Surface Reconstruction Using Anchored Pose Graph Volumes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Surface reconstruction in large physical spaces faces challenges due to cumulative inaccuracies in tracking and mapping errors, leading to inconsistent SR data, especially when operating on a small scale or ignoring these errors.
Innovation Solution
The approach involves creating multiple separate SR volumes anchored to nodes in a pose graph, allowing initial overlap to avoid gaps, and adjusting their poses based on updates to the graph, with the option to re-anchor or delete volumes as necessary to maintain accurate surface reconstruction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If surface reconstruction is performed in large physical spaces using depth maps and pose graphs, then the coverage area increases, but cumulative tracking and mapping errors cause degradation in measurement precision
Solution Approach 1:
The patent divides the large physical space into multiple overlapping local coordinate systems, each covering a specific region. Instead of maintaining a single global coordinate system that accumulates errors over large distances, the system segments the space into manageable chunks with their own local poses and transforms, thereby limiting error propagation within each segment.
Solution Approach 2:
The patent introduces pose graphs as intermediary structures that connect and relate different local coordinate systems. The pose graph serves as a mediator that manages transformations between local frames, allowing the system to maintain accurate relative positions while covering large areas without direct accumulation of errors across the entire space.
2Area of stationary object
If multiple SR volumes are created to cover large physical spaces, then the coverage area increases, but managing and coordinating these volumes increases device complexity
Solution Approach 1:
The patent segments the surface reconstruction task into multiple independent SR volumes, each associated with a specific node in the pose graph. This segmentation allows each volume to be managed independently with its own coordinate system and data structures, reducing the overall complexity compared to managing a single large-volume system.
Solution Approach 2:
The patent creates a universal framework where SR volumes can be dynamically created, anchored, and managed through the pose graph structure. This multi-functional system handles volume creation, positioning, updating, and deletion through standardized operations, reducing management complexity despite the increasing number of volumes.
3Measurement precision
If volumes are re-anchored to maintain accuracy, then measurement precision improves, but the frequency of updates increases processing time
Solution Approach 1:
The patent implements periodic re-anchoring of SR volumes to the pose graph, where volumes are updated at specific intervals or when certain conditions are met, rather than continuously. This periodic action maintains measurement precision by regularly correcting positions while avoiding unnecessary continuous updates that would waste processing time.
Solution Approach 2:
The system uses feedback from the pose graph updates to determine when and how to re-anchor SR volumes. By monitoring changes in the pose graph and only triggering volume re-anchoring when necessary, the system maintains accuracy while minimizing unnecessary processing operations and reducing overall processing time.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Depth maps of a physical space are obtained using a depth sensor carried by a rig such as a robot or a head mounted display device worn by a user. Visible light images are also obtained. The images and orientation readings are used to create a pose graph which includes nodes connected by links. The nodes are associated with different poses of the rig and the corresponding images. Links between the nodes represent correspondences between the images, and transforms between coordinate systems of the nodes. As new images are captured, the pose graph is updated to reduce an accumulation of errors. Furthermore, surfaces in the physical space can be reconstructed at any time according to the current state of the pose graph. Volumes used in a surface reconstruction process are anchored to the nodes such that the positions of the volumes are adjusted as the pose graph is updated.