Polygonal Mesh Generation Using Hierarchical Resolution Maps

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional mesh generation algorithms are limited in processing large datasets with numerous points, making it impractical to generate meshes for large-scale polygonal data sets, especially those with irregularly spaced data points, as they require all points to be in system memory.

Innovation Solution

A new mesh construction technique that uses a hierarchical resolution map and binning grid to process and store points efficiently, allowing for dynamic construction of mesh partitions, which can be triangulated independently and combined to form a continuous mesh without requiring all points to be in memory at once.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If conventional mesh generation algorithms are used to process large datasets, then complete mesh coverage is achieved, but system memory requirements become prohibitive

Engineering Contradiction:
Improvenumber of data pointsVSAvoidmemory requirements
Core Design Contradiction:
Quantity of substanceVSDevice complexity

Solution Approach 1:

The patent divides the large point cloud dataset into multiple smaller partitions, each of which can be processed independently in memory. The resolution map is also segmented into tiles corresponding to these partitions. This segmentation allows the system to handle billions of points by processing manageable subsets sequentially or in parallel, avoiding the need to load all points into memory simultaneously.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a hierarchical resolution map that adds a dimensional layer of abstraction between the raw point cloud and the final mesh. This resolution map stores aggregated statistical information about point density and distribution at different scales, enabling the system to reconstruct mesh geometry without retaining all original point data in memory.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Manufacturing precision

If all points are loaded into memory for mesh generation, then accurate mesh representation is achieved, but processing time and memory usage increase

Engineering Contradiction:
Improvemesh accuracyVSAvoidprocessing time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary processing to create a resolution map that captures essential geometric and statistical properties of the point cloud before actual mesh generation. This pre-computed resolution map contains information about point density, variance, and spatial distribution that guides subsequent mesh construction, allowing accurate mesh representation to be achieved from partitioned data without reloading all original points.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent transforms the raw point cloud data into a resolution map with different parameters (mean position, variance, point density) that preserve the essential geometric characteristics. This parameter transformation enables accurate mesh reconstruction from compressed representations, reducing memory requirements and processing time while maintaining mesh accuracy.

Inventive Principle:
Principle #35Parameter changes

3Area of stationary object

If conventional algorithms process large point clouds, then complete mesh coverage is achieved, but the algorithm becomes impractical for large datasets

Engineering Contradiction:
Improvemesh coverage areaVSAvoidprocessing efficiency
Core Design Contradiction:
Area of stationary objectVSProductivity

Solution Approach 1:

The patent segments both the point cloud data and the resolution map into spatial partitions that can be processed independently. Each partition corresponds to a specific region of the overall mesh coverage area. This segmentation enables parallel processing of multiple regions simultaneously, dramatically improving processing efficiency for large-scale datasets while maintaining complete mesh coverage across the entire area.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements a system where each partition can be processed autonomously using only the resolution map data for that region, without requiring access to the entire dataset. The resolution map serves as a self-contained guide that enables each partition to generate its mesh independently, allowing distributed and parallel processing that scales efficiently to large coverage areas.

Inventive Principle:
Principle #25Self-service

Data Source

PatentUS9390556B2Systems and methods for generating a large scale polygonal mesh
Publication Date: 2016.07.12 TELEDYNE DIGITAL IMAGING INC(CA)
  • US9390556B2 patent drawing
  • US9390556B2 patent drawing
  • US9390556B2 patent drawing

AI summary

A polygonal mesh is generated from a collection of points that are organized for a mesh partition in accordance with a tile that includes one or more bins used to process the points that define the mesh. The resolution of the tile is related to the number of bins for the tile. The organization of the tiles in a partition of the mesh permits the mesh to be constructed with partitions that are independent of each other and that can be joined to form a continuous mesh. The resolution of the mesh can be dynamic with respect to the partitions due to the variable resolution assigned to each tile. Portions of the mesh for which points are not provided can be assigned finalization points to permit a continuous mesh to be constructed.