Point Cloud Registration Using Overlap Cell Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing point cloud registration methods are inefficient in aligning multiple point clouds in a common coordinate system, particularly for large datasets, leading to slow runtime and suboptimal alignment accuracy.
Innovation Solution
A method involving 2D or 3D overlap volume division into cells, followed by a comparison process to determine similarity scores and select the best matching target sets, and applying a transformation matrix to register the point clouds, utilizing techniques like Moore neighborhoods and convex hull comparisons for enhanced alignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional point cloud registration methods are used, then the process can align point clouds in a common coordinate system, but the runtime becomes slow and alignment accuracy is suboptimal for large datasets
Solution Approach 1:
The patent divides the overlapping region between source and target point clouds into multiple cells (e.g., 2D grid cells or 3D volumetric cells). This segmentation allows the registration process to focus on smaller, manageable regions rather than processing the entire point cloud at once, thereby reducing computational time while maintaining alignment accuracy through localized comparisons.
Solution Approach 2:
The patent performs detailed comparison and similarity scoring only in the overlapping region where the source and target point clouds actually intersect. By concentrating computational resources on the local overlapping area rather than the entire space, the method achieves high alignment accuracy in relevant regions without the time penalty of processing unrelated data.
2Reliability
If the entire point cloud is processed for registration, then comprehensive alignment can be achieved, but computational complexity and processing time increase significantly
Solution Approach 1:
The patent extracts only the overlapping region between source and target point clouds for registration processing. By taking out and focusing solely on the relevant overlapping portion rather than processing the entire point cloud, the method maintains registration reliability for the actual alignment task while significantly reducing computational complexity.
Solution Approach 2:
The patent applies partial action by processing only the necessary overlapping region rather than the entire point cloud. This selective processing approach ensures sufficient alignment accuracy for reliable registration while avoiding the excessive computational burden of analyzing the complete point cloud data.
3Measurement precision
If detailed comparison of all points is performed, then high alignment accuracy can be achieved, but processing time increases
Solution Approach 1:
The patent segments the overlapping region into cells and performs detailed point-by-point comparison only within these cells. This segmentation strategy maintains high alignment accuracy through thorough local comparison while improving processing speed by avoiding the time-consuming task of comparing all points across the entire point cloud.
Solution Approach 2:
The patent performs preliminary actions by first identifying and isolating the overlapping region, then dividing it into cells before conducting detailed comparisons. This preliminary organization of data reduces the complexity of subsequent processing and enables faster alignment while maintaining accuracy.
Data Source
AI summary
A computer-implemented method comprising: determining a 2D overlapping portion between a source 2D projection of a source 3D point cloud and a target 2D projection of a target 3D point cloud, the 2D overlapping portion corresponding to an overlapping region between the source and target geographical areas; dividing the 2D overlapping portion into a plurality of cells and determining at least one overlap cell which includes at least a threshold number of points from the source 2D projection and at least a threshold number of points from the target 2D projection; performing a comparison process for said overlap cell and selecting a best matching target set; determining a transformation between the points of the source set and the points of the best matching target set or vice versa; and applying the transformation to register the source 3D point cloud to the target 3D point cloud or vice versa.