Convex Hull Filtration for Internal Workpiece Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for filtering measurement data sets to verify internal features of workpieces, such as bores, lack a unified filtration approach that can effectively eliminate outliers and smooth deviations, unlike the convex hull-based method used for external features.
Innovation Solution
A method involving mirroring measurement points of an internal feature onto an auxiliary feature's boundary, determining a convex hull, projecting onto it, and mirroring back to generate a filtered data set, allowing for the application of a convex hull-based filtration method to both internal and external features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a convex hull based filtration method is used for outer features, then outliers are eliminated and outermost points are kept, but this method cannot be directly applied to inner features
Solution Approach 1:
The patent applies inversion by mirroring measurement points of inner features onto an auxiliary feature's boundary, transforming them into outer feature points. This allows the convex hull method (designed for outer features) to be applied to inner features by inverting the problem space, then mirroring back the filtered results to obtain the final filtered inner feature points.
Solution Approach 2:
The patent introduces an auxiliary feature as an intermediary element. Measurement points of the inner feature are mirrored onto the auxiliary feature's boundary, which serves as a mediator that enables the application of convex hull filtration. The auxiliary feature acts as a temporary reference frame that facilitates the transformation and filtering process.
2Manufacturing precision
If different filter types are used for inner and outer features, then each feature type can be filtered appropriately, but the filtration process becomes complex and non-unified
Solution Approach 1:
The patent achieves universality by creating a single unified filtration approach that works for both inner and outer features. By mirroring inner feature points onto an auxiliary boundary and applying convex hull filtration, the same mathematical method (convex hull) serves multiple purposes - filtering both inner and outer features through a unified process rather than requiring separate filter types.
Solution Approach 2:
The inversion technique transforms the inner feature filtering problem into an outer feature problem, allowing the standard convex hull method to be universally applied. This eliminates the need for completely different filter types for inner versus outer features, simplifying the overall filtration process while maintaining effectiveness.
3Reliability
If morphological filters are used for inner features, then outliers can be eliminated, but the filter does not preserve innermost points as effectively as convex hull for outer features
Solution Approach 1:
By mirroring inner feature points onto the auxiliary boundary (inverting the problem), the patent transforms the filtering task into one where convex hull naturally preserves the most relevant points (analogous to outermost points in outer feature filtering). When mirrored back, these correspond to the innermost points of the original inner feature, thus preserving them effectively while eliminating outliers.
Data Source
Figure 1a~1c
Figure 2
Figure 3a~3b
AI summary
The present invention relates to a method for filtering a measurement data set which is usable for specifying and/or verifying an internal feature (10) of a workpiece (100), the method comprising: - providing a measurement data set comprising a plurality of measurement points (P) of the internal feature (10); - providing an auxiliary feature (30) which represents an ideal estimate for the internal feature (10) of the workpiece (100); - mirroring each measurement point (P) of the measurement data set on a boundary element of the auxiliary feature, thereby generating a first modified data set comprising a plurality of first modified measurement points (P'); - determining a convex hull (H) of the first modified measurement points (P') and projecting the first modified measurement points (P') onto the determined convex hull (H), thereby generating a second modified data set comprising a plurality of second modified measurement points (P"); and - mirroring each second modified measurement point (P") on the boundary element of the auxiliary feature (30), thereby generating a filtered measurement data set comprising a plurality of filtered measurement points (P'").