JP2010211680A - Method of correcting model data - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To simply and efficiently correct model data to match it with measurement data. <P>SOLUTION: A die fabricated based on reference model data is corrected, and the corrected die is measured with a measuring instrument to obtain three-dimensional measured die data. Then, a measurement data surface 10 of the three-dimensional measured die data and a model surface 16 of the model data are placed in proximity to each other using a computer. A normal line is determined from a vertex 24 of a polygon set on the model surface 16 (S102). A first intersection between the normal line and the measurement data surface 10 is determined (S103). The vertex 24 is moved along the normal line to a predetermined ratio position to the first intersection so as to obtain a movement correction surface (steps S104, S105). Similar processing is repetitively performed more than once (S106 to S116). An optimization processing is performed to an upper layer surface obtained by projection processing (S117). <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention corrects a die or an actual model created based on reference model data, and further obtains three-dimensional measurement data by measuring with a measuring instrument, and then displays the three-dimensional measurement data by a computer. The present invention relates to a method for correcting model data that is compared with a first surface that is arranged close to a second surface indicated by model data.

  Conventionally, when manufacturing a mold for press molding, a mold is designed by using CAD or the like from shape data of a molded product, and mold data is generated. An NC program for machining a die is created based on the obtained die data, and the die as the first stage is machined using an NC machine tool. The mold at this stage does not necessarily produce a desired product, and verification based on a molded product actually obtained through trial use is generally performed to correct the mold. It has been broken.

  In press-molding of complex shapes such as automobiles, clearance between the upper die and lower die, which cannot be understood from the prototype or molding simulation, may occur, and the prototype may be wrinkled or cracked. Therefore, it is necessary to perform a repeated process of making a prototype again after correcting the mold.

  The final mold, that is, the first mold is only one, but after making the door mold on one side of the car, the other door has the same shape or the same at multiple production sites When manufacturing a product, it is advisable to provide one or more second molds that are the same (or symmetrical) as the first mold.

  In order to shorten the time required for manufacturing the second mold, it is conceivable to measure the shape of the mold after correction three-dimensionally and reflect the obtained three-dimensional measurement data in the mold model data. The present applicant has proposed a means for reflecting the three-dimensional measurement data in the mold model data in Patent Document 1 as described above. That is, the surface indicated by the mold three-dimensional measurement data is arranged close to the surface indicated by the die model data, and the absolute value of the distance between a plurality of pairs of corresponding points is calculated between them, and the absolute value of the distance is calculated. Correct the mold model data. According to this method, CAD data of a smooth surface can be obtained, and the corresponding points between the surfaces can be prevented from being twisted.

  In the method described in Patent Document 1, the reference points of a plurality of polygons on the surface indicated by the mold three-dimensional measurement data and the corresponding points on the surface indicated by the corresponding mold model data are defined.

  On the other hand, even when designing the appearance of a vehicle, a designer may prepare model data at any stage and the clay model (so-called mock-up) created based on the model data may be modified several times. Even in such a case, there is a desire to reflect the corrected clay model in the model data.

JP 2008-176441 A

  In the method described in Patent Document 1, in order to define the reference point on the surface indicated by the mold 3D measurement data and the corresponding point on the surface of the mold model data, the reference point on the surface of the mold 3D measurement data is used. A normal is set for it. Since the three-dimensional measurement data of the mold is obtained by measuring the first mold of the actual product, it has a slightly rough surface due to the influence of minute processing marks and measurement errors caused by the measuring instrument. Therefore, it is desirable not to set the normal line as it is from the reference point, but to set the normal line after performing a predetermined smoothing process (relaxation smoothing process or the like) once. However, such a smoothing process is complicated and takes a long time.

  The present invention has been made in consideration of such problems, and based on the measurement data obtained by measuring the actual mold and the actual model corrected manually, it corresponds to the actual product before correction. It is an object of the present invention to provide a model data correction method capable of easily and efficiently correcting model data obtained at the beginning so as to conform to measurement data.

