Classifications

• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T19/00—Manipulating 3D models or images for computer graphics
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
  • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
  • G06V30/10—Character recognition
  • G06V30/19—Recognition using electronic means
  • G06V30/196—Recognition using electronic means using sequential comparisons of the image signals with a plurality of references
  • G06V30/1983—Syntactic or structural pattern recognition, e.g. symbolic string recognition
  • G06V30/1988—Graph matching

Definitions

• the present invention relates to the technical field of drafting and designing solid bodies by using computer aided drafting and computer aided design tools. Particularly, the present invention relates to a method for comparing a first computer aided 3D model with a second computer aided 3D model.
• the invention further relates to a computer program product with a computer- readable medium and a computer program, stored on the computer-readable medium with a program code, which is suitable for carrying out such a method when the computer program is run on a computer.
• SolidWorks Cooperation provides utilities as described at http : //www. solidworks . com/pages/products/solutions/docume nts/Utilities_FINAL.pdf .
• two solutions are offered.
• a feature based solution a comparison takes place between two versions of the same part. In this case, it has to be started from a common source.
• the solution is based upon using internal feature information, which restricts the solution to comparing two models produced with the SolidWorks tool .
• a second approach a geometry compare tool is used to compare imported bodies coming from different sources. This approach appears to be based upon comparing matching volumes and using this information to highlight differences on the body. Volume matching is prone to resolutions issues and feature recognition reliability.
• US Patent 6,625,607 discloses a method of comparing parts by comparing metadata representing the part topologies to identify identical or near identical parts. The comparison method takes place at a part level. It can distinguish that two parts are different, but can not distinguish at a detail level what those differences are.
• a computer-implemented method for comparing a first computer-aided 3D model with a second computer- aided 3D model the first model being described by a first topology comprising vertices, edges and faces of the first model and the second model being described by a second topology, comprising vertices, edges and faces of the second model, is disclosed herein, the method being executed by a scanning component, a vertex mapping component, an edge mapping component and a face mapping component and comprising the operations of: scanning the first topology and the second topology, deriving from the first topology a first topology map relating vertices, edges and faces of the first model and from the second topology a second topology map relating vertices, edges and faces of the second model, determining by means of the topological maps all possible mappings
• the method disclosed herein solves the geometrical problem, mentioned above, primarily in a non-geometric manner.
• the method addresses both, the issue of comparing any two arbitrary solid models, regardless of the system used to generate those models, provided they can be converted into a so-called boundary representation and of providing mapping knowledge necessary to migrate information between these two arbitrary models which may be of interest to a user.
• the aforementioned boundary representation is today the common denominator of nearly all of today's solid modelling solutions. For more details about solid modelling representation see, for example, the standard book "An Introduction to Solid Modeling" from Martii Mantyla .
• the topology of each model is first scanned.
• the topology represents generally a connectivity of each of the boundary elements, which make-up a solid model, i.e. a 3D model.
• solid model and “3D model” are used synonymously.
• This provides a connectivity network relating each face, edge and vertex of the modelled body.
• the method as disclosed herein, provides a mapping knowledge, necessary to migrate information between two arbitrary models, which is of interest to a user. This mapping knowledge is critical to successfully moving information from one arbitrary model to another and allowing the design to continue without loss of information.
• One further embodiment of the method according to the present invention comprises the step of producing a map, which relates all mapped faces, edges and vertices between the first and second models, the map being used by verifications routines. All mapping knowledge necessary to migrate information between the first and the second models is integrated.
• the possible mappings between the vertices of the first and the second models is performed by using coordinates as matching criteria.
• Still another embodiment of the method according to the present invention further comprises as a subsequent iteration the step of repeating all preceding steps with the first and second models being reversed. Since new faces may exist on the second model, which were never considered by a preceding mapping routine, the processing can be completed by reversing the mapping process.
• the present invention also refers to a computer program product with a computer-readable medium and a computer program stored on the computer-readable medium with a program code, which is suitable for carrying out a method according to the present invention, when the computer program is run on a computer.
• the present invention also relates to a computer program with a program code, which is suitable for carrying out a method according to the present invention, when the computer program is run on a computer.
• the present invention also refers to a computer-readable medium with a computer program stored thereon, the computer program comprising a program code, which is suitable for carrying out a method according to the present invention when the computer program is run on a computer.
• Figure 1 is a schematical view of a topological map of a cube as it can be used in an embodiment of the method according to the present invention.
• Figure 2 shows multiple vertices within a resolution region.
• Figure 1 shows a topological map of a cube as a simple example for a connectivity network as it can be used within an embodiment of the method according to the present invention.
• vl-v8 represents each of the 8 vertices of a cube
• el-el2 represent each of the 12 edges on the cube
• fl-f6 represent each face on the cube.
• Each arc segment in the above diagram represents the connectivity of these various elements in a cube.
• the object may be a cube, a rectangle or it may be any arbitrary 6 sided object bounded by 4 edges each with a total of 8 vertices.
• a first model and a second model source and target bodies
• anchor points the topology of the bodies of interest can be used to provide a mapping from one element to another on the two corresponding 3D models, i.e. solid models.
• the vertices of the model can be used to anchor the topological representation of the two models being compared for differences.
• At most one vertex from each body can share the same position. Once a source and target vertex is found sharing the same position in 3D-space, the vertex is considered mapped. After all possible vertex mappings have been identified; those mappings are then used to identify possible edge mappings. For each edge on a solid model, it may reference at most one or two vertices. If all vertices of an edge are mapped, then this edge can now be marked as mapped. For each face on a solid model, it may be bounded by one or more edges . Given one (special case condition) or more mapped edges, the corresponding face can be uniquely identified and therefore mapped.
