3D Point Cloud Registration Using Dominant Plane Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional 3D/2D image registration methods face challenges in environments where imaging parameters such as lighting, positioning, and motion are not easily controllable, leading to suboptimal performance, especially in industrial applications where real-time processing is required.
Innovation Solution
A method that estimates the mathematical relationship between 3D roto-translations of dominant planes in a 3D point cloud and bi-dimensional homographies in a 2D image plane, using a trained classifier to determine the motion of dominant planes, allowing for real-time 3D point cloud registration with multiple 2D images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional feature-based matching or intensity-based matching methods are used for 3D/2D image registration, then registration accuracy can be achieved under controlled conditions, but real-time performance is lost due to high computational demand
Solution Approach 1:
The patent segments the 3D point cloud into multiple dominant planes, and for each plane, identifies corresponding 2D image regions. This segmentation allows parallel processing of different planes, reducing overall computational time while maintaining registration accuracy for each plane independently.
Solution Approach 2:
The patent extracts dominant planes from the 3D point cloud and their corresponding 2D image regions, isolating the most significant geometric features for registration. This extraction focuses computational resources on critical elements rather than processing all points, enabling real-time performance.
2Reliability
If conventional registration methods process all 3D points and 2D pixels, then comprehensive registration is achieved, but computational complexity increases significantly
Solution Approach 1:
The patent applies different processing strategies to different regions: dominant planes receive focused registration attention with higher computational resources, while non-dominant regions are processed more simply or excluded. This local quality approach maintains registration reliability for critical surfaces while reducing overall computational complexity.
Solution Approach 2:
The patent performs registration on only the most significant dominant planes rather than all possible surfaces. This partial action focuses computational effort on the most impactful elements, achieving sufficient registration reliability for the application while dramatically reducing computational complexity.
3Loss of information
If multiple 2D images are registered to 3D point cloud, then imaging coverage and information completeness improve, but processing time and computational load increase
Solution Approach 1:
The patent performs preliminary identification and classification of dominant planes in the 3D point cloud before processing multiple 2D images. This preliminary action organizes the data structure in advance, enabling efficient matching of multiple images to the same dominant planes without redundant computations, thus reducing processing time while maintaining information completeness.
Solution Approach 2:
The patent merges the registration results from multiple 2D images onto the same dominant planes in the 3D point cloud. Instead of treating each image independently, the results are combined and integrated, reducing redundant processing while maintaining complete imaging information from all views.
Data Source
AI summary
A computer-implemented method of performing a three-dimensional 3D point cloud registration with multiple two-dimensional (2D) images may include estimating a mathematical relationship between 3D roto-translations of dominant planes of objects in a 3D point cloud and bi-dimensional homographies in a 2D image plane, thereby resulting in a 3D point cloud registration using multiple 2D images. A trained classifier may be used to determine correspondence between homography matrices and inferred motion of the dominant plane(s) on a 3D point cloud for paired image frames. A homography matrix between the paired images of the dominant plane(s) on the 2D image plane may be selected based on the correspondence between the inferred motions and measured motion of the dominant plane(s) on the 3D point cloud for the paired image frames. The process may be less computationally intensive than conventional 2D-3D registration approaches.