JP2005190045A - Repair method for data exchanged between cad systems - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a repair method for data exchanged between different CAD systems, capable of automatically repairing a flaw present inside the CAD data exchanged between the CAD systems. <P>SOLUTION: This repair method for the data exchanged between the CAD systems automatically repairing the flaw present inside the two-dimensional CAD data exchanged between the different CAD systems has steps: for range-designating a portion WR of the flaw on a two-dimensional image expressed by the two-dimensional CAD data; for designating an area W+ wherein the range-designated portion WR of the flaw is moderately expanded; for finding function data by use of the CSRBF method to an area WL obtained by subtracting the portion WR of the flaw from the area W+; and for performing interpolation of the portion WR of the flaw by use of the found function data to repair the area W+. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a data restoration method capable of repairing defects (that is, flaws) in CAD data exchanged between different CAD systems.

  CAD data exchange refers to exchanging two-dimensional or three-dimensional CAD data between different CAD systems. The exchange of data between different CAD systems has long been a challenge, and traditionally, common data exchange standards such as IGES (Initial Graphics Exchange Specification) (some versions have become US standards) and ISO STEP (ISO 10303) is defined as a CAD data exchange standard of (International Standards Organization).

However, in fact, when CAD data is exchanged, especially when it is solid, there are problems of geometric transformation errors and phase expression inconsistencies. Furthermore, in order to fully guarantee the information in units of features in both directions, shape generation is substantially possible. There is a condition that the internal representation is exactly the same including the history (see Non-Patent Document 1).
"CAE for Digital Process", JSME Workshop Material, No.98-22, March 1998

  Therefore, when creating a mesh from CAD data that has been exchanged, there is often a problem that the connection between elements is lost. In addition, there is a problem that it is difficult to create a mesh due to the presence of holes unnecessary for analysis, or the mesh is unnecessarily too fine.

  Conventionally, in the CAD system field, corrections caused by CAD data defects caused by CAD data exchange, more specifically, the inability to completely restore the structural shape when data is exchanged between different CAD systems. In addition, CAD is usually developed in a form that leads to CAM and is much finer than the information required for analysis. For example, corrections such as filling in holes that are unnecessary for analysis are all done manually. If it is a correction, or if a cloning approach is used, the cloning approach has the problem that even a small wound is forced to be tedious and computationally expensive.

  Hereinafter, a portion to be repaired (that is, an inconvenient portion as described above) in the CAD data exchanged between different CAD systems is defined as a scratch.

  The present invention has been made under the circumstances described above, and an object of the present invention is to provide a CAD system capable of automatically repairing a flaw existing in CAD data exchanged between different CAD systems. It is to provide a method for repairing data exchanged between the two.

  The present invention relates to a method for repairing data exchanged between CAD systems, and the above object of the present invention is to automatically repair flaws existing in two-dimensional CAD data exchanged between different CAD systems. A method for repairing data exchanged between systems, comprising: a step of specifying a range of a wound portion WR on a two-dimensional image represented in the two-dimensional CAD data; A step of designating a region W + widened, a step of obtaining function data using the CSRBF method for a region WL obtained by subtracting the scratched portion WR from the region W +, and using the obtained function data And interpolating the flaw portion WR to repair the region W +.

  The above object of the present invention is a method for repairing data exchanged between CAD systems, which automatically repairs a flaw existing in two-dimensional CAD data exchanged between different CAD systems, A first step of specifying a range of a scratched part on the two-dimensional image represented in the two-dimensional CAD data; a second step of determining a predetermined point within the range of the specified scratched part; A third step of specifying a predetermined peripheral region portion adjacent to a predetermined point of the flaw portion determined in the second step, and the predetermined peripheral region specified in the third step Using the CSRBF method for the part, a fourth step for obtaining function data, and using the function data obtained in the fourth step, interpolating a predetermined point of the flaw part, Predetermining the scratched part A fifth step of repairing a point, a sixth step of determining a next predetermined point of the flawed portion, and a next predetermined point of the flawed portion determined in the sixth step A seventh step of specifying a predetermined peripheral region portion adjacent to the eighth step, and an eighth step of obtaining function data using the CSRBF method for the predetermined peripheral region portion specified in the seventh step And the function data obtained in the eighth step is used to interpolate the next predetermined point of the flawed portion, and to repair the next predetermined point of the flawed portion. And a predetermined point next to the wound portion repaired in the ninth step until the repair of all the scratches of the wound portion specified in the first step is completed. The tenth step is returned to the sixth step. And the predetermined peripheral region portion specified in the seventh step includes a predetermined point of the scratch portion. In the ninth step, the scratch portion In the repair of the next predetermined point, the data after the repair of the predetermined point of the flaw portion is used.

  The above-mentioned object of the present invention is a method for repairing data exchanged between CAD systems, which automatically repairs a flaw existing in three-dimensional CAD data exchanged between different CAD systems, A wound contour extraction step of finding all boundary edges that can become boundaries at a polygonal scratch portion on the three-dimensional image represented in the three-dimensional CAD data, and forming a closed fold line from the obtained boundary edges; The polygonal flaw represented by the closed fold line is triangulated to sew and repair by contour stitching step and the flawed portion obtained by triangulation obtained in the contour stitching step by a predetermined algorithm. And performing a smoothing step, and the predetermined algorithm is applied to each triangle obtained in the stitching step of the contour. A step of obtaining detailed point data inside the flaw, a step in which a flawed portion WR is moderately widened is set as W +, a region in which the flawed portion WR is subtracted from W + is set as WL, and a region WL , Using the CSRBF method to obtain function data, and using the obtained function data, interpolating with detailed point data inside the flaw to repair the region WL. Is achieved.