• mapping information removes the requirement for iterating through the elements of the source and target models looking for geometrical sameness.
• All mapped faces found to be geometrically different are considered to be modifications between the source and target objects. All unmapped elements can be considered as representing the differences between the two models.
• the invention converts what has been traditionally an iterative process and establishes a methodology to convert the problem into a linear approach. This greatly reduces the computational effort required to resolve the problem. Furthermore, it solves an inherently geometrical problem by in fact performing a very limited number of geometrical checks. Combining these two benefits yields a solution which has the following benefits:
• any embodiment of the method according to the present invention operates equally well on manifold or non-manifold bodies, and in any combination.
• the method uses no intrinsic knowledge within the host system to resolve differences.
• the method according to the present invention is relatively insensitive to the geometric complexity of the objects being compared.
• Geometric perturbations which are introduced by translators can be resolved in a method so as not to interfere with this process.
• the method is inherently robust as only a portion of the topology needs to be mapped to resolve the differences between the source and target bodies, i.e. the first model and the second model.
• mapping information one can determine how to migrate information forward to the corresponding body following whatever rules are imposed by a host system.
• the first model is referred to as a source body and the second model is referred to as a target body.
• the four basic components of the method are described in the following. Each component successively deals with more specific elements of the methodology.
• the four components consist of:
• the first component is the starting point for the method.
• the first component primarily consists of the preprocessing and post-processing activities associated with the method.
• the pre-processing activities are described in steps 1-3 while the post processing activities are described in parts A to C.
• the source body i.e. the first model
• a topology map consisting of at least all faces, edges and vertices on the body is established. If desired, additional information such as loops and shells can also be collected at this time.
• the face structure is populated with the handle of the face along with the handles for all edges which bound this face. For performance reasons, all face structures should be collected into a data structure which allows for fast lookup which will be used heavily during the mapping process.
• a hash table was constructed to contain all face structures whose hash-key was the handle to the face.
• edge data structure Once the edge data structure has been constructed, all vertices of that edge are selected. The edge structure is populated with the handle of the edge along with the handles for all vertices which bound this edge. For performance reasons, all edge structures should be collected into a data structure which allows for fast lookup which will be used heavily during the mapping process. In the specific implementation, a hash table was constructed to contain all edge structures whose hash-key was the unique handle to the edge. If this edge had been previously scanned, this step can be skipped.
• the 3-space coordinate of this vertex is determined.
• the vertex structure is populated with the handle of the vertex along with the coordinate of this vertex.
• all vertex structures should be collected into a data structure which allows for fast lookup which will be used heavily during the mapping process.
• a hash table can be constructed to contain all vertex structures whose hash-key is the unique handle to the vertex. If this vertex has been previously scanned, this step can be skipped.
• the topological scanners described above could also include shell and loop information. This additional information is not required, but can improve the mapping process by providing additional information.
• Shells can be used to differentiate faces associated with one or more internal voids within a solid model.
• Loop information can be used to differentiate edges of a face which are representing one or more internal boundaries and the single external perimeter of a given face.
• the target body i.e. the second model
• a topology map consisting of all faces, edges and vertices on the body is established using the same procedure described above.
• Step 3 the source and target bodies are reversed (i.e. instead of mapping elements from the source body onto the target body, map the elements from the target body onto the source body) . Since new faces may exist on the target body which were never considered by the mapping routines, the processing is completed by reversing the mapping process. The second iteration through the mapping process should only consider those elements which remain unmapped. This greatly reduces the amount of possible elements for the mapping process to consider and therefore should operate in only a fraction of the time of the original iteration through Step 3.
• mapping process Once the mapping process has been completed, a map is produced which relates all known faces, edges and vertices between the source and target bodies. This information is then used by verifications routines to establish that the mapping process has indeed found a reasonable solution. This is done by performing geometrical verifications of the mapped faces. All faces on the source and target body are classified as one of three conditions:
• the specific implementation can use this information in various ways which may be useful for showing differences, reconciling non-geometrical information between the source and target bodies, or even possibly to migrate the geometrical differences forward or backward as suggested above.
• the approach provided by the present invention and described here takes a linear approach to the problem of finding differences between the source and target bodies. This avoids the highly iterative process of solely trying to map faces geometrically. Minimal geometries checks are done and only done when clarification is required to resolve ambiguities found during the mapping process. This creates a methodology which is very stable and quite insensitive to unstable geometric situations.
• Geometrical checks can be a tricky thing to do when trying to map corresponding elements. Edges and faces can be shortened, lengthened, or even change form from linear to circular or some other type. The fewer geometrical checks that are required to be performed during the mapping process, the faster the comparison can be made. On the other hand, some checking should be done so that proper anchoring of the topology can occur at all levels (not just at the vertex level) .
• the embodiment of the method according to the present invention which is described above is highly effective when a sufficient number of mapped vertices exist to anchor the edges of the compared bodies. However, when certain geometrical changes are made, there may be a very limited or no common vertices.
• loop information (if available) is used which would help to establish further mapping information.
