Point Cloud Matching With Selective Shape Model Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for matching point cloud data, such as ICP and NDT, face challenges in achieving high-speed and high-accuracy processing.
Innovation Solution
A method and apparatus that generate input models by reducing data amount, extract relevant models based on shape information, and iteratively adjust position and posture to minimize cost, using ellipsoid or planar models generated from voxelized point clouds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ICP or NDT methods are used for matching point cloud data, then localization can be performed, but the processing speed and accuracy are insufficient
Solution Approach 1:
The patent segments the point cloud data into multiple voxel groups, where each voxel is further divided into sub-voxels. This segmentation allows the matching process to operate on divided, manageable units rather than the entire point cloud at once, improving both processing speed and accuracy through localized optimization.
Solution Approach 2:
The patent extracts shape information (such as ellipsoid parameters) from each voxel and sub-voxel to create simplified models. By extracting only the essential geometric characteristics rather than processing all raw point data, the system achieves high-speed processing while maintaining matching accuracy through the use of these extracted shape models.
2Productivity
If voxel division is applied to point cloud data for NDT-based localization, then processing can be performed, but the computational complexity remains high
Solution Approach 1:
The patent divides the voxel into multiple sub-voxels, creating a hierarchical structure that simplifies computational complexity at each level. By processing smaller sub-voxel units rather than large voxels directly, the system reduces computational burden while maintaining processing capability through systematic division of the problem space.
Solution Approach 2:
The patent creates simplified ellipsoid models that copy the essential shape characteristics of each voxel and sub-voxel. These copied geometric models serve as computationally efficient representations that capture the necessary spatial information without requiring complex processing of the original point cloud data, thereby reducing computational complexity while preserving processing capability.
Data Source
AI summary
An apparatus for matching point cloud data according to an embodiment includes a generation unit configured to generate an input model and an input model group including a plurality of the input models, the input model being obtained by reducing a data amount of the point cloud data, an extraction unit configured to extract some of the input models from the input model group according to shape information of the input models, a calculation unit configured to compare an extracted extraction model with a reference model based on reference point cloud data and calculate a cost, a determination unit configured to determine whether or not the cost has converged, and a change unit configured to change a position and a posture so that the cost is reduced when the cost has not converged.