Multi-Sensor Extrinsic Parameter Estimation via IR Pattern Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing devices with multiple sensors, such as head-mounted displays, face challenges in accurately and efficiently determining the spatial relationships between sensors, which can change due to adjustments or damage, affecting the inter-pupillary distance and overall performance.
Innovation Solution
The implementation of a system that uses infrared pattern projection and matching to estimate the extrinsic parameters of sensor-to-sensor relationships, allowing for accurate determination of relative positions and orientations between sensors, including those in head-mounted displays, by projecting IR patterns, capturing images, and matching 3D positions to adjust for inter-pupillary distance and other spatial configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If factory calibration is used to specify spatial relationships between sensors, then initial accuracy is achieved, but the spatial relationships may change due to adjustments or damage
Solution Approach 1:
The system performs preliminary calibration by projecting an IR pattern and capturing images before actual use. This preliminary action establishes the initial spatial relationships between sensors, which can then be updated if changes occur due to adjustments or damage.
Solution Approach 2:
The system continuously monitors spatial relationships by capturing images with multiple sensors and comparing the captured IR pattern positions. When discrepancies are detected indicating sensor movement, the system provides feedback to recalculate and update the extrinsic parameters, maintaining accuracy despite physical changes.
2Adaptability or versatility
If sensors are adjusted to account for different inter-pupillary distances, then adaptability is improved, but measurement accuracy of spatial relationships deteriorates
Solution Approach 1:
The system dynamically adjusts extrinsic parameters based on captured images rather than using fixed factory-calibrated values. When inter-pupillary distance adjustments are made, the system recaptures the IR pattern and recalculates spatial relationships, ensuring measurement precision is maintained despite physical reconfiguration.
Solution Approach 2:
The system changes the extrinsic parameters (spatial relationships) based on actual captured data rather than relying on fixed factory settings. By projecting the IR pattern and analyzing its appearance in multiple sensor images, the system determines updated parameters that reflect the current physical configuration, accommodating different IPD values while maintaining accuracy.
3Measurement precision
If traditional methods are used to determine sensor spatial relationships, then device complexity is reduced, but measurement precision and efficiency deteriorate
Solution Approach 1:
The system introduces an infrared pattern projector as an intermediary tool to facilitate accurate measurement. By projecting a known IR pattern and capturing it with multiple sensors, the system creates a reference framework that enables precise determination of spatial relationships without requiring complex direct measurement apparatus.
Solution Approach 2:
The system uses a virtual 3D model of the projected IR pattern as a copy or reference against which actual sensor captures are compared. By matching the captured pattern positions with the known virtual pattern geometry, the system efficiently calculates extrinsic parameters through geometric relationships rather than direct physical measurement.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This method enables precise and efficient estimation of sensor-to-sensor spatial relationships, ensuring correct positioning and alignment, even in dynamic conditions, thereby enhancing the performance and user experience of multi-sensor devices like head-mounted displays.
Implementation Method 1
The projector projects IR pattern elements onto an environment surface
Implementation Method 2
The first IR sensor captures a first image including first IR pattern elements corresponding to the projected IR pattern elements
Data Source
AI summary
In one implementation, a device has a processor, a projector, a first infrared (IR) sensor, a second IR sensor, and instructions stored on a computer-readable medium that are executed by the processor to estimate the sensor-to-sensor extrinsic parameters. The projector projects IR pattern elements onto an environment surface. The first sensor captures a first image including first IR pattern elements corresponding to the projected IR pattern elements and the device estimates 3D positions for first IR pattern elements. The second IR sensor captures a second image including second IR pattern elements corresponding to the projected IR pattern elements and the device matches the first IR pattern elements and the second IR pattern elements. Based on this matching, the device estimates a second extrinsic parameter corresponding to a spatial relationship between the first IR sensor and the second IR sensor.