Epipolar-Constrained Bundle Adjustment for AR Headset Calibration

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Augmented reality systems face challenges in accurately accounting for deformations in wearable devices due to movements and deformations, leading to inaccuracies in camera calibration and environmental mapping.

Innovation Solution

Implementing bundle adjustment using epipolar constraints to refine camera positions and extrinsic parameters in real-time, accounting for deformations caused by movements and rotations in wearable augmented reality devices.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If bundle adjustment is performed using traditional reprojection error minimization, then computational efficiency is maintained, but accuracy of camera calibration deteriorates due to unaccounted deformations in wearable devices

Engineering Contradiction:
Improvecamera calibration accuracyVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent changes the optimization parameters by introducing epipolar geometry constraints as additional error terms in the bundle adjustment process. Instead of only minimizing reprojection error, the system minimizes a combined error function that includes epipolar constraints, which account for camera deformations. This parameter change enables accurate calibration despite wearable device movements without requiring complex deformation modeling.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent replaces direct mechanical deformation measurement (which would require complex sensors and models) with a computational substitution using epipolar geometry constraints. By substituting the mechanical deformation problem with a geometric constraint-based mathematical model, the system achieves accurate calibration while maintaining computational efficiency suitable for real-time operation.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If epipolar constraints are added to bundle adjustment to account for deformations, then calibration accuracy improves, but computational load increases

Engineering Contradiction:
Improveextrinsic parameters accuracyVSAvoidprocessing energy consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent applies partial action by selectively enforcing epipolar constraints only on corresponding feature points that are visible in multiple camera views, rather than attempting to model all possible deformations. This partial application of constraints provides sufficient accuracy improvement while avoiding the excessive computational burden of comprehensive deformation modeling, enabling real-time operation on wearable devices.

Inventive Principle:
Principle #16Partial or excessive action

3Reliability

If traditional bundle adjustment is used without deformation compensation, then processing speed is maintained, but reliability of environmental mapping deteriorates

Engineering Contradiction:
Improveenvironmental mapping reliabilityVSAvoidprocessing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent introduces epipolar geometry constraints as an intermediary mathematical model that mediates between the observed image data and the underlying camera deformation. This intermediary constraint system translates deformation effects into measurable geometric relationships, enabling reliable environmental mapping without requiring direct measurement or complex modeling of the physical deformation processes.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP4260162B1Bundle adjustment using epipolar constraints
Publication Date: 2026.04.22 MAGIC LEAP INC
  • EP4260162B1 patent drawingFigure 1
  • EP4260162B1 patent drawingFigure 2
  • EP4260162B1 patent drawingFigure 3

AI summary

Methods, systems, and apparatus for performing bundling adjustment using epipolar constraints. A method includes receiving image data from a headset for a particular pose. The image data includes a first image from a first camera of the headset and a second image from a second camera of the headset. The method includes identifying at least one key point in a three-dimensional model of an environment at least partly represented in the first image and the second image and performing bundle adjustment. Bundle adjustment is performed by jointly optimizing a reprojection error for the at least one key point and an epipolar error for the at least one key point. Results of the bundle adjustment are used to perform at least one of (i) updating the three-dimensional model, (ii) determining a position of the headset at the particular pose, or (iii) determining extrinsic parameters of the first camera and second camera.