但し、ｒ（Ｐ_ｉ，Ｐ_ｊ）は、複数の離散点の中の任意の２点Ｐ_ｉとＰ_ｊの距離
ｒ_０は、初期値として与える任意の点Ｐ_ｉを中心とした半径
と定義され、そのときの関数データは、

但し、λ_ｉ（ｉ＝１，２，．．．，Ｎ）は点Ｐ_ｉにおける基底関数の係数
λ_Ｎ＋１，λ_Ｎ＋２，λ_Ｎ＋３，λ_Ｎ＋４は１次項の係数
で表わされる関数により与えられ、前記係数λ_ｉ，λ_Ｎ＋１，λ_Ｎ＋２，λ_Ｎ＋３，λ_Ｎ＋４の計算は、補間行列を対角化する高速アルゴリズムを用いて行い、前記高速アルゴリズムは、前記複数の離散点が含まれる初期データから１つの点がリストに加えられ、前記リストに加えられた点は前記初期データから除かれる第１ステップと、前記第１ステップで加えられた点に対する近傍点が、前記初期データの中から検索されて前記リストに加えられ、前記リストに加えられた点は前記初期データから除かれる第２ステップと、前記第２ステップで加えられた点に対する近傍点が、前記初期データの中から検索されて前記リストに加えられ、前記リストに加えられた点は前記初期データから除かれる第３ステップと、前記リストに加えられる点が無くなったら、前記初期データに残っている離散点から１つの点が前記リストに加えられ、前記リストに加えられた点は前記初期データから除かれる第４ステップと、前記初期データに離散点がすべて無くなるまで、これらの前記第１ステップから前記第４ステップが繰り返され、バンド性を持つ対角行列が作成される第５ステップとを有するようにすることによって達成される。 Furthermore, the object of the present invention is that the basis function in the CSRBF method is:

However, _{r (P} i, _{P j),} the distance r ₀ of any two points _{P i} and _{P j} in a plurality of discrete points, radius defined around the arbitrary point _{P i} which gives the initial value The function data at that time is

However, λ _i (i = 1, 2,..., N) is a coefficient of the basis function at the point P _i λ _{N + 1} , λ _{N + 2} , λ _{N + 3} , λ _{N + 4} is given by a function expressed by a coefficient of the first-order term, The coefficients λ _i , λ _{N + 1} , λ _{N + 2} , λ _{N + 3} , and λ _{N + 4} are calculated using a high-speed algorithm that diagonalizes the interpolation matrix, and the high-speed algorithm uses the initial data including the plurality of discrete points. A point is added to the list, the point added to the list is removed from the initial data, and a neighboring point for the point added in the first step is searched from the initial data. A second step in which the points added to the list are excluded from the initial data, and whether neighboring points to the points added in the second step are in the initial data. Are added to the list, and points added to the list are removed from the initial data, and when there are no more points to be added to the list, 1 is added from the discrete points remaining in the initial data. Four points are added to the list, the points added to the list are removed from the initial data, and the first step to the fourth step until all the discrete points are removed from the initial data. Is repeated to have a fifth step in which a diagonal matrix with banding is created.

  According to the method for repairing data exchanged between CAD systems according to the present invention, it is possible to repair a flaw caused by the fact that the structural shape cannot be completely restored when data is exchanged between different CAD systems. Can do.

  Further, according to the present invention, it is possible to automatically perform correction such as filling a fine hole which is present in the CAD data and is unnecessary for CAM analysis. In other words, it is only necessary to apply the present invention to such unnecessary fine holes as scratches.

  Furthermore, according to the present invention, for example, it took about 3 months to develop a collision analysis model for an automobile, but there is an excellent effect that a shortening of about 1 month can be obtained.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  In the restoration method of data exchanged between CAD systems according to the present invention, a structure represented in CAD data by using a CSRBF (compactly supported radial basis functions) technique and a LCSRBF (local compactly supported radial basis functions) technique. The biggest feature is to divide the area into parts and search for shapes that are unsuitable for analysis and make corrections so that they are smooth.

  The present invention focuses on solving the above-mentioned conventional problems occurring in the CAD system field by image restoration technology.

  The CSRBF technique used in the present invention is one of the techniques for reconstructing the surface of an object, and generates an implicit function f (x, y, z) from discrete points (x, y, z). It is. The basis function in the CSRBF method is defined as follows. That is,

但し、ｒ（Ｐ_ｉ，Ｐ_ｊ）は、複数の離散点の中の任意の２点Ｐ_ｉとＰ_ｊの距離で、ｒ_０は、初期値として与える任意の点Ｐ_ｉを中心とした半径である。そして、任意の点Ｐ_ｉを中心とした半径ｒ_０の球に対し、その球内のすべての離散点について、中心Ｐ_ｉとの距離を求める。このとき、関数データは、以下のように表わされる関数により与えられる。

Here, r (P _i , P _j ) is the distance between any two points P _i and P _j among a plurality of discrete points, and r ₀ is the radius around any point P _i given as an initial value. It is. Then, for a sphere having a radius r ₀ centered on an arbitrary point P _i , the distance from the center P _i is obtained for all discrete points in the sphere. At this time, the function data is given by a function expressed as follows.