  The model data correction method according to the present invention corrects a mold manufactured based on model data serving as a reference, and measures the corrected mold with a measuring instrument to obtain mold three-dimensional measurement data. Then, the computer projects the mold three-dimensional measurement data and the model data close to each other, and projects the first surface indicated by the model data onto the second surface indicated by the mold three-dimensional measurement data. And the projecting step includes a first step of obtaining an average normal obtained by including a normal or a peripheral portion thereof from a plurality of reference points set on the first surface as base points, and the method A second step for obtaining an intersection between the line or the average normal line and the second surface, and moving the plurality of reference points along the normal line or the average normal line to a predetermined percentage position up to the intersection point, respectively. Let And having a third step of obtaining a movement balancing plane.

  Further, the present invention corrects the actual model produced based on the model data serving as a reference, obtains the actual model three-dimensional measurement data by measuring the corrected actual model with a measuring instrument, and then uses a computer. A projecting step of arranging the actual model measurement data and the model data close to each other, and projecting a first surface indicated by the model data onto a second surface indicated by the actual model measurement data; A first step for obtaining an average normal obtained from a plurality of reference points set in the first surface as a base point and including a normal or a peripheral portion thereof, and the normal or the average normal and the first A second step of obtaining intersections with the two surfaces, respectively, and moving the plurality of reference points along the normal line or the average normal line to a predetermined ratio position to the intersection point, thereby moving the movement correction surface. And having a third step that.

  Thus, according to the projection process in which a normal line or an average normal line is provided from a plurality of reference points set on the first surface and the reference point is moved along the normal line, the mold three-dimensional measurement data surface and the model No special smoothing is required for any of the data. Therefore, it is possible to easily and efficiently correct the model data so as to match the measurement data. Here, the predetermined ratio means including 100%.

  The movement correction surface may be updated as the first surface, and the first to third steps may be repeated a plurality of times.

  The reference point is a vertex of a polygon constituting the first surface, and the average normal vector is a vector obtained by weighted averaging of normals at the vertexes of the polygon within a predetermined range around the reference point including the reference point. There may be.

  After the projecting step, there may be an optimization step of creating a mesh based on the pseudo curved surface so that the finally formed movement correction surface meets a predetermined accuracy condition.

  According to the method for correcting model data according to the present invention, it is possible to easily and efficiently correct the model data originally obtained corresponding to the actual product before correction so as to match the measurement data.

It is a flowchart which shows the procedure of the correction method of the model data which concerns on this Embodiment. It is a schematic diagram which shows a mode that a normal line is set with respect to a model surface. It is a flowchart (the 1) which shows the procedure of lamination deformation. It is a flowchart (the 2) which shows the procedure of lamination deformation. It is a schematic diagram which shows a mode that the point within two nodes is extracted from a predetermined | prescribed division | segmentation point. It is a figure which shows a weighting function. It is a schematic diagram which shows a mode that a normal line is set from the 1st layer surface. It is a figure which shows two-dimensionally the movement correction surface of multiple times by lamination | stacking deformation | transformation typically. It is a figure which shows the movement correction surface of multiple times by lamination | stacking deformation | transformation typically three-dimensionally. It is a schematic diagram of the example which a twist generate | occur | produces in a normal line between surfaces. It is a schematic diagram which shows the mode of an optimization process. It is a flowchart which shows the procedure of the correction method of the model data which concerns on a modification.

  Hereinafter, a model data correction method according to the present invention will be described with reference to FIGS.

  First, in step S1 of FIG. 1, the molded product to be obtained is designed, and data of the molded product model is created.

  In step S2, mold model data data is created by CAD based on the molded product model data.

  In step S3, NC data for the NC processing machine is created based on the data of the mold model data.

  In step S4, using the obtained NC data, a mold is formed by an NC processing machine.

