Camera Pose Relocalization With Parallel Anchor Image Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The relocalization process in camera pose tracking consumes a relatively long time, leading to a blocked main thread and delayed relocalization results, which cannot be directly applied to current images.
Innovation Solution
A relocalization method involving a front-end program that switches to a new anchor image when conditions are met, parallelly performing camera pose tracking, while a back-end program performs relocalization on the new anchor image, allowing for immediate correction of positioning results using both pose parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If relocalization is performed on the target image relative to the first anchor image, then positioning accuracy is improved, but processing time increases and main thread is blocked
Solution Approach 1:
The system performs preliminary actions by switching to a new anchor image when tracking quality degrades, preparing for potential relocalization needs. The front-end program continuously monitors tracking quality and proactively switches anchor images before complete tracking failure occurs, reducing the urgency and complexity of emergency relocalization operations.
Solution Approach 2:
The relocalization process is segmented into multiple independent tracking processes, each with its own anchor image. Instead of performing a single comprehensive relocalization, the system divides the tracking space into multiple segments (different anchor images), allowing parallel processing and reducing the computational burden on the main thread.
2Stability of the object's composition
If feature point tracking is performed on current image relative to i th anchor image, then tracking continuity is maintained, but tracking accuracy degrades over time
Solution Approach 1:
The system dynamically switches between different anchor images based on tracking quality metrics. When the current tracking process shows degradation, the system transitions to a new anchor image, making the tracking system adaptive rather than static. This dynamic adjustment maintains both continuity and accuracy by resetting the tracking reference before accuracy degrades too much.
Solution Approach 2:
The system implements feedback mechanisms by monitoring tracking quality and using this information to determine when to switch anchor images. The front-end program evaluates tracking performance and provides feedback that triggers anchor image switching, creating a closed-loop control system that maintains tracking accuracy while ensuring continuity.
3Measurement precision
If anchor image is switched to maintain tracking quality, then tracking accuracy is improved, but system complexity increases
Solution Approach 1:
The front-end program is designed with multi-functionality, handling both normal tracking operations and anchor image switching within a single unified system. This universal design avoids the need for separate specialized modules for each function, managing complexity while maintaining the capability to switch anchor images when needed for improved tracking accuracy.
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
This application discloses a relocalization method and apparatus in a camera pose tracking process, a device, and a storage medium, belonging to the field of augmented reality (AR). The method includes: switching, by a front-end program, a target image to an (i+1)th anchor image in a case that the target image satisfies a relocalization condition, and transmitting the target image to a back-end program; obtaining, by the front-end program, a current image acquired after the target image, and performing feature point tracking on the current image relative to the target image to obtain a first pose parameter; performing, by the back-end program, relocalization on the target image to obtain a second pose parameter, and transmitting the second pose parameter to the front-end program; and calculating, by the front-end program, a current pose parameter of the current image according to the first pose parameter and the second pose parameter. This application can correct a positioning result of a current image even if it takes a relatively long time for a back-end program to obtain a relocalization result of a target image, thereby ensuring a relatively adequate relocalization effect.