但し、λ_ｉ（ｉ＝１，２，．．．，Ｎ）は点Ｐ_ｉにおける基底関数の係数で、λ_Ｎ＋１，λ_Ｎ＋２，λ_Ｎ＋３，λ_Ｎ＋４は１次項の係数である。本発明で用いられるＣＳＲＢＦ法は、この関数を用いて表面再構成を行うものである。

Here, λ _i (i = 1, 2,..., N) is a coefficient of a basis function at a point P _i , and λ _{N + 1} , λ _{N + 2} , λ _{N + 3} , and λ _{N + 4} are coefficients of a primary term. The CSRBF method used in the present invention performs surface reconstruction using this function.

  The function data specifically refers to image data in which the function of f (x, y, z) is stored. This may be data in which only the above function is stored, or may include other information.

Among the functions given as described above, in the calculation of the spline coefficients λ _i , λ _{N + 1} , λ _{N + 2} , λ _{N + 3} , λ _{N + 4} , the CSRBF method used in the present invention uses an interpolation matrix (SLAE matrix) that is calculated. ) Is converted into a diagonal matrix with band characteristics based on an octree. This is done to reduce the amount of calculation in the computer even when there are many discrete points. If all spline coefficients are superimposed, a smooth function system that passes through all points can be created.

Hereinafter, the flow of a high-speed algorithm for diagonalization in the CSRBF method used in the present invention will be described in detail with reference to FIG. First, one point is added to an output list from initial data (input list) including a plurality of discrete points (step 21), and the point added to the list is deleted from the initial data (step 22). . Then, a neighboring point for the added point is searched from the initial data (step 23). If there is a nearby point (step 24), the process returns to step 21 again, the point is added to the list, and the point added to the list is deleted from the initial data (step 22). Then, a neighboring point for the added point is searched again from the initial data (step 23). Steps 21 to 24 are repeated until there are no neighboring points in the initial data for the added point. When there are no neighboring points (step 24), it is checked whether there are still discrete points in the initial data (step 25). If so, return to step 21 and add one of the points to the list. Similarly, the points added to the list are deleted from the initial data (step 22), and the neighboring points for the added points are searched again from the initial data (step 23). In this way, steps 21 to 25 are repeated until there are no discrete points in the initial data. Thereby, a diagonal matrix (band matrix) having a band property is generated. By calculating a newly rearranged matrix using this high-speed algorithm, the calculation time of the coefficients λ _i , λ _{N + 1} , λ _{N + 2} , λ _{N + 3} , and λ _{N + 4} can be reduced. A program for realizing the algorithm as described above is, for example, as shown in FIG.

The case where the above high-speed algorithm is applied to an image having six discrete points {1, 2, 3, 4, 5, 6} in a cube will be described. Assume that for each of these six points, neighboring points in a sphere having an arbitrary radius r ₀ are given as follows.

Neighboring points of point 1: 1, 3, 6
Neighboring points of point 2: 2, 5
Neighboring points of point 3: 1, 3, 4
Neighboring points of point 4: 3, 4
Neighboring points of point 5: 2, 5
Neighboring points of point 6: 1, 6
When the fast algorithm of the CSRBF method used in the present invention is applied to such 6 discrete points, the discrete point sequence is {1, 3, 6, 4, 2, 5} as shown in FIG. From this, a diagonal matrix (SLAE matrix) having band characteristics as shown in FIG. 4 is generated. In addition, x in FIG. 4 shows the distance between two points. By using such a high-speed algorithm, it is possible to calculate necessary parameters at high speed by the CSRBF method. Therefore, the function calculation by the CSRBF method used in the present invention can be performed at a very high speed.

When the radius r ₀ in the basis function is large, most of the points are included in the sphere having the radius r ₀ centered on the arbitrary point P _i . Therefore, in this case, more storage space must be created for the diagonal matrix than the matrix before rearrangement. However, if the radius r ₀ is reduced, the best resulting matrix with less storage space can be used. The radius r ₀ may be determined empirically so that a highly accurate image can be generated in a short time.

  The function f (x, y, z) is a notation for a three-dimensional image. However, in the case of a two-dimensional image, z = 0, and the function is given by f (x, y). . Further, the color information of the image is given as a function value. For example, when f (10, 20) = 30, a point having an x coordinate of 10 and a y coordinate of 20 indicates a point having color information of 30.

  Further, the discrete points may be points distributed in a lattice pattern or points distributed discretely. For example, when the original image is an image of 160 × 120 pixels, when a function is created for the entire image, the number of discrete points is 19,200. However, when the preprocessing described later is performed, the number is smaller than this. It becomes a number. For example, when only 10% of points are used, there are 1920 discrete points. The 1920 discrete points may be arranged in a grid pattern. However, as described later, when feature points are extracted, they exist irregularly along the feature points.