  In step S5, press molding is performed using the molded mold to obtain a molded product as a prototype.

  In step S6, the prototype and the die press surface are observed and examined, and the die is corrected manually. At this time, the prototype is observed and examined for wrinkles and cracks, dimensional errors, etc., and the mold is corrected by comprehensively considering the condition of the surface of the press. These steps S5 and S6 may be repeated a plurality of times.

  In step S7, the corrected mold is measured three-dimensionally with a measuring instrument (such as a three-dimensional digitizer) to obtain three-dimensional measurement data composed of point groups. This measuring instrument may be a contact type or a non-contact type.

  In step S8, the point group of the three-dimensional measurement data is set to a large number of polygons by a predetermined means using a computer. These polygons indicate the measured mold surface shape. The polygon is mainly represented by a triangular plane.

  Further, the computer compares the polygonalized three-dimensional measurement data with the mold model data, and, as shown in FIG. 2, a measurement data surface (second surface) indicated by a polygon based on the mold three-dimensional measurement data. ) 10 is arranged sufficiently close to the model surface (first surface) 16 indicated by the mold model data. In this close placement processing, for example, it is preferable that the entire surface is sufficiently approached so that the average of the distance between the measurement data surface and the model surface is substantially minimized. The measurement data surface and the model surface may partially intersect. By making this close arrangement, the surfaces of the mold that are not corrected are substantially in surface contact with each other.

  As shown in FIG. 2, the measurement data plane 10 is an aggregate of polygons 22 having a large number of measurement points 18 as vertices. Since the measurement data surface 10 is obtained by measuring the first mold of the actual product, the measurement data surface 10 is a slightly rough surface due to the influence of minute processing marks and measurement errors caused by the measuring instrument.

  The model surface 16 is also composed of a large number of polygons 22. In FIG. 2 and the subsequent drawings corresponding thereto, the planar measurement data surface 10 and the model surface 16 are schematically shown as lines.

  In step S9, the distance between the measurement data surface and the model surface is determined at a plurality of correction points.

  In step S10, the difference between the measurement data plane and the model plane at a plurality of reference points is determined, and a range to be corrected is cut out. This range to be corrected is automatically performed by the computer.

  In step S11, a stacking deformation process is performed. The lamination deformation process will be described later.

  In step S12, based on the absolute value of the distance from each measurement point of the obtained mold three-dimensional measurement data to the mold model, that is, the difference data, the mold model is deformed to produce a mold correction model. Because this process transforms the mold model data based on the difference data, model data is created that inherits the adjacent information and curves of the original data, so even if there are missing points at the measurement points, It is possible to easily restore and restore model data corresponding to the shape of the model.

  The mold correction model obtained in this way considerably reflects information on the shape of the mold corrected in step S6 based on a prototype obtained by actually making a prototype at least once. Therefore, the number of man-hours required for correction when manufacturing a repeat mold is greatly reduced. That is, NC data is created based on the mold correction model, and the repeat type created by the NC machine tool based on the NC data reflects the correction made in step S6. Therefore, a highly accurate molded product can be formed.

  Next, the lamination deformation process in step S11 will be described. This process is called a lamination deformation process because three intermediate layers are set on the original measurement data plane 10 while being laminated and deformed.

  First, in step S101 of FIG. 3, a reference point for stacking deformation on the model surface 16 is set. Here, the vertex 24 of the polygon 22 is used as a reference point as shown in FIG.

  In step S102, a normal vector line 26 with respect to the measurement data surface 10 is set from each vertex 24 of the model surface 16. That is, the normal vector line 26 is set so that the angle δ formed by the normal vector line 26 and each surrounding model surface 16 becomes equal.

  Since the vertex 24 is the vertex of three or more polygons 22, the normal vector line 26 may be set so that the angles with respect to the adjacent polygons 22 are as equal as possible.

