3D Scan Alignment Using Voxel Maps Without External References
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for aligning 3D scan datasets face challenges such as precise alignment without references, lack of dimension verification, and scalability constraints, particularly when dealing with large objects like airplanes or objects scanned from multiple angles.
Innovation Solution
A computing system that uses a processor to generate a voxel-based representation of a target object, compute transformation matrices for scan point sets, and apply these matrices to align scan points within an outer surface voxel layer, thereby transforming misaligned points to accurate positions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If external tracking systems or SLAM techniques are used for alignment, then alignment capability is improved, but device complexity and operational cost increase
Solution Approach 1:
The system uses the scanned object itself as the reference framework through voxel map generation. The object's own geometry provides the alignment reference, eliminating the need for external tracking systems, markers, or complex SLAM infrastructure. This self-referential approach achieves precise alignment while reducing device complexity and operational cost.
Solution Approach 2:
The system creates a voxel-based digital copy (voxel map) of the scanned object from the point cloud data. This voxel map serves as a reference model that can be efficiently processed and compared against new scan data, enabling precise alignment without requiring complex external tracking systems or physical reference markers.
2Productivity
If traditional point-to-surface alignment methods are used, then alignment process is simplified, but processing speed and scalability deteriorate with large datasets
Solution Approach 1:
The system segments the continuous point cloud data into discrete scan point sets that can be processed independently. Each scan point set is aligned separately using voxel-based methods, enabling parallel processing and significantly improving processing speed for large-scale datasets while maintaining alignment precision.
Solution Approach 2:
The system transforms the data representation from traditional point-cloud coordinates to a voxel-based spatial grid system. This parameter change enables more efficient data structures and algorithms that scale better with data volume, improving processing speed and productivity while handling large datasets from complex objects like airplanes.
3Measurement precision
If scans are acquired from multiple angles without reference points, then coverage and comprehensiveness are improved, but alignment accuracy and dimension verification become difficult
Solution Approach 1:
The system uses the scanned object's own geometry as the reference framework. By generating a voxel map from the scanned data itself, the system creates an internal reference system that enables accurate alignment and dimension verification without requiring external reference points, markers, or tracking infrastructure.
Solution Approach 2:
The system transforms the 3D point cloud data into a voxel-based representation that adds a discrete spatial grid dimension. This voxel map provides a structured reference framework that enables precise alignment verification across multiple scan angles, converting the alignment problem into a voxel-coordinate transformation problem that can be solved without external references.
Data Source
AI summary
A computing system for aligning laser scan data is provided, including a processor configured to obtain multiple scan point sets, each set including a plurality of scan points that represent an outer surface of a target object. The processor is configured to generate a voxel map including an outer surface voxel layer for the target object, and compute a respective transformation matrix for each of the scan point sets. The processor is configured to apply the respective transformation matrix for each scan point set to the scan points in the respective scan point set to thereby transform a position of each of the scan points in each scan point set that lies partially or wholly outside of the outer surface voxel layer to a modified position within the outer surface voxel layer, to thereby generate respective transformed scan point sets containing the scan points at the modified positions.