  The function f (x, y, z) for the three-dimensional image of the CSRBF method used in the present invention described above, and the function f (x, y) for the two-dimensional image (that is, z = 0, so f (x, y, z) = f (x, y)) is simply written as a function f (P) as follows.

  First, a method for restoring data exchanged between CAD systems according to the first embodiment of the present invention will be described. The first embodiment of the present invention relates to a method for repairing flaws on two-dimensional CAD data exchanged between different CAD systems. In the present embodiment, the two-dimensional CAD data represents a two-dimensional image with a smooth color change.

  Further, the scratch referred to in the present invention means an inconvenient part on CAD data exchanged between different CAD systems, that is, for example, CAD exchanged when data is exchanged between different CAD systems. A portion in the data where the structural shape cannot be completely restored (that is, the data in that portion is lost), and a fine hole unnecessary in the CAM analysis existing in the CAD data can be considered as a scratch. .

  The procedure of the method for restoring data exchanged between the CAD systems of the first embodiment of the present invention is as follows.

  First, it is assumed that a scratched portion on a two-dimensional image represented by two-dimensional CAD data is represented by WR. An area where the scratched portion WR is slightly widened is denoted by W +, and an area obtained by subtracting WR from W + is denoted by WL. At this time, the area W + and the scratched part WR need to be given by the user as initial values. For example, Adobe (registered trademark) Photoshop (registered trademark) GUI interface (registered trademark), Photopaint (registered trademark) By using a so-called selection tool of image processing software such as Gimp (registered trademark), it is possible to easily specify and determine the region W + and the range of the scratched portion WR. If the region W + and the scratched portion WR are determined, the region WL is also determined.

  Next, the above-described CSRBF method is applied to the determined region WL to restore the two-dimensional CAD data (two-dimensional image). In other words, the function f (P) is obtained by applying the CSRBF method to the region WL, and repair is performed by continuously and smoothly interpolating the damaged portion.

The procedure of the above repair method is summarized as follows.
Step 1
A scratched part WR on the two-dimensional image (on the two-dimensional CAD data) and a region W + where the scratched part WR is slightly expanded are determined.
Step 2
The above-described CSRBF method is applied to the region WL, which is a region obtained by subtracting the scratched portion WR from the region W +, and the function f (P) is obtained.
Step 3
Using the obtained function f (P), interpolation of the flaw portion WR is performed to restore the region W +.

  The method for recovering data exchanged between the CAD systems according to the first embodiment of the present invention having the above-described steps 1 to 3 is the two-dimensional CAD data (two-dimensional image) in which the color change occurs relatively smoothly. ) Effective for repairing the upper wound. The reason is that the CSRBF method used in the method for restoring data exchanged between the CAD systems according to the first embodiment of the present invention has a smooth function f from a discrete point on two-dimensional CAD data (two-dimensional image). This is because (P) is obtained and surface reconstruction is performed.

  The data repair method for exchanging data between the CAD systems according to the first embodiment of the present invention is to repair the entire wound at once, and the wound portion is corrected from the periphery of the specified wound region. It is possible. Further, in the present invention, since the region WL, which is a predetermined peripheral region surrounding the wound portion WR, is expressed as a function, the wound portion WR can be repaired with high accuracy in a short time.

  On the other hand, for the scratches on the two-dimensional CAD data representing a two-dimensional image in which the color change sharply occurs like a texture pattern, the data exchanged between the CAD systems of the first embodiment of the present invention When the restoration method is applied, the CSRBF method that performs smooth function display is used as it is, and the edge of the two-dimensional image becomes blurred, and there is a problem that it is not suitable in terms of accuracy.

  Next, a method for restoring data exchanged between CAD systems according to the second embodiment of the present invention will be described. The second embodiment of the present invention relates to a method for repairing flaws on two-dimensional CAD data exchanged between different CAD systems using an LCSRBF (local compactly supported radial basis functions) method that uses the CSRBF method locally and repeatedly. Is. In this embodiment, the two-dimensional CAD data represents a two-dimensional image in which a color change occurs sharply like a texture pattern.

  Therefore, the above-described problem can be solved by using the method for restoring data exchanged between the CAD systems according to the second embodiment of the present invention. This is because the LCSRBF method used in the method for restoring data exchanged between CAD systems according to the second embodiment of the present invention uses the CSRBF method locally to express a function. In the LCSRBF method, the basis functions and equations used in the formulation are the same as those in the CSRBF method, but the application method is different. In order to reproduce the original more faithfully, the LCSRBF method is different from the CSRBF method, which emphasizes the smoothness of the image, and reconstructs the result obtained by using the CSRBF method in the previous step. It is characterized by that.

  Here, using the data restoration method exchanged between the CAD systems of the second embodiment of the present invention, for example, on the texture pattern of the image size represented by the two-dimensional CAD data of 128 × 128 pixels from the lower left. Consider a repair procedure when there is a single straight scratch to the upper right. The lower left pixel where the linear scratch begins is the 'scratch start point' (ie, 'scratch first point'), the pixel next to the start point is 'second scratch point', and this second point is adjacent This pixel is designated as the “third point”, and the pixel at the upper right where the flaw is repeated is designated as the “wound end point” (that is, the “128th point of the flaw”). In such a case, the linear flaw on the texture pattern is repaired for each point from the start point to the end point by the following procedure.

