Feature Tree Extraction from Mesh Data via Geometric Graph Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for extracting a feature tree from a mesh are inefficient, as they often require an exact B-Rep model as input, which can be difficult to compute and resource-intensive, and may fail with incomplete or complex mesh data, leading to inaccurate or invalid results.
Innovation Solution
A computer-implemented method that computes a geometric and adjacency graph from the mesh, instantiates surfaces based on primitive types and parameters, applies sweep operations to generate solids, and computes combinations of add or remove operations to identify the most similar solid, thereby constructing a feature tree directly from the mesh without the need for an exact B-Rep model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If an exact B-Rep model is used as input for feature tree extraction, then the extraction accuracy is improved, but the computational complexity and resource requirements increase significantly
Solution Approach 1:
The patent extracts only the essential geometric and topological features from the mesh data needed for feature tree construction, rather than requiring a complete exact B-Rep model. This selective extraction reduces computational complexity while maintaining sufficient accuracy for CAD operations.
Solution Approach 2:
The mesh data is segmented into meaningful regions and features that can be processed independently to construct the feature tree. This segmentation allows the algorithm to work with smaller, more manageable data subsets rather than the entire complex B-Rep model.
2Measurement precision
If an exact B-Rep model is computed from mesh data, then the feature tree extraction accuracy is improved, but the processing time and computational resources increase
Solution Approach 1:
The patent performs preliminary analysis of the mesh data to identify key geometric features and their relationships before constructing the feature tree. This preliminary action extracts essential information directly from the mesh without requiring time-consuming B-Rep computation.
Solution Approach 2:
The algorithm performs partial computation by focusing only on the portions of the mesh data that are necessary for feature tree extraction, rather than processing the entire dataset required for complete B-Rep modeling. This selective processing reduces time loss while maintaining extraction accuracy.
3Measurement precision
If traditional methods are used to extract feature trees from meshes, then the results are accurate, but the methods fail with incomplete or complex mesh data
Solution Approach 1:
The patent changes the input parameters from requiring exact B-Rep models to accepting raw mesh data with varying quality. The algorithm adapts its processing based on the actual mesh characteristics, making it robust to incomplete or complex geometries while maintaining extraction accuracy.
Solution Approach 2:
The patent introduces an intermediary processing stage that bridges raw mesh data and feature tree extraction. This intermediary layer handles the conversion and validation of mesh data, ensuring reliable extraction even when the input mesh is incomplete or complex, without requiring perfect B-Rep input.
4Manufacturing precision
If complete B-Rep reconstruction is performed, then valid feature limits are recovered, but the computational resources and time required increase
Solution Approach 1:
The patent extracts only the critical feature limit information directly from mesh analysis, rather than performing complete B-Rep reconstruction. This selective extraction recovers necessary manufacturing precision information while significantly reducing computational resource consumption.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
The invention notably relates to a computer-implemented method for extracting a feature tree from a mesh. The method comprises providing a mesh, computing 110) a geometric and adjacency graph of the provided mesh, wherein each node of the graph represents one region of the mesh and comprises a primitive type and parameters of the region; each connection between two nodes is an intersection between the respective surfaces of the regions represented by the two connected nodes. The method also comprises instantiating for each node of the graph, a surface based on the identified primitive type and parameters of the region. The method further comprises for each node of the graph: computing one outer intersection loop and possibly one or more inner intersection loops, each intersection loop being obtained from an ordered list of nodes of the graph that are connected to the said each node; computing at least one sketch from the outer intersection loop and possibly one or more sketches respectively from possibly one or more inner intersection loops of the instantiated surface. The method also comprises applying a sweep operation to each computed sketch therefore generating a solid for each sketch, computing all possible combinations of add or remove operations between the obtained solids, and obtaining a final solid for each computed combination, computing a similarity score of each final solid with the input mesh, and identifying the most similar final solid with the input mesh based on the similarity score, the combinations of add or remove operations of the identified most similar final solid corresponding to the feature tree of the mesh.