinate axes a ridge line of the input arbitrary shape is to be a straight line segment parallel to the coordinate axis, and converting the arbitrary shape into the determined straight line A shape conversion device comprising: means for converting into a shape formed by line segments; and means for generating a grid in a shape formed by the straight line segments. 2. The shape conversion apparatus according to claim 1, wherein the means for converting into a shape composed of line segments parallel to the coordinate axes includes an ambiguous calculation unit for performing calculation using a membership function. 3. A means for approximating a boundary line or ridge line of an arbitrary shape with a plurality of straight line segments, and expressing the degree of approximation to the coordinate axis of each line segment by a variable from 0 to 1, and using a variable from an ambiguity rule to represent the variable. Means for correcting the whole and assigning each line segment in parallel to any one of the coordinate axes, and finally converging to one approximate model. 4. A means for approximating a boundary line or ridge line of an arbitrary shape with a plurality of straight line segments, and converting said line segments so that adjacent line segments are on a straight line or perpendicular to each other. Forming means. 5. A means for approximating a boundary line or ridge line of an arbitrary shape with a plurality of straight line segments, calculating an angle between each line segment and a coordinate axis, and calculating the angle based on the calculated angle and a preset fuzzy rule. Means for allocating the line segment to any one of the coordinate axes by a membership function to create the approximate model of the arbitrary shape. 6. The fuzzy rule of the means for creating an approximate model is that at least the line segment is assigned to the direction of the coordinate axis having the smallest angle with the line segment, and that two adjacent line segments are The shape conversion apparatus according to claim 5, wherein the angle is formed in a different direction as much as the angle formed is smaller than the fixed angle, and the same direction is allocated as the angle formed is larger than the fixed angle. . 7. A first means for inputting an arbitrary shape, and a second means for generating a model in which a grid is formed from a model formed by a plurality of straight line segments having an orthogonal and parallel relationship from the input arbitrary shape. Means, and third means for generating, from the generated model, an arbitrary shape in which the arbitrary shape is gridded,
A shape conversion device comprising: 8. The shape conversion apparatus according to claim 7, wherein the second means generates an approximate model of an arbitrary shape composed of only linear components parallel to the coordinate axes of the rectangular coordinate system from the arbitrary shape; Means for generating a mapping model having a lattice shape from the approximate model. 9. The shape conversion device according to claim 8, wherein the means for generating the mapping model sets a grid so that each linear component of the approximate model is an integral multiple of a unit length. Characteristic shape conversion device.