  First, repair of the 'start point of the wound' (that is, 'the first point of the scratch') is performed using a function f (P) obtained from the decimal point around the 'start point of the scratch' using the CSRBF method. Next, for the second point of the wound, which is the next point after the repaired first point of the wound, find the decimal points around the second point of the wound, Using the CSRBF method, a new function f (P) is created and repaired. When repairing the 'second point of the wound', the repaired value is used for the 'first point of the wound' which is the surrounding point.

  In this way, for the 'scratch L + 1st point', which is the next point after the repaired 'scratch Lth point' (2 ≦ L ≦ 127), around the 'scratch L + 1st point' A small number of points are obtained, and using these points, a new function f (P) is created and repaired using the CSRBF method. When repairing the ‘L + 1 point of the wound’, the repaired value is used for the ‘Lth point of the wound’, which is the surrounding point. Repeat until the 'end of wound'.

  It should be noted that it is preferable to employ a number of pixels within a radius t = M / N centered on the point to be repaired that is half or more of the surrounding decimal points used when repairing each point of the scratch. Here, N_t indicates the image size (N_t = 128 in this example). M is given by the user with an initial value of 1 to N_t. Here, M = 3 is used as a value satisfying both the calculation cost and the obtained image accuracy as a result of numerical calculation using various M for repairing scratches on a large number of texture patterns.

  As described above, the method for restoring data exchanged between the CAD systems of the second embodiment of the present invention using the LCSRBF method repeatedly uses the CSRBF method for each scratch point from the scratch start point to the scratch end point. After the 'second point of the wound', the greatest feature is that the repair result of the previous point is reflected. By using such an LCSRBF method, it becomes possible to accurately repair a scratch on a texture pattern, which is difficult with the CSRBF method.

The restoration procedure of the restoration method of the data exchanged between the CAD systems of the second embodiment of the present invention is as follows.
Step 10
The range of the scratched part on the two-dimensional image (on the two-dimensional CAD data) is designated.
Step 11
First, a predetermined point within the range of the specified scratch portion is determined. For example, the endmost point within the range of the scratched part is determined as a predetermined point of the scratched part. A predetermined peripheral area portion adjacent to a predetermined point of the determined flaw portion is specified, and a function f (P) is generated using the CSRBF method for the specified peripheral area portion. Then, using the generated function f (P) for the peripheral region portion, a predetermined point of the scratch portion is interpolated to repair the scratch.
Step 12
Next, the next predetermined point of the flaw portion is determined. For example, a point adjacent to a predetermined point on the scratched part is determined as a predetermined point next to the scratched part. A predetermined peripheral region portion adjacent to the predetermined portion of the scratch portion is specified in the same manner as the predetermined point of the scratch portion. Then, a function f (P) is generated using the CSRBF method for the specified peripheral region portion, and the next predetermined point of the flaw portion is interpolated using the generated function f (P). Repair the wound. In repairing a predetermined point next to the scratched portion, data after repairing the predetermined point of the scratched portion is used.
Step 13
By the time the repair of all the wounds in the range-specified wound part is completed, a predetermined point next to the repaired wound part is set as a predetermined point of the wound part, and the above step 12 is repeated.

  According to the method for recovering data exchanged between the CAD systems of the second embodiment of the present invention, which includes the above steps 10 to 13, the image damage such as a texture pattern in which the color change sharply occurs is shown. Also for the repair, a highly accurate repair is possible.

  Finally, a method for restoring data exchanged between CAD systems according to the third embodiment of the present invention will be described. The third embodiment of the present invention relates to a method for repairing flaws on three-dimensional CAD data exchanged between different CAD systems. Note that the data restoration method exchanged between the CAD systems according to the third embodiment of the present invention is effective for restoration of polygonal flaws on the three-dimensional CAD data (three-dimensional image).

A method for restoring data exchanged between CAD systems according to the third embodiment of the present invention is a method of extracting a flaw outline of a polygonal flaw on three-dimensional CAD data (three-dimensional image), stitching the outline, and space space mapping method. It is characterized in that it is performed in three stages, such as a smoothing method based on. The procedure at each stage is as follows.
<1> Extraction of Wound Contours Wound contour extraction means finding all polygonal flaw outlines on the three-dimensional CAD data (three-dimensional image), and the procedure is as follows.
Step 100
All sides (boundary sides) that can become boundaries at the polygonal scratches on the 3D CAD data (3D image) are found.
Step 101
A broken line closed from the boundary obtained in step 100 is constructed according to the procedure shown in steps 102 to 104 below.
Step 102
First, one boundary edge e ₁ is selected from all the boundary edges obtained in step 100 above. At this time, the start point of the boundary side e ₁ is v ₁ and the end point is v ₂ .
Step 103
Then, the v ₂ Locate boundary side as the start point and e _2, the end point of the boundary edge e ₂ and v _3.
Step 104
This procedure is repeated until the vertex v ₁ is returned to form one closed fold line.

At this time, it is assumed that each boundary side is included in only one closed fold line. Therefore, the same boundary side cannot be used for two or more closed folding lines. Here, closed fold lines obtained at a finite number of vertices may intersect, and a plurality of closed fold line assemblies may be considered.
Step 105
Of the closed fold lines obtained in steps 102 to 104, those having an intersection point are combined at the intersection point and recreated into one closed fold line. Repeat this step until there are no intersections.

