Device Position Estimation Beyond Camera View Using Angular-Rate Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing device position estimation techniques fail when the device moves outside the view of the imaging apparatus, preventing accurate tracking and position estimation.
Innovation Solution
An information processing apparatus that estimates device position using a combination of image data and sensor data, including a photographed image acquisition unit, a sensor data acquisition unit, and estimation processing units to derive the device's position in three-dimensional space, even when the device is not within the camera's view, by utilizing the position of a predetermined part on the user's body as a rotation center.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If device position estimation is performed using only photographed image data, then measurement precision is improved when the device is visible, but reliability deteriorates when the device moves outside the camera's field of view
Solution Approach 1:
The patent combines image data from the camera with sensor data (acceleration and angular velocity) from the device to perform position estimation. This fusion of multiple data sources allows the system to maintain accurate tracking even when the device moves outside the camera's field of view, resolving the contradiction between measurement precision and reliability
Solution Approach 2:
The patent introduces sensor data as an intermediary element that bridges the gap when visual tracking is unavailable. The sensor data serves as a mediator to continue estimating device position and posture when the device is not captured in the photographed image, maintaining tracking continuity
2Reliability
If the imaging apparatus field of view is expanded to cover all possible device positions, then reliability of tracking is improved, but device complexity increases
Solution Approach 1:
The patent replaces the mechanical/optical approach of expanding camera field of view with a computational approach using sensor data fusion. Instead of adding more cameras or using wide-angle lenses (which would increase device complexity), the system uses acceleration and angular velocity sensors combined with algorithms to estimate device position beyond the camera's visible range
3Reliability
If sensor data is used to estimate device position when the device is not photographed, then reliability is improved, but measurement precision may deteriorate due to sensor drift
Solution Approach 1:
The patent uses feedback by continuously comparing sensor data with image data when the device is visible, and using the image data to correct sensor drift. The system switches between using image data (when available) and sensor data (when not available), with the ability to reset or correct sensor-based estimates when visual tracking is restored, maintaining both reliability and precision
Data Source
AI summary
A photographed image acquisition unit 212 acquires an image obtained by photographing a device. A sensor data acquisition unit 214 acquires sensor data indicating an angular speed of the device. A position and posture deriving unit 244 derives the position of the device in a three-dimensional space from a position coordinate of the device in the photographed image when the device is included in the photographed image. A part position estimation unit 246 estimates a position of a predetermined part in a body of a user on the basis of the estimated position of the device. A position and posture deriving unit 244 derives, as the position of the device, a position rotated by a rotation amount corresponding to the sensor data with the position of the part estimated by the position and posture deriving unit 244 used as a rotation center when the device is not included in the photographed image.