Large Geometric Data Visualization Using Spatial Tree Partitioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems face challenges in efficiently processing and visualizing large amounts of three-dimensional data, particularly with LIDAR scanning technologies, where data sets can exceed computer memory limits, leading to slow processing and loss of information due to the need for data separation and preprocessing.
Innovation Solution
The system allows for real-time partitioning and management of large data sets, enabling users to select and display subsets of data without loading unnecessary blocks, using a spatial tree structure for storage and allowing actual point data to be displayed, with tools for hiding, showing, and making data transparent to focus on specific sections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If large data sets are loaded into computer memory for visualization, then complete data can be processed and displayed, but memory limits are exceeded and processing becomes slow
Solution Approach 1:
The patent divides the large point cloud data set into multiple manageable blocks that can be stored in a spatial tree structure. Each block contains a subset of points, allowing the system to load only necessary blocks into memory during visualization, thus avoiding memory overflow while maintaining processing efficiency.
Solution Approach 2:
The patent introduces a spatial hierarchy dimension through the spatial tree structure, organizing points not just in 3D space but also in a tree-based hierarchical space. This allows efficient navigation and selective loading of data blocks based on the current view requirements, improving processing speed without sacrificing data completeness.
2Ease of operation
If data is separated and preprocessed to fit memory limits, then visualization becomes possible, but information is lost and processing time increases
Solution Approach 1:
The patent performs preliminary organization of point cloud data into a spatial tree structure during data ingestion, creating blocks and hierarchical relationships in advance. This preliminary action enables efficient runtime queries and selective loading without requiring data separation or downsampling, thus preserving complete information while enabling visualization.
Solution Approach 2:
The patent applies different levels of detail and processing to different regions of the data based on their importance and visibility requirements. High-priority regions are processed with full precision, while less important regions use coarser representation, maintaining data accuracy where needed while reducing overall processing burden.
3Quantity of substance
If all data blocks are loaded for display, then complete visualization is achieved, but system resources are overwhelmed
Solution Approach 1:
The patent implements dynamic loading and unloading of data blocks based on the current camera position, field of view, and user interactions. As the user navigates through the point cloud, relevant blocks are automatically loaded into memory while irrelevant blocks are unloaded, maintaining data completeness conceptually while keeping actual memory usage proportional to current needs.
Solution Approach 2:
The patent extracts and displays only the necessary subset of data blocks that are relevant to the current view, separating them from the rest of the data set. This extraction is performed efficiently using the spatial tree structure, which allows rapid identification and loading of only those blocks intersecting with the current frustum, reducing system resource usage while maintaining visualization quality.
Data Source
AI summary
A representation of a physical object can be displayed even where the amount of geometric data is too large to be stored in resident memory. A primary viewing window displays point data for the object using a substantially even sampling of data at an appropriate point density for the system. At least one auxiliary viewing window displays a two-dimensional representation of the point data. A user can select a portion of the data in the auxiliary window(s), such as by selecting cells of an overlaid grid, to be displayed in the primary window using a rendering such as a “visible” rendering. The remainder of the displayed data can be displayed using a rendering such as a “hidden” or “transparent” rendering. The resolution of the selected region can be increased while maintaining a substantially even spacing among points for the region. The resolution of the unselected region can be decreased accordingly.