As a specific example, for example, as shown in FIG. 5 (A), two closed lines L ₁ = (v ₁ , v ₂ ,... It is assumed that v _n ), L ₂ = (w ₁ , w ₂ ,..., w _m ) is configured.閉折line _{L 1} and _{L 2} has two intersections, one of which, vertex v _p1 in閉折line _{L 1,} has become vertex w _p2 in閉折line _{L 2.} At this time, as shown in FIG. 5 (B), L ₁ and L ₂ are combined at step S105 from the above step 105 to form one closed fold line. In this way, by connecting closed fold lines at intersections, one closed fold line including self-intersection can be created, and the number of intersections can be reduced by one.
<2> Contour stitching At the contour stitching stage, polygonal flaws represented by closed fold lines are divided into triangles and stitched together for repair. Assume that a closed fold line L = (v ₁ , v ₂ ,..., V _n ) is obtained in the wound contour extraction stage. At this time, it is assumed that there is no knot on the closed fold line L. Here, the closed folding line L is divided into n-2 triangles so as to satisfy the following three conditions. T is a triangulation of the closed folding line L.
Condition 1:
The vertex of the triangulation T is the vertex of the closed fold line L.
Condition 2:
The sides of T have an appropriate common part. That is, no two triangles overlap, but may be adjacent at some vertex or some side.
Condition 3:
The boundary of T coincides with L.

The number of triangulations considered in the above three conditions is enormous. For example, if there are n points v ₁ , v ₂ ,..., V _n , 2 ^{m + 1} (2m + 1) / (m + 2) !, m ≧ 1 possibility of triangulation for m = n−2 diagonals There is. It is necessary to select the optimal triangulation from such a huge number, but if the approach for that is bad, the calculation cost will increase and even calculation of not much point data will not be possible .

、辺ｖ_１,ｖ_ｋを含み三角形△_1knに隣接する面の法線ベクトルを

、辺ｖ_ｋ,ｖ_ｎを含み三角形△_1knに隣接する面の法線ベクトルを

、辺ｖ_ｎ,ｖ_１を含み三角形△_1knに隣接する面の法線ベクトルを

としたとき、ベクトル

その周囲のベクトル

とのスカラー乗法の和を最小にするものを最適な三角形分割としている。つまり、隣るようなＴ_1nを見つけている。 In order to obtain the optimum triangulation, an algorithm for partially verifying the entire work proposed by Cormen et al. In 1992 was used. Using this technique, for example, two triangulation parts (T _1k , T _kn ) so that one triangulation (T _1n ) obtained for a certain closed fold line is connected by one triangle (Δ _1kn ). ) And check whether each of the divided triangulations T _1k and T _kn is optimal. Actually, here, the normal vector of triangle △ _1kn is

, The normal vector of the surface that includes sides v ₁ and v _k and is adjacent to the triangle Δ _1kn

, The normal vector of the surface that includes sides v _k and v _n and is adjacent to the triangle Δ _1kn

, The normal vector of the surface that includes sides v _n and v ₁ and is adjacent to the triangle Δ _1kn

And the vector

Its surrounding vector

The optimal triangulation is the one that minimizes the sum of the scalar multiplications. That is, T _1n that is _adjacent is found.

This approach is designed to reduce the computation time because the resulting set of triangles is included in more than one triangulation to avoid repeated verification on that set of faces as much as possible. Can do. FIG. 6 shows the result of actually reconstructing a polygonal flaw by dividing it into triangles. FIG. 6A shows the original image, and FIG. 6B shows the result of reconstruction by optimal triangulation.
<3> Smoothing Method Based on Spatial Space Mapping Method The repair result of the polygonal flaw on the three-dimensional CAD data (three-dimensional image) performed in the above-described flaw outline extraction stage and the contour stitching stage is a cross-sectional view. And far from the original. The actual object is smooth, and as shown in FIG. 6 (B), it is not sufficient for the repaired portion to have an angular shape as obtained in the above-described wound contour extraction step and contour stitching step.

  Therefore, the surface is reconstructed and smoothed by the following procedure for the flaw portion obtained by the triangulation obtained in the stitching step of the contour. Here, similarly to the method for restoring data exchanged between the CAD systems according to the first embodiment of the present invention, the wound portion is WR, the area where the wound portion WR is slightly expanded is W +, and WR is subtracted from W +. Let the region be WL.

については、逆写像関数

で再構成点

を求めている。 The procedure of the smoothing method based on the spatial space mapping method is as follows.
Step 200
For each triangle obtained at the contour stitching stage, detailed point data inside the flaw is obtained in steps 201 to 203 below.
Step 201
Find the average triangle size when the original object is triangulated.
Step 202
If the triangle is larger than the average size in the scratched part, one point is supplemented to the center of gravity of the triangle, and the center of gravity and the three vertices of the triangle are connected to divide into three regions (that is, three new triangles).
Step 203
Step 202 is also applied to the newly formed triangle. Step 202 is repeated until there are no more triangles larger than the average size.
Step 204
The CSRBF method is applied to the region WL to obtain the function f (P).
Step 205
Using the obtained function f (P), interpolation is performed with the detailed point data inside the flaw obtained in steps 201 to 203 to repair the region WR. At this time, the height component at each point of the region WR

For the inverse mapping function

Reconstruction point at

Seeking.

