Virtual Surface Mapping With Occupancy Volumes for AR Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing artificial reality systems struggle with accurately modeling environments by segmenting scenes into virtual surfaces and managing positional information of these surfaces relative to spatial anchors, particularly when there are mismatches between stored positional information and real-time SLAM data.
Innovation Solution
A computing device generates virtual surfaces and occupancy volumes based on sensor data, using RANSAC and machine-learning algorithms to detect planar surfaces, and updates spatial relationships using Truncated Signed Distance Fields (TSDFs) to adjust positional information dynamically.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional environment modeling methods are used without per-surface volume representation, then the computational complexity is lower, but the accuracy of positional information and scene realism deteriorates
Solution Approach 1:
The environment model is segmented into multiple virtual surfaces, each with its own associated volume representation. This segmentation allows the system to focus computational resources on individual surfaces and their local volumes, improving positional accuracy without requiring complex global computations for the entire scene.
Solution Approach 2:
The patent introduces a volumetric dimension associated with each two-dimensional virtual surface. By adding this third dimension (volume) to each surface, the system achieves more accurate spatial representation and positional information while maintaining manageable computational complexity through localized volume representations.
2Reliability
If detailed environment modeling with per-surface volumes is implemented, then the scene realism is improved, but the computational resources required increase
Solution Approach 1:
Each virtual surface is equipped with its own dedicated volume representation that captures local spatial characteristics. This local quality approach ensures high scene realism in each region without requiring complex global models, as each surface-volumepair independently contributes to overall realism with computationally efficient local representations.
3Adaptability or versatility
If virtual surfaces are updated based on real-time SLAM data, then the adaptability to dynamic environments is improved, but the difficulty of detecting and measuring positional relationships increases
Solution Approach 1:
The patent introduces spatial anchors as intermediary elements that mediate between virtual surfaces and the real-time SLAM coordinate system. These anchors provide stable reference points that simplify the detection and measurement of positional relationships, enabling the system to adapt to dynamic environments while maintaining manageable complexity in tracking positional changes.
Data Source
AI summary
In one embodiment, a method includes receiving sensor data of a scene captured using one or more sensors, generating (1) a number of virtual surfaces representing a number of detected planar surfaces in the scene and (2) a point cloud representing detected features of objects in the scene based on the sensor data, assigning each point in the point cloud to one or more of the number of virtual surfaces, generating occupancy volumes for each of the number of virtual surfaces based on the points assigned to the virtual surface, generating a datastore including the number of virtual surfaces, the occupancy volumes of each of the number of virtual surfaces, and a spatial relationship between the number of virtual surfaces, receiving a query, and sending a response to the query, the response including an identified subset of the plurality of virtual surfaces in the datastore that satisfy the query.