  Further, in order to further improve the accuracy, the vector line 26 may be obtained by a weighted average of predetermined surrounding areas.

  That is, as shown in FIG. 5, a point 28b of one ball node and a point 28c of two ball nodes are extracted with respect to the reference point 28a. One ball node is a point connected by a single line to the point 28a, and is indicated by a black circle in FIG. A 2-ball node is a point connected by two or less lines, and is indicated by a white circle in FIG. In the example shown in FIG. 5, the point 28b of 1 ball node is 8 points, and the point 28c of 2 ball node is 11 points. Therefore, there are 19 points within 2 ball nodes.

Thereby allowing the identification of the corresponding point vector 34 as a point n _j and numbered j (j = 1~19) for each point within two ball node, linear distance d _j from the point 28a of each point n _j Ask for.

Based on the following equation (1), a point vector n ′ _j is obtained by weighting the vector n _j of points within 2 ball nodes according to the distance d _j and performing weighted averaging.

Here, the parameter m is the total number of points within 2 ball nodes, and m = 19 in the example of FIG. The function f to the distance d _i and the argument is the weighting function represented in Figure 6. The function f, when the absolute value of the distance d _i is equal to or smaller than the threshold d _Max is defined by the function g, 0 and becomes when the absolute value exceeds the threshold value d _Max. The function g is a function in a range of 0 ≦ g ≦ 1 indicating a substantially normal distribution, and is defined as g = 0 when | d _j | = d _Max and g = 1 when d _j = 0. In FIG. 6, the positive range and the negative range of the distance d _i represent the front side and the back side of the target surface separately.

In the point representative vector n ′ obtained by the equation (1), a vector corresponding to a point of 3 ball nodes or more, or a point corresponding to a point where the distance d _j is too large is removed. Weighted according to d _j and averaged. Therefore, the influence of the short-distance vector increases, and a point representative vector n ′ that appropriately indicates the peripheral shape is obtained.

  Next, in step S103, a first intersection point 38, which is each intersection of the line 26 and the measurement data surface 10, is obtained, and a distance from the vertex 24 to the first intersection point 38 is obtained.

  In step S104, the vertex 24 and the first intersection point 38 are divided into, for example, four equal parts, and the first division point 40 closest to the vertex 24 is obtained. In other words, the first division point 40 is a division point obtained by dividing the measurement point 18 and the first intersection 38 at a ratio of 1: 3. The number of divisions may be one or more (that is, the division ratio of division points may be 100%).

  In step S105, while maintaining the polygon connection relation based on the initial measurement point 18, as shown in FIG. 7, a polygon is set for each corresponding first division point 40, and the first layer surface (movement correction) is performed. Surface) 42 is formed. That is, the plurality of vertices 24 are respectively moved along the line 26 to the first division point 40 which is a predetermined ratio position to the first intersection 38 to obtain a movement correction plane.

  In steps S101 to S105, the smoothing process is not particularly necessary for the measurement data plane 10 and the model plane 16, and the original polygon plane can be applied as it is, and high-speed processing is possible.

  In step S106, as shown in FIG. 7, a line 44, which is a weighted average normal line, is set from each first division point 40 to the measurement data plane 10. This is the same processing as in step S101 described above, and is equivalent to updating the first layer surface 42 obtained as the movement correction surface as the original model surface 16.

  In step S107, the second intersection point 46, which is the intersection point between the line 44 and the model surface 16, is obtained, and the distance from the first division point 40 to the second intersection point 46 is obtained. This is the same processing as step S102 described above.

  In step S <b> 108, the first division point 40 and the second intersection 46 are divided into three equal parts, and the second division point 48 closest to the first division point 40 is obtained. In other words, the second division point 48 is a division point obtained by dividing the first division point 40 and the second intersection 46 at a ratio of 1: 2.

    In step S109, while maintaining the polygon connection relation based on the initial measurement point 18, a polygon is set for the obtained second division point 48 to form the second layer surface 49.

