Multi-Camera Map Drift Correction for AR and VR Pose Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for correcting mapping drift in augmented and virtual reality systems require users to revisit previously mapped areas, which is often impractical due to physical limitations or dynamic objects, leading to inaccuracies in pose estimation.
Innovation Solution
Utilizing multiple cameras with overlapping fields of view to extract feature points and compute mapping drift, allowing for continuous correction without revisiting previous areas by comparing position coordinates and updating the primary map based on secondary map data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If loop closing is used to correct drift by revisiting previously mapped areas, then mapping accuracy is improved, but user convenience and system adaptability deteriorate due to physical limitations and dynamic objects
Solution Approach 1:
The patent introduces a secondary map as an intermediary structure that enables drift correction without requiring users to revisit previously mapped areas. The secondary map is built using data from secondary cameras and serves as a reference for computing drift in the primary map, thereby resolving the contradiction between maintaining accuracy and ensuring user convenience
Solution Approach 2:
The patent replaces the mechanical requirement of physically revisiting mapped areas with a computational approach using multiple cameras and map comparison algorithms. Instead of relying on user movement to close loops, the system uses image data from primary and secondary cameras to compute drift through feature point matching and coordinate comparison
2Measurement precision
If multiple cameras with overlapping fields of view are used to compute drift continuously, then mapping accuracy is improved, but device complexity increases
Solution Approach 1:
The patent divides the mapping system into two independent segments: a primary map built from primary cameras and a secondary map built from secondary cameras. This segmentation allows each camera system to operate independently, simplifying the overall system architecture while enabling continuous drift correction through comparison between the two segmented maps
Data Source
AI summary
A method performed by an electronic device for correcting a mapping drift is provided. The method includes receiving, by the electronic device via one or more primary cameras of the electronic device, a first plurality of images associated with a scene, extracting, by the electronic device, a plurality of first feature points from each of the received first plurality of images, receiving, by the electronic device via one or more secondary cameras of the electronic device, a second plurality of images, extracting, by the electronic device, a plurality of second feature points from each of the received second plurality of images, wherein a field of view of the one or more secondary cameras overlaps with a field of view of the one or more primary cameras, computing, by the electronic device, a mapping drift in a primary map associated with the first plurality of images by comparing position coordinates of the extracted plurality of first feature points with position coordinates of the extracted plurality of second feature points, and correcting, by the electronic device, the mapping drift in the primary map based on the computed mapping drift.