Range Query Processing Complex Geometric Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for processing range queries on complex geometric data are computationally expensive and inefficient, often returning many irrelevant objects due to the high complexity of similarity metrics, especially when dealing with spatial trajectories.
Innovation Solution
A method that involves creating a buffered query object with an increased geometric footprint based on the query distance, generating a buffered query filter using a filter function that maps overlapped grid cells to a corresponding element, and filtering out complex geometric data objects that are not a subset of or have an empty intersection with this filter, allowing for efficient pruning of irrelevant data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If MBR (minimum bounding rectangle) intersection is used for pruning, then the amount of relevant objects to database query is reduced, but many dissimilar objects are still returned that share spatial similarity but are not ultimately relevant to the search query
Solution Approach 1:
The patent segments the complex geometric objects into multiple simple geometric components (e.g., trajectories divided into line segments, polygons into edges). This segmentation allows the use of simple MBR intersection tests on individual components rather than requiring complex full-object similarity calculations, thereby maintaining pruning efficiency while reducing false positives from dissimilar objects that share overall spatial proximity.
Solution Approach 2:
The patent introduces an additional dimension of analysis by examining geometric similarity at the component level (individual line segments, edges, or points) rather than only at the object level. This dimensional shift from whole-object MBR comparison to component-level geometric comparison enables more precise filtering that distinguishes between objects that are merely spatially proximate versus those that are genuinely similar in shape and structure.
2Measurement precision
If direct comparison of complex geometric objects using similarity metrics is performed, then accurate similarity measurement is achieved, but computational resources required are extremely high due to the complexity of the metrics
Solution Approach 1:
The patent performs preliminary filtering using computationally inexpensive MBR intersection tests on segmented geometric components before applying expensive similarity metrics. By pre-processing the data to identify and discard obviously dissimilar object pairs through simple bounding box checks, the system avoids performing resource-intensive exact similarity calculations on all possible pairs, thereby significantly reducing overall computational resource consumption while preserving measurement accuracy for the subset of candidates that pass the preliminary filter.
3Measurement precision
If all pairs of geometry elements are put into relation with each other for calculating similarity metrics, then complete similarity assessment is achieved, but the algorithms are almost never sub-quadratic in computational complexity
Solution Approach 1:
The patent extracts and compares only the essential geometric components (key line segments, critical edges, or representative points) rather than relating all pairs of geometry elements. By identifying and focusing on the most discriminative geometric features that define object similarity, the system achieves meaningful similarity assessment without the quadratic complexity of exhaustive pairwise comparison of all geometry elements.
Data Source
Figure 1A~1B
Figure 2A~2B
Figure 3
AI summary
A computer-implemented method for range query processing of complex geometric data in a database with respect to a query distance to a query object, said method comprising the steps of determining a buffered query object with respect to the query object, wherein a geometric footprint of the buffered query object is increased with respect to the query object based on the query distance; determining a buffered query filter for the buffered query object using a filter function, wherein the filter function comprises determining a mapping of an overlapped cell of a grid to a corresponding element of the buffered query filter, the overlapped cell being overlapped by the buffered query object on the grid; and filtering out complex geometric data objects in the database for which the filter function generates corresponding filter objects, which are not a subset of the buffered query filter and/or which have an empty intersection with the buffered query filter.