  Thereafter, as shown in FIG. 4, in the same manner, a normal vector of the polygon is set from the second division point 48 (step S110), the third intersection point is obtained (step S111), and the second division point is obtained. 48 and the third intersection point are divided into two equal parts, a third division point is obtained (step S112), a polygon is set for the third division point, and a third layer surface 56 (see FIG. 9) is formed. (Step S113).

  Further, a normal vector of the polygon is set from the third division point (step S114), and a corresponding point 50 (see FIG. 8) that is an intersection of the normal vector and the measurement data plane 10 is obtained (step S115). A polygon is set for the point 50 to set the upper layer surface 58 (step S116).

  The processes so far are collectively shown in FIGS. It will be understood that the original model surface 16 is projected onto the measurement data surface 10 through four stages. In such layer conversion processing, projection is performed stepwise by providing movement correction surfaces at predetermined ratios instead of performing projection at once along the line 26 that is the original normal line. Therefore, even when the plurality of lines 26 intersect each other at a portion having a large curvature on the measurement data surface 10 and the model surface 16, the positional relationship between the plurality of polygons 22 on the original model surface 16 is maintained as it is. And projected onto the measurement data plane 10.

  If the lamination deformation process is not performed, a surface straight line 52 is set from the measurement point 18 to the measurement data surface 10 in a portion where the radius of curvature of the measurement data surface 10 or the model surface 16 is small as shown in FIG. However, the relationship between the obtained corresponding point 54 and the measurement point 18 may be twisted, and the mold correction model cannot be set with high accuracy. On the other hand, in the present embodiment, there is no such inconvenience due to the stacking deformation process, and the corresponding point 50 is set on the measurement data surface 10 while the mutual positional relationship between the plurality of measurement points 18 on the measurement data surface 10 is substantially maintained. As a result, an appropriate correspondence can be obtained.

  Next, in step S117, as shown in FIG. 11, the finally formed upper layer surface 58 is optimized so as to meet a predetermined accuracy condition (for example, the tolerance tr is reduced according to the specified value MT). In this optimization process, an appropriately smooth pseudo-curved surface 59 is set for a location that does not meet the accuracy condition, and an appropriate pitch is recalculated based on the pseudo-curved surface 59 and then the mesh is reconstructed. Good. Further, the obtained mesh surface may be reprojected to the measurement data. Data that has been optimized and guaranteed accuracy in this way can be used as CAM data for die machining.

  2, 7, and 8, the measurement data surface 10 is provided only on one side with respect to the model surface 16, but the measurement data surface 10 may be provided on the opposite side of the model surface 16. It may be well or partially crossed. In the above-described lamination deformation process, an example in which three layers of intermediate surfaces are provided is shown, but two or more layers may be provided. The division ratio serving as a reference for the division point determined in the middle can be arbitrarily set. For example, a location that is always the middle point (1: 1) may be set as the division point.

  As described above, according to the model data correction method according to the present embodiment, no special smoothing process is required for either the measurement data plane 10 or the model plane 16 in the projection process (steps S101 to S116). Therefore, the model surface 16 can be easily and efficiently corrected so as to match the measurement data surface 10. According to the result of the trial by the present inventor, the model data correction method according to the present embodiment has a predetermined size compared to the method of correcting a surface while performing smoothing according to the procedure described in Patent Document 1. When applied to the metal mold, the processing time was shortened to about 1/6 and the accuracy was maintained as before.

  Data corrected in this way can also be used for FEM analysis.

  Next, a method for applying the present invention to the vehicle exterior design stage will be described. That is, at the time of appearance design, model data is prepared at any stage, and the designer may modify the clay model created based on the model data several times. Even in such a case, the corrected clay model can be reflected in the model data.