  The CSRBF method is a technique for performing smooth function conversion from given point data and performing image reconstruction. Therefore, the CSRBF method is suitable for smoothing the image surface of the present invention.

  FIG. 7 shows a method of repairing data exchanged between CAD systems according to the third embodiment of the present invention for the problem of scratch repair when there are 16 holes on 3D CAD data (3D image). It is the result of applying.

  FIG. 7A is an original image having 16 holes before the present invention is applied. The left and right figures in Fig. 7 (A) are the original images viewed from different angles. The left diagram of FIG. 7B shows the result of triangulating a polygonal flaw at the contour stitching stage. The right figure of FIG. 7 (B) shows the result of detailed point data inside the triangular flaw obtained by the above steps 201 to 203 in the smoothing method stage based on the spatial space mapping method. FIG. 7C shows the result of smoothing using the CSRBF method, that is, the final result of applying the method for restoring data exchanged between the CAD systems of the third embodiment of the present invention.

  By using the method for recovering data exchanged between the CAD systems of the third embodiment of the present invention, it is possible to automatically find out a flaw portion on the three-dimensional CAD data and reconstruct the flaw portion. . In addition, the original image shown in FIG. 7A is composed of 19467 polygons, and the 16 holes that are scratches are composed of 372 sides. The repair time of the scratch on the image as shown in FIG. 7A is 7.3 seconds, which is a practical level of speed. Further, from the final result of FIG. 7C, it is well understood that the accuracy is sufficient.

  FIG. 8 shows a method for repairing data exchanged between CAD systems according to the third embodiment of the present invention for a problem of repairing scratches when there are 211 holes on 3D CAD data (3D image). It is the result of applying. FIG. 8A shows an original image having 211 small holes before the present invention is applied. FIG. 8B shows the result of applying a part of the procedure of the method for restoring data exchanged between the CAD systems according to the third embodiment of the present invention. That is, FIG. 8B is a result of performing only the procedures of the wound contour extraction stage and the contour stitching stage in the data restoration method exchanged between the CAD systems of the third embodiment of the present invention. Smoothing is not performed using the CSRBF method. The repair time of the scratch on the image as shown in FIG. 8A is 3 seconds, which is a practical level of speed. Further, from the result of FIG. 8B, it is well understood that the accuracy is sufficient.

  Comparison of the repair example of FIG. 7 and the repair example of FIG. In other words, when the scratches are small, only the procedure of the wound contour extraction step and the contour stitching step in the method for recovering data exchanged between the CAD systems of the third embodiment of the present invention is performed, so that high-precision repair is performed. Results are obtained. On the other hand, when the scratch becomes large to some extent, smoothing is further performed using the CSRBF method, so that a highly accurate restoration result can be obtained.

  Also, in some cases, images on the surface of a three-dimensional model may be used for visualization, for example to represent material (wood, metal, etc.). Such an image is called a texture pattern image. By using the data restoration method exchanged between the CAD systems according to the second embodiment of the present invention, it is possible to repair such scratches on the texture pattern image. An example is shown in FIG. FIG. 9A shows an original image before applying a method for restoring data exchanged between CAD systems according to the second embodiment of the present invention. FIG. 9B shows an image restored using the data restoration method exchanged between the CAD systems according to the second embodiment of the present invention.

  Furthermore, in another case, for example, in order to simplify the analysis of the three-dimensional model, it is necessary to remove some features of the three-dimensional model. At that time, the problem can be solved by using the method for restoring data exchanged between the CAD systems of the third embodiment of the present invention. An example is shown in FIG. FIG. 10A shows the original three-dimensional model before applying the present invention, and there are two cavities near the bottom of the plate in front of the three-dimensional model. When analyzing the three-dimensional model, these two cavities can be omitted. FIG. 10B shows a state in which the two cavities are selected and a cut is made to remove the two cavities. FIG. 10C shows the final result after applying the method for recovering the data exchanged between the CAD systems of the third embodiment of the present invention, where two cavities are removed from this previous plate. You can see that it has been removed.

  Note that the method for restoring data exchanged between CAD systems of the present invention can be easily incorporated into existing devices, and is therefore effective as a smoothing for removing a dent generated in a laser hit portion by a laser scanner. is there. As described above, the method for repairing data exchanged between CAD systems according to the present invention has an advantage that it can perform high-speed and high-precision repair of scratches on the exchanged CAD data and has a wide application range. Yes.

It is a flowchart for demonstrating the flow of the high-speed algorithm in the CSRBF method used by this invention. It is an example of the program for implement | achieving the high-speed algorithm in the CSRBF method used by this invention. It is a figure for demonstrating the process in which a discrete point is rearranged by the high speed algorithm in the CSRBF method used by this invention. It is a figure showing an example of the diagonal matrix with the band property produced | generated by the high-speed algorithm in the CSRBF method used by this invention. It is a figure which shows one specific example of the closed fold line calculated | required in the outline extraction stage of a wound in the restoration method of the data exchanged between CAD systems of 3rd Example of this invention. It is a figure which shows one specific example obtained in the stitching | fitting stage of the outline in the restoration method of the data exchanged between the CAD systems of 3rd Example of this invention. It is a figure which shows the repair result of the damage | wound on three-dimensional CAD data to which the repair method of the data exchanged between CAD systems of 3rd Example of this invention is applied. It is a figure which shows the repair result of the damage | wound on three-dimensional CAD data to which the partial procedure of the repair method of the data exchanged between CAD systems of 3rd Example of this invention is applied. It is a figure which shows the repair result of the damage | wound on the image of a texture pattern to which the restoration method of the data exchanged between CAD systems of 2nd Example of this invention is applied. It is a figure which shows the result by which the one part characteristic of the unnecessary three-dimensional model was removed by using the restoration method of the data exchanged between CAD systems of 3rd Example of this invention.

