3D Bounding Box Partitioning for Faster Spatial Queries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for indexing and querying spatial data in complex product structures, such as those used in CAD tools, suffer from large storage requirements and slow query times due to the need to store full permutations of geometric bounds, leading to numerous false positives and inefficient caching solutions.
Innovation Solution
A method involving partitioning and merging axis-aligned bounding boxes along the x, y, and z axes, using partition identifications to group and merge boxes based on geometric bounds, reducing the need to store full permutations and minimizing false positives while maintaining efficient storage and query times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full permutations of geometric bounds are stored to accurately bound each permutation of item positions, then spatial query accuracy is improved, but storage requirements increase significantly
Solution Approach 1:
The patent segments the set of bounding boxes by partitioning them using intervals along the x, y, and z axes, creating partition identification tuples (xpar, ypar, zpar) that group bounding boxes into manageable subsets. This segmentation allows the system to process and store spatial data in a structured manner without requiring full permutation storage, thereby reducing storage requirements while maintaining query accuracy.
Solution Approach 2:
The patent merges bounding boxes that share the same partition identification tuple by computing a unified geometric bound that encompasses all boxes in the group. This merging operation reduces the total number of stored bounding boxes while ensuring that the merged bound accurately represents the spatial extent of all original boxes, thus decreasing storage requirements without sacrificing spatial query precision.
2Measurement precision
If full permutations of geometric bounds are stored to ensure accurate spatial queries, then query accuracy is improved, but query time increases due to data volume
Solution Approach 1:
By segmenting bounding boxes into partition groups using interval-based partitioning along three axes, the patent enables queries to operate on smaller, organized subsets of data rather than scanning full permutations. This segmentation reduces the computational burden during query operations, thereby decreasing query time while preserving the accuracy needed for precise spatial searches.
Solution Approach 2:
The patent merges bounding boxes within each partition group to create consolidated geometric bounds. This merging reduces the total number of bounding box evaluations required during spatial queries, directly decreasing query time. The merged bounds maintain the necessary geometric accuracy to ensure that spatial queries return correct results without requiring examination of every individual bounding box permutation.
3Device complexity
If unconfigured product structure index is used to simplify indexing, then device complexity is reduced, but false positives increase due to multiple part positions
Solution Approach 1:
The patent applies interval-based partitioning to segment the unconfigured product structure index into discrete partition groups along the x, y, and z axes. This segmentation organizes the simplified index structure into manageable segments, allowing the system to maintain low indexing complexity while reducing false positives through structured organization of spatial data.
Solution Approach 2:
The patent merges bounding boxes within each partition group to create consolidated geometric bounds that accurately represent the spatial extent of multiple part positions. This merging operation reduces false positives by ensuring that queries against the unconfigured index evaluate against accurate merged bounds rather than individual scattered boxes, thereby improving reliability while maintaining the simplicity of the unconfigured index approach.
Data Source
AI summary
A computer implemented method of bounding spatial data associated with the geometric bounds of an item mapped into one or more 3-D axis-aligned bounding boxes is disclosed. The geometric bounds bound each permutation of all possible positions of the item geometrically. The method comprises firstly partitioning a set of bounding boxes using a first group of intervals along the x axis direction and allocating a partition identification xpar. Secondly partitioning the set of bounding boxes using a second group of intervals along the y axis direction and allocating a partition identification ypar is carried out. Thirdly, partitioning the set of bounding boxes using a third group of intervals along the z axis direction and allocating a partition identification zpar is carried out. The set of bounding boxes is then partitioned by partition identification tuples (xpar, ypar, zpar). Preferably bounding boxes with the same partition identification tuple are then merged.