3D Scan Registration via Cluster Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for registering multiple scans in laser scanning and measurement systems often face interruptions due to performance limitations, leading to incomplete or ambiguous registrations, especially when there are not enough targets in overlapping areas or complex topologies.
Innovation Solution
The method involves generating clusters from scans based on quality criteria, allowing for automatic or user-assisted selection and registration of pairs, with optional user confirmations, to improve the registration process by ensuring sufficient information is gathered and reducing ambiguity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pairwise examination of all scans is performed for registering, then registration accuracy is improved, but processing time and computational resources increase significantly
Solution Approach 1:
The patent segments the registration process into two distinct phases: a clustering phase that groups scans into neighborhoods, and a registration phase that processes only pairs within those neighborhoods. This segmentation reduces the computational complexity from examining all possible pairs to examining only local pairs, thereby maintaining registration accuracy while significantly reducing processing time.
Solution Approach 2:
The patent applies partial action by examining only a subset of scan pairs (those within the same neighborhood) rather than all possible pairs. The neighborhood concept ensures that sufficient local information is gathered for accurate registration without the excessive computational burden of global pairwise comparison.
2Productivity
If only neighborhood scans are examined for registering, then processing performance is improved, but registration completeness and accuracy deteriorate
Solution Approach 1:
The patent divides the complete set of scans into multiple neighborhoods through clustering, ensuring that each scan is assigned to an appropriate neighborhood based on spatial or temporal criteria. This segmentation allows efficient local processing while maintaining global coverage through the union of all neighborhoods.
Solution Approach 2:
The patent merges multiple local registration results from different neighborhoods into a complete global registration. By combining the results from various neighborhoods, the system achieves both processing efficiency (through localized operations) and registration completeness (through global integration).
3Extent of automation
If automatic clustering is performed without user intervention, then automation level is improved, but ability to handle ambiguous or insufficient information deteriorates
Solution Approach 1:
The patent implements a dynamic registration system that can switch between fully automatic mode and user-assisted mode based on the quality of information available. When scan information is sufficient and unambiguous, the system operates automatically. When information is insufficient or ambiguous, the system dynamically transitions to request user input, thereby maintaining high automation levels while ensuring registration reliability.
Solution Approach 2:
The patent incorporates feedback mechanisms where the system evaluates the quality of available scan information and automatically determines whether user intervention is needed. This feedback loop allows the system to maintain optimal automation levels while handling challenging registration scenarios through selective user engagement.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances the performance and reliability of the registration process by ensuring that scans are accurately joined and registered, even in challenging environments, by allowing for both automatic and user-intervention strategies, thereby improving the generation of a complete three-dimensional point cloud.
Implementation Method 1
a laser scanner is taken to a new location after a scan to generate an additional scan
Implementation Method 2
A distance meter in the device measures a distance to the object O
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, if specified quality criteria are fulfilled and the generated clusters are then joined, for which purpose clusters are selected, registered for test purposes and registering is confirmed if appropriate, wherein the clusters to be joined are visualized with an optional possibility for the user to intervene, for supporting the selection of clusters.