• this information can be used to establish a high quality mapping even if the geometry of the edges should change.
• mapping vertices which represents one step in an embodiment of the method according to the invention, is based upon the simple process of matching coordinates between the target and source body. After the scanning step, i.e. the scanning process has been completed, this part of the process attempts to anchor the topology map so that it is possible to map elements on one body to another. Within the resolution of any given part, it should be impossible for two vertices to share the same coordinate. When comparing bodies of different resolutions, the lower of the two is used for matching coordinate information.
• the method for vertex mapping involves :
• one vertex within the resolution region R may be mapped to a specific vertex with a specified coordinate.
• the remaining vertices would remain unmapped. This condition would ultimately lead to these items being classified as differences which is exactly the solution desired.
• a lookup table is constructed which is indexed by the 3-space coordinate of the vertices. To further optimise the solution, as each vertex is matched, the matched vertex could be eliminated from the lookup table so that the further processing will be accelerated.
• the process of mapping edges representing a further step of an embodiment of the method according to the present invention uses the results of the vertex mapping procedure to establish the relationship between edges on the source and target bodies.
• This procedure's predictability depends on how aggressively one tries to resolve unmapped vertices using propagation rules.
• the methodology described here applies several propagation rules and iterates the entire map edge component until no additional mappings are resolved. Each specific implementation may attempt variations of what is described here.
• the edge mapping method involves:
• the first step of the method is to locate the corresponding edge on the target body using the mapped vertex information.
• One option is to resolve the ambiguity geometrically.
• Edges may not be in the same 3d space or they may have been trimmed/extended from their original configuration. Therefore, an alternative method to resolve the ambiguity would be to use additional topology information to resolve the situation. In this case, one can query all the faces associated with the ambiguous target edge. If all connected faces are mapped, then the ambiguous edge can be uniquely identified without any geometrical checks.
• the third step involved in the edge mapping component is that of mapping edges which have one unmapped vertex associated with it. Using basic topological rules, additional mappings may be found.
• Each unresolved edge is looped through and it is determined if it has at least one mapped vertex along with one unmapped vertex. If this situation exists, further edges may be resolved.
• the edge data structure is first populated with the mapping information and this edge is removed from the list of unresolved edges.
• steps 2 and 3 are repeated in their entirety. Additional iterations could resolve additional mappings since additional vertices may have been mapped. This iteration is continued until no further edges have been resolved.
• mapping extensions are available if the host system has the concept of shared geometry. It is common practice in many systems where different topological elements can refer to the same geometrical definition. This may occur, for example, when two linear edges are colinear and both edges refer to the same underlying curve definition. If this is the case, then additional mappings may be resolved. For each mapped edge, the corresponding curve definition can be identified. If multiple edges are found to refer to this one curve, these are established as secondary mappings. For each secondary mapping on the source body, the corresponding edge on the target body is identified. Once again, the corresponding curve is determined and all edges which refer to this curve are identified. With this additional information, secondary mappings between the source and target body can be further established.
• mapping faces representing a further step of an embodiment of the method according to the present invention uses the results of the edge mapping procedure to establish the relationship between faces on the source and target bodies.
• the predictability of this procedure depends on how aggressively one tries to resolve unmapped edges using propagation rules. The higher the quality of mapping edges, the greater the success in mapping faces.
• mapping faces consists of the following steps: 1. A map of edges is created, which provides a mapping of all available edges between the source and target bodies. This mapping information is used as the basis for the following steps described here.
• Face list List of edges bounding this face
• a single mapped edge is bound by up to two faces.
• the first step is to remove any mapped faces from the two possible candidates. Only unmapped faces are possible candidates. If there continues to be more than one possible solution, simple geometric checks could be used to identify the proper mapping.
• the first level check is to determine which of the two possible faces on the target body have the same geometric types (i.e. they are both planes) .
• the second level check is to determine if they share a common point on the surface .
• mappings may be resolved using an iterative approach described below. a) All unmapped faces on the source body are scanned though.
• Step 4 is repeated in its entirety until no further unmapped faces are resolved.
• mapping extensions are available if the host system has the concept of shared geometry. It is common practice in many systems where different topological elements can refer to the same geometrical definition. It was found that certain modelling operations can cause a single face to have a correspondence to one or more faces . Similarly, operations can cause multiple faces to be most accurately represented by a single face. Consider the case of a planar face being split into two faces by a groove. This leads to the concept that it may be desired to allow a single face to be mapped into multiple faces on the other body. To deal with this possibility, surface information is the key to resolve this condition. If topological elements share the same underlying geometry information, this information can be used to assist in the mapping processes.
• Secondary mappings For both face and edge data structures. This new data field allows information to be carried when multiple mappings are encountered. Before populating this field, the scanning routines need to look for topological elements which have shared geometric definitions. For example, there may exist the situation where two faces both refer to the same planar definition. There are two conditions which can cause the secondary mappings field to get populated. The first occurs when the process encounters a face or edge which shares its definition with another. The element that is currently being mapped is assigned as the primary mapping element. Other elements which share the same geometrical definitions are assigned as secondary mappings provided they have not been previously mapped. The second condition occurs when multiple topological elements are found to map onto a single geometrical element. This can occur when two faces on the source body are merged into a single face on the target body. Should this condition occur, the secondary mappings field is used to carry this information which can then be used for reconciliation.