Claims (4)

A method of repairing data exchanged between CAD systems, which automatically repairs flaws present in 2D CAD data exchanged between different CAD systems,
Specifying a range of a wound portion WR on a two-dimensional image represented in the two-dimensional CAD data;
Designating a region W + in which the range-designated scratch portion WR is appropriately widened;
Obtaining function data using a CSRBF method for a region WL obtained by subtracting the scratched portion WR from the region W +;
Using the obtained function data, interpolating the scratched portion WR, and repairing the region W +;
A method for recovering data exchanged between CAD systems, comprising: A method of repairing data exchanged between CAD systems, which automatically repairs flaws present in 2D CAD data exchanged between different CAD systems,
A first step of specifying a range of a portion of a scratch on a two-dimensional image represented in the two-dimensional CAD data;
A second step of determining a predetermined point within the range-specified wound portion;
A third step of identifying a predetermined peripheral area portion adjacent to a predetermined point of the scratch portion determined in the second step;
A fourth step of obtaining function data using the CSRBF method for the predetermined peripheral region portion identified in the third step;
Using the function data obtained in the fourth step, a fifth step of interpolating a predetermined point of the flaw portion to repair a predetermined point of the flaw portion;
A sixth step of determining a next predetermined point of the wound portion;
A seventh step of identifying a predetermined peripheral area portion adjacent to a predetermined point next to the flaw portion determined in the sixth step;
An eighth step of obtaining function data using the CSRBF method for the predetermined peripheral region portion specified in the seventh step;
Using the function data obtained in the eighth step, a ninth step of interpolating a predetermined point next to the flawed portion and repairing a predetermined point next to the flawed portion; ,
Until the completion of the repair of all the scratches of the wound portion specified in the first step, the next predetermined point of the scratch portion repaired in the ninth step A tenth step as a predetermined point and returning to the sixth step;
Have
The predetermined peripheral area portion specified in the seventh step includes a predetermined point of a wound portion,
In the ninth step, when repairing a predetermined point next to the scratched portion, data after repairing the predetermined point of the scratched portion is used between CAD systems. How to recover the exchanged data. A method of repairing data exchanged between CAD systems, which automatically repairs flaws present in 3D CAD data exchanged between different CAD systems,
A wound contour extraction step of finding all boundary edges that can become boundaries at a polygonal scratch portion on the three-dimensional image represented in the three-dimensional CAD data, and forming a closed fold line from the obtained boundary edges; ,
Contour stitching step for performing repair by triangulating and dividing the polygonal flaw represented by the closed fold line;
A smoothing step for performing smoothing by reconstructing the surface with a predetermined algorithm for the scratched part by triangulation obtained in the stitching step of the contour, and
Have
The predetermined algorithm is:
For each triangle obtained in the stitching step of the contour, obtaining detailed point data inside the wound;
A region where a wound portion WR is appropriately widened is defined as W +, and a region obtained by subtracting the scratch portion WR from W + is defined as WL,
Obtaining function data for the region WL using the CSRBF method;
Using the determined function data, interpolating with detailed point data inside the scratch, and repairing the region WL; and
A method for recovering data exchanged between CAD systems, comprising: The basis function in the CSRBF method is

However, _{r (P} i, _{P j),} the distance r ₀ of any two points _{P i} and _{P j} in a plurality of discrete points, radius defined around the arbitrary point _{P i} which gives the initial value The function data at that time is

However, λ _i (i = 1, 2,..., N) is a coefficient of the basis function at the point P _i λ _{N + 1} , λ _{N + 2} , λ _{N + 3} , λ _{N + 4} is given by a function expressed by a coefficient of the first-order term,
The calculation of the coefficients λ _i , λ _{N + 1} , λ _{N + 2} , λ _{N + 3} , λ _{N + 4} is performed using a fast algorithm that diagonalizes the interpolation matrix,
The fast algorithm is
A first step in which one point is added to a list from the initial data including the plurality of discrete points, and the point added to the list is removed from the initial data;
A second step in which neighbor points for the points added in the first step are retrieved from the initial data and added to the list, and points added to the list are removed from the initial data;
A third step in which neighbor points for the points added in the second step are retrieved from the initial data and added to the list, and points added to the list are removed from the initial data;
When there are no more points to be added to the list, a fourth step in which one point from the discrete points remaining in the initial data is added to the list and the points added to the list are removed from the initial data;
The first step to the fourth step are repeated until there are no discrete points in the initial data, and a fifth step in which a diagonal matrix having a band property is created;
A method for recovering data exchanged between CAD systems according to any one of claims 1 to 3.

Images