Flexible Stereo Depth Correction Using VIO Bias Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Flexible devices used in augmented and virtual reality systems experience bending, which leads to inaccurate depth sensing due to shifts from factory-calibrated configurations, causing errors in stereo image alignment and unstable yaw estimation, especially in indoor settings.
Innovation Solution
A method is introduced to estimate and compensate for bending in flexible stereo-to-depth devices using Visual-Inertial Odometry (VIO) stereo matches, validating rectification by projecting stereo VIO features to a rectified coordinate system and triggering a bending compensation process when features do not lie on the same raster line, thereby correcting pitch-roll and yaw biases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If factory calibration parameters are used for depth sensing, then initial alignment accuracy is achieved, but parameter drift occurs over time due to mechanical stress and temperature changes
Solution Approach 1:
The system continuously monitors the alignment between VIO features and depth map features, detects drift when misalignment exceeds thresholds, and triggers rectification processes to correct the calibration parameters, creating a closed-loop feedback system that maintains accuracy over time
Solution Approach 2:
The system dynamically adjusts calibration parameters based on detected drift conditions, changing the relative position and orientation parameters between cameras and sensors to compensate for mechanical stress and temperature-induced deviations from factory calibration
2Measurement precision
If stereo vision systems are used for depth sensing, then depth information is obtained, but bending of flexible devices causes misalignment between stereo images
Solution Approach 1:
The system makes the stereo rectification dynamic by continuously checking alignment conditions and applying rectification transformations when bending is detected, allowing the system to adapt to changing device shapes while maintaining depth sensing accuracy
3Measurement precision
If VIO features are projected to rectified coordinate system for validation, then bending detection accuracy is improved, but computational resources increase
Solution Approach 1:
The system performs partial rectification validation by checking only critical alignment conditions (whether features lie on the same raster line) rather than full rectification, triggering bending compensation only when necessary, thus reducing computational overhead while maintaining detection accuracy
Data Source
AI summary
A method for correcting a bending of a flexible device is described. In one aspect, the method includes accessing feature data of a first stereo frame that is generated by stereo optical sensors of the flexible device, the feature data generated based on a visual-inertial odometry (VIO) system of the flexible device, accessing depth map data of the first stereo frame, the depth map data generated based on a depth map system of the flexible device, estimating a pitch-roll bias and a yaw bias based on the features data and the depth map data of the first stereo frame, and generating a second stereo frame after the first stereo frame, the second stereo frame based on the pitch-roll bias and the yaw bias of the first stereo frame.