Landscapes

• Engineering & Computer Science (AREA)
• Physics & Mathematics (AREA)
• General Physics & Mathematics (AREA)
• Theoretical Computer Science (AREA)
• Computer Vision & Pattern Recognition (AREA)
• Computer Graphics (AREA)
• Computer Hardware Design (AREA)
• General Engineering & Computer Science (AREA)
• Software Systems (AREA)
• Multimedia (AREA)
• Processing Or Creating Images (AREA)

Priority Applications (4)

Applications Claiming Priority (2)

Publications (1)

Family

ID=38068538

Family Applications (1)

Country Status (5)

Cited By (4)

Families Citing this family (11)

Citations (2)

Family Cites Families (4)

• 2007
  • 2007-02-28 WO PCT/EP2007/001694 patent/WO2007098929A1/en not_active Ceased
  • 2007-02-28 DE DE602007001275T patent/DE602007001275D1/de active Active
  • 2007-02-28 US US12/224,497 patent/US8422763B2/en not_active Expired - Fee Related
  • 2007-02-28 JP JP2008556708A patent/JP4969590B2/ja not_active Expired - Fee Related
  • 2007-02-28 EP EP07711706A patent/EP1989685B1/en not_active Ceased

Patent Citations (2)

Non-Patent Citations (4)

Also Published As

Similar Documents

Legal Events