Geospatial Data Clustering via Tile-Based Link Partitioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for parallel processing of large geospatial data sets in mapping applications face challenges in efficient partitioning, leading to duplication errors and increased computational resources and time, especially when dealing with non-uniformly distributed data.
Innovation Solution
A method for partitioning geospatial data by determining link tile associations, generating families of links, and identifying clusters centered on these links, which are then assigned to tiles for aggregation, ensuring accurate and efficient parallel processing across multiple processors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If parallel processing nodes are used to speed up clustering of large geospatial data, then processing speed is improved, but data partitioning complexity increases and duplication errors occur
Solution Approach 1:
The patent divides the geospatial data into discrete tiles based on geographic boundaries. Each processing node is assigned specific tiles to process independently, creating clear segmentation of the data workload. This segmentation approach enables parallel processing while maintaining manageable complexity through standardized tile-based units.
Solution Approach 2:
The patent introduces a coordination mechanism that acts as an intermediary between processing nodes and the central system. This intermediary manages the assignment of tiles to processing nodes and handles the aggregation of results, reducing the complexity burden on individual nodes while enabling coordinated parallel processing across the distributed system.
2Productivity
If traditional partitioning methods are used for parallel processing, then processing can be distributed, but duplication errors increase and computational resources are wasted
Solution Approach 1:
The patent applies local quality by assigning specific geographic tiles to specific processing nodes based on their spatial boundaries. Each processing node processes only the data within its assigned tiles, ensuring that computational resources are focused on localized data subsets without duplication. This spatially-aware partitioning eliminates redundant processing while maintaining distributed processing capabilities.
3Measurement precision
If all geospatial data is made available at once for clustering algorithms, then clustering accuracy is improved, but processing time becomes unacceptable for large data sets
Solution Approach 1:
The patent segments the large geospatial dataset into multiple smaller tiles that can be processed independently and in parallel. Each processing node receives a subset of tiles and performs clustering locally, producing accurate results for its assigned region. The final comprehensive clustering accuracy is achieved by aggregating these local results, while the parallel processing dramatically reduces total processing time compared to sequential processing of the entire dataset.
Data Source
AI summary
System and methods are provided for parallel geospatial clustering. A link tile association is determined between one or more links and a tile of a plurality of tiles. Map matched geospatial data is acquired for the tile based on the link tile association. A family of links is generated from the one or more links. A cluster of the map matched geospatial data is identified that is at least partially associated with the family of links. The cluster is determined to be centered on a link of the family of links that originates in the tile. The cluster is assigned to the tile. The cluster is reported for aggregation with other clusters from other parallelized processes.