  In step S201 in FIG. 12, the designer designs the appearance of the vehicle in a virtual space using a computer. After several reviews, the first stage design is decided. This appearance design is recorded as model data. Modern computers have high processing capabilities and can easily and rapidly perform three-dimensional design.

  The clay model obtained in this way has a fairly sophisticated design. However, in the design on the computer, it is only seen through a monitor or printed matter, and three-dimensional examination is indispensable, so the following processing is performed.

  In step S202, a clay model (actual model) is produced based on the model data.

  In step S203, the clay model is observed, and a three-dimensional examination of the external design is performed to make a predetermined correction. This correction is performed manually by a designer or a predetermined operator. These steps S202 and S203 may be repeated a plurality of times. The clay model may be initially a small model and then a full size model may be created.

  In step S204, the shape of the modified clay model is measured three-dimensionally with a measuring instrument, and three-dimensional measurement data composed of point groups is obtained. This is substantially the same as the processing in step S7 described above, except that the target is not a mold but an actual model.

  Subsequent steps S205 to S209 are the same processing as steps S8 to S11 (see FIG. 1). Therefore, in the stacking deformation in step S209, the same processing as in FIGS. 3 and 4 is performed.

  The data thus obtained can be used as mold model data when the mold shown in FIG. 1 is created. It can also be used when a clay model is created again for some purpose. It may be used for FEM analysis.

  The model data correction method described above is not limited to a method for a vehicle body, but can be applied to smaller products.

  The model data correction method according to the present invention is not limited to the above-described embodiment, and it is needless to say that various configurations can be adopted without departing from the gist of the present invention.

DESCRIPTION OF SYMBOLS 10 ... Measurement data surface 16 ... Model surface 18 ... Measurement point 22 ... Polygon 24 ... Vertex 38 ... 1st time intersection 42 ... 1st layer surface 49 ... 2nd layer surface 56 ... 3rd layer surface 58 ... Upper layer surface

Claims (5)

The mold manufactured based on the model data as a reference is corrected, and the corrected mold is measured by a measuring instrument to obtain the mold three-dimensional measurement data, and then the mold is measured by the computer. A projecting step of closely arranging data and the model data, and projecting a first surface indicated by the model data onto a second surface indicated by the mold three-dimensional measurement data;
The projecting step includes
A first step of obtaining an average normal obtained by including a normal line or its peripheral part, with a plurality of reference points set in the first surface as base points; and
A second step for determining an intersection between the normal line or the average normal line and the second surface;
A third step of obtaining a movement correction surface by moving a plurality of the reference points along the normal line or the average normal line to a predetermined ratio position to the intersection point;
A model data correction method characterized by comprising: The actual model produced based on the reference model data is corrected, and the corrected actual model is measured by a measuring instrument to obtain actual model three-dimensional measurement data. A projecting step of closely arranging the model data and projecting a first surface indicated by the model data onto a second surface indicated by the actual model measurement data;
The projecting step includes
A first step of obtaining an average normal obtained by including a normal line or its peripheral part, with a plurality of reference points set in the first surface as base points; and
A second step for determining an intersection between the normal line or the average normal line and the second surface;
A third step of obtaining a movement correction surface by moving a plurality of the reference points along the normal line or the average normal line to a predetermined ratio position to the intersection point;
A model data correction method characterized by comprising: The method for correcting model data according to claim 1 or 2,
Updating the movement correction surface as the first surface;
The method of correcting model data, wherein the first step to the third step are repeatedly executed a plurality of times. In the correction method of the model data of any one of Claims 1-3,
The reference point is a vertex of a polygon constituting the first surface,
The model data correcting method, wherein the average normal vector is a vector obtained by weighted averaging of normals at the vertices of polygons within a predetermined range around the reference point including the reference point. In the correction method of the model data according to any one of claims 1 to 4,
After the projecting step, a model data correcting method comprising an optimization step of creating a mesh based on a pseudo curved surface so that the finally formed movement correction surface meets a predetermined accuracy condition .

Images