Tracking Camera Recalibration for Extrinsic Drift Compensation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing calibration techniques for input sensors in dynamic environments, such as moving vehicles, fail to account for mechanical and thermal stresses that cause extrinsic drifts, leading to inaccurate spatial reconstruction and compromised user experience in augmented reality systems.

Innovation Solution

A system and method for on-the-fly recalibration of tracking cameras using features from reference and latest images, focusing on non-adjustable parts of the enclosed space, to adjust intrinsic and extrinsic parameters and mitigate positional drifts.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If explicit calibration processes are performed during manufacturing, then initial sensor accuracy is achieved, but long-term reliability deteriorates due to extrinsic drifts from mechanical stresses and temperature changes

Engineering Contradiction:
Improvesensor accuracyVSAvoidlong-term reliability
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The system performs preliminary calibration during manufacturing to establish baseline extrinsic parameters, then uses these pre-calibrated parameters as reference for subsequent drift detection and compensation operations throughout the device's operational lifetime

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system continuously monitors feature positions in captured images, compares them against reference positions from initial calibration, detects drifts, and automatically adjusts extrinsic parameters based on this feedback to maintain long-term accuracy

Inventive Principle:
Principle #23Feedback

2Measurement precision

If calibration is performed manually during manufacturing, then initial spatial reconstruction accuracy is achieved, but productivity decreases due to time-consuming calibration processes

Engineering Contradiction:
Improvespatial reconstruction accuracyVSAvoidcalibration efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The system performs calibration automatically using its own imaging capabilities and processing units, eliminating the need for manual intervention. The tracking camera captures images, the processor identifies features, calculates drifts, and adjusts parameters autonomously throughout operation

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

Instead of performing calibration as a discrete manual step during manufacturing, the system continuously performs calibration-related operations (image capture, feature extraction, drift detection, parameter adjustment) throughout its operational lifetime, transforming calibration from a batch process into a continuous automated process

Inventive Principle:
Principle #20Continuity of useful action

3Adaptability or versatility

If tracking cameras are used in dynamic environments like moving vehicles, then adaptability to real-world applications is improved, but measurement precision deteriorates due to mechanical stresses and thermal expansion causing extrinsic drifts

Engineering Contradiction:
Improvereal-world application capabilityVSAvoidtriangulation accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The system transitions from static calibration parameters established during manufacturing to dynamic parameter adjustment. Extrinsic parameters are continuously updated based on real-time drift detection, allowing the system to adapt to changing mechanical and thermal conditions in dynamic environments like moving vehicles

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system detects changes in extrinsic parameters (position, orientation) caused by mechanical stresses and thermal expansion, then compensates for these parameter changes by adjusting the calibration data to maintain triangulation accuracy despite environmental variations

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4648011A1On-the-fly recalibration of tracking camera
Publication Date: 2025.11.12 DISTANCE TECHNOLOGIES OY
  • EP4648011A1 patent drawingFigure 1~2A
  • EP4648011A1 patent drawingFigure 2B~2C
  • EP4648011A1 patent drawingFigure 3~4

AI summary

A first set of reference images (200) is captured using tracking camera(s) (102, 312). Features (202, 204) are extracted from said first set. A first set of features that pertain to at least non-adjustable part(s) (210a-f) of an enclosed space (206) is selected. Positions of the features of the first set are determined. Latest image(s) is/are captured using the tracking camera(s). Features are extracted from the latest image(s). A second set of features that pertain to at least the non-adjustable part(s) and that match with at least a subset of the first set of features, is selected. Positions of the features of the second set are determined. A difference in a position of a given feature of the second set and a position of a corresponding feature of the first set is determined. The tracking camera(s) is/are calibrated.