Camera Orientation Tracking via Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for tracking an object's orientation using a video camera are limited by the need for fixed infrastructure, drift issues with inertial sensors, environmental disturbances affecting magnetic compasses, and requirements for clear celestial views and pre-mapped environments, especially when the camera is not fixed and the object lacks positional markers.
Innovation Solution
A system combining inertial orientation sensors with optical sensors, using sensor fusion algorithms to determine object orientation and position in a global reference frame, even without fixed infrastructure, by integrating data from acceleration, angular rotation, and magnetic field sensors, and employing a processor to estimate orientation using Kalman or complementary filters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If inertial sensors (gyroscopes) are used for orientation tracking, then tilt angles (pitch and roll) can be measured, but yaw angle drifts over time without compensation
Solution Approach 1:
The patent combines multiple sensing modalities (inertial sensors, magnetic compass, optical flow sensors, GPS) into a unified orientation estimation system. The inertial measurement unit provides high-frequency tilt data, while the magnetic compass and other sensors compensate for yaw drift, creating a redundant system that maintains accuracy over time through sensor fusion algorithms.
2Reliability
If magnetic compasses are used to supplement inertial measurement, then yaw angle can be compensated, but the system becomes sensitive to environmental disturbances and ferrous materials
Solution Approach 1:
The patent introduces multiple intermediary sensing mechanisms (optical flow sensors, GPS position data, visual feature tracking) that indirectly contribute to orientation estimation without being directly affected by magnetic disturbances. These intermediaries provide alternative references that can validate or correct magnetic compass readings when environmental conditions are unfavorable.
Solution Approach 2:
The system continuously monitors the quality and consistency of magnetic compass readings against data from other sensors. When disturbances are detected (through inconsistency with inertial predictions or optical flow data), the system automatically reduces reliance on the magnetic compass and increases weighting of alternative sensors, creating a dynamic feedback loop that adapts to environmental conditions.
3Adaptability or versatility
If optical recognition of the environment is used, then orientation can be determined without infrastructure, but the features must be mapped beforehand
Solution Approach 1:
The system performs preliminary action by continuously maintaining an updated model of the visual environment and feature relationships. Rather than requiring complete pre-mapping, the system learns and adapts to the environment in real-time, building a dynamic understanding of spatial features that can be used for immediate orientation determination while allowing for future adaptability to new environments.
4Measurement precision
If a fixed camera system with infrastructure is used, then accurate tracking is achieved, but the system requires additional infrastructure and loses flexibility
Solution Approach 1:
The system makes the camera self-sufficient by integrating all necessary sensing capabilities (inertial sensors, magnetic compass, optical flow) directly onto the moving platform. This eliminates the need for external fixed infrastructure while maintaining tracking accuracy through the combined data from multiple self-contained sensors that work together to provide robust orientation and position estimation.
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 approach provides accurate and reliable object tracking without the need for infrastructure, compensates for sensor drift, and maintains orientation estimation during periods of invalid optical solutions, enhancing tracking efficiency and robustness in dynamic environments.
Implementation Method 1
Inertial measurement of orientation using accelerometers and/or gyroscopes
Implementation Method 2
Magnetic compassing to supplement inertial measurement
Implementation Method 3
A sensor fusion algorithm, see, e.g., http:**en.wikipedia.org/wiki/Sensor_fusion, may be used
Implementation Method 4
employing a processor to estimate orientation using Kalman or complementary filters
Implementation Method 5
employing a processor to estimate orientation using Kalman or complementary filters
Data Source
AI summary
An object tracking system has an inertial orientation sensor attached to a camera. The sensor uses a rigid body position and orientation (with or without markers) visible to the camera for determining the orientation and the position of the object in the global reference frame, when the camera is not rigidly fixed. Another orientation sensor is attached to the object in order to keep tracking of its orientation when a valid tracking of the object cannot be obtained from the camera. The data from the orientation sensor attached to the object and the data from the orientation sensor attached to the camera is used to increase the accuracy of the optical tracking of the object.