Hybrid Motion Capture Combining Visual and Inertial Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional motion capture systems face challenges in accurately tracking and representing three-dimensional motion, especially when markers are obscured or when subtle changes in position and orientation are difficult to detect, leading to ambiguities and physically impossible motions in virtual representations.
Innovation Solution
Combining visual information from images with sensor information, such as inertial data, to improve the accuracy of motion tracking and reduce ambiguities by using inertial sensors attached to the subject, which provide motion information that can be synchronized and combined with visual data to determine the precise pose and orientation of objects in three-dimensional space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If visual information from images is used to track motion, then motion capture can be achieved, but ambiguities and physically impossible motions occur when markers are obscured or subtle changes are difficult to detect
Solution Approach 1:
The patent combines visual information from images with sensor information from inertial sensors to create a hybrid motion capture system. The visual data collector captures images of markers while the physical data collector gathers inertial sensor data, and a motion data combiner merges these data streams to determine accurate pose and orientation information, resolving ambiguities that occur when using visual information alone
Solution Approach 2:
The patent introduces an intermediary processing system that includes a motion data combiner and calibration module. This intermediary system calibrates the relationship between visual markers and sensor positions, then combines both data sources to produce reliable virtual representations, acting as a mediator that transforms raw visual and sensor data into accurate motion capture information
2Measurement precision
If inertial sensors are attached to the subject to provide motion information, then sensor data can be collected, but drift and calibration issues arise
Solution Approach 1:
The patent implements a feedback mechanism where visual information from tracked markers is used to correct and calibrate inertial sensor data. The system continuously compares sensor-derived pose information with visually-confirmed marker positions, using the visual feedback to correct drift and maintain accurate calibration of the inertial sensors throughout the motion capture process
Solution Approach 2:
The patent replaces purely mechanical or purely optical motion capture systems with a hybrid approach that substitutes some mechanical tracking (physical markers) with sensor-based measurement (inertial sensors), then uses visual feedback to maintain accuracy. This substitution allows for more robust motion capture while using visual information to correct the inherent drift of inertial sensors
3Measurement precision
If conventional motion capture systems are used with retro-reflective markers, then three-dimensional position can be determined, but the system becomes complex and requires multiple cameras
Solution Approach 1:
The patent makes the motion capture system more universal by enabling it to function with fewer cameras through the addition of inertial sensors. The system can operate in modes where visual information alone is used, where sensor information alone is used, or where both are combined, providing multi-functional capability that reduces the need for complex multi-camera arrangements while maintaining three-dimensional position accuracy
Data Source
AI summary
A system includes a visual data collector for collecting visual information from an image of one or more features of an object. The system also includes a physical data collector for collecting sensor information provided by at one or more sensors attached to the object. The system also includes a computer system that includes a motion data combiner for combining the visual information the sensor information. The motion data combiner is configured to determine the position of a representation of one or more of the feature in a virtual representation of the object from the combined visual information and sensor information. Various types of virtual representations may be provided from the combined information, for example, one or more poses (e.g., position and orientation) of the object may be represented.