Optical Sensor Pose Estimation with Decoupled Measurement Noise
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating the pose of optical positioning devices, such as digital pens or stylus devices, face inaccuracies due to ambiguity between camera translation and rotations, which are computationally complex and difficult to implement effectively.
Innovation Solution
A method that decorrelates measurement noise terms by estimating the position of an aiming point, orientation, and distance of the optical sensor with respect to the target surface, using a simplified measurement model and sensor fusion with additional sensing systems like inertial measurement units, and optionally Time-of-Flight sensors, to reduce inaccuracies and simplify computations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional pose estimation methods are used, then computational accuracy is maintained, but device complexity and computational load increase
Solution Approach 1:
The patent segments the pose estimation problem into two independent parts: (1) estimating the optical axis direction using only image moment information, and (2) estimating the distance along the optical axis separately using depth sensor data. This segmentation eliminates the need for complex iterative optimization while maintaining accuracy, as each sub-problem can be solved with simple, direct computations rather than full pose estimation algorithms.
Solution Approach 2:
The patent introduces an intermediary coordinate system aligned with the optical axis as a mediator between the image coordinate system and the world coordinate system. By transforming measurements into this intermediate frame, the patent simplifies the relationship between camera parameters and pose estimates, enabling direct calculation without complex iterative solving.
2Productivity
If fast sampling rate is used for real-time pose estimation, then writing responsiveness is improved, but energy consumption and processing requirements increase
Solution Approach 1:
The patent employs computationally inexpensive algorithms that can be executed at high frequencies without significant energy cost. By using simple moment-based calculations for optical axis estimation and basic depth sensor readings for distance measurement, the system achieves fast sampling rates (suitable for real-time writing) while consuming minimal energy compared to complex iterative pose estimation methods.
3Device complexity
If simple measurement model is used, then computational effort is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent changes the parameterization of the pose estimation problem by expressing the optical axis direction in terms of image moment ratios (which are invariant to scale and translation) rather than traditional camera matrix parameters. This parameter transformation enables simple computation while preserving measurement precision, as the moment-based parameters directly capture the essential geometric relationships without requiring complex calibration data.
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
There is described a method of estimating a pose of an optical positioning device (OD) comprising an optical sensing system (OS) including an optical sensor (C) aimed at a target surface (S), the method including determining an estimation of a pose of the optical sensor (C) with respect to the target surface (S) based on optical measurements carried out by means of the optical sensing system (OS) and providing a corresponding estimation of the pose of the optical positioning device (OD) based at least in part on the estimation of the pose of the optical sensor (C). The estimation of the pose of the optical sensor (C) includes estimating positional parameters (AP, θ, d) of the optical sensor (C) that have entirely decorrelated measurement noise terms, namely (i) a position of an aiming point (AP) of the optical sensor (C) on the target surface (S) along an optical axis (AX) of the optical sensor (C), (ii) an orientation (θ) of the optical axis (AX) of the optical sensor (C) with respect to the target surface (S), and (iii) a distance (d) separating the aiming point (AP) and the optical sensor (C) along the optical axis (AX).