3D Scan Registration via Cluster Segmentation and Pairwise Testing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for registering multiple scans in laser scanning and measurement systems often experience interruptions due to performance limitations, leading to incomplete or ambiguous registrations, especially when not all scans are examined pairwise.
Innovation Solution
The method involves generating clusters from scans based on quality criteria, allowing for automatic or user-assisted selection and registration of clusters, with optional user confirmations to ensure accurate joining of clusters and scans, thereby improving the registration process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all scans are examined pairwise for registration, then registration accuracy is improved, but processing time and computational resources increase significantly
Solution Approach 1:
The patent segments the scan data into clusters based on spatial proximity and temporal continuity. By dividing the complete set of scans into smaller cluster groups, the system performs pairwise examination only within each cluster rather than across all scans globally. This segmentation maintains registration accuracy for scans that are spatially and temporally related while dramatically reducing the total number of pairwise comparisons required.
Solution Approach 2:
The patent applies partial action by examining only a subset of scan pairs that are most likely to be related based on temporal and spatial criteria. Instead of performing exhaustive pairwise examination of all possible scan combinations, the system selectively processes scans within defined clusters, achieving sufficient registration accuracy for the application while avoiding the computational burden of complete examination.
2Productivity
If only neighborhood scans are examined for registration, then processing performance is improved, but registration completeness and reliability deteriorate
Solution Approach 1:
The patent performs preliminary organization of scans into clusters based on temporal and spatial relationships before the registration process. By pre-grouping scans that are likely to be related into clusters, the system ensures that relevant scans are identified and processed together, preventing interruptions and maintaining registration completeness while preserving processing efficiency.
Solution Approach 2:
The system incorporates feedback mechanisms where registration results from initial cluster processing inform subsequent processing decisions. If registration within a cluster achieves satisfactory results, the system can proceed efficiently; if not, it can adjust cluster definitions or processing parameters to improve reliability, thus maintaining both performance and completeness.
3Extent of automation
If automatic clustering is performed without user intervention, then automation level is improved, but handling of ambiguous or insufficient information worsens
Solution Approach 1:
The patent implements a dynamic registration system that adapts its automation level based on the quality and clarity of the data. When scan information is clear and unambiguous, the system operates fully automatically with high productivity. When information is ambiguous or insufficient, the system dynamically transitions to a semi-automatic mode, presenting options to the user for confirmation or correction, thus maintaining reliability across varying data conditions.
Solution Approach 2:
The system changes operational parameters based on data quality assessment. When information is sufficient and clear, automation parameters are set to maximum efficiency. When information is ambiguous or insufficient, the system adjusts parameters to reduce automation level and increase user involvement, ensuring reliable handling of challenging cases while maintaining high automation for routine cases.
Data Source
AI summary
A method for optically scanning and measuring a scene by a three-dimensional (3D) measurement device in which multiple scans are generated to then be registered in a joint coordinate system of the scene. At first at least one cluster is generated from at least one scan, further scans are registered for test purposes in the coordinate system of the cluster, and registering is then confirmed if specified quality criteria are fulfilled and the generated clusters are then joined, for which purpose pairs are formed of selected scans and/or clusters to form pairs, the pairs are registered for test purposes and registering is confirmed if appropriate.