LED Group Selection for Optical Tracking Pose Determination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for solving the initial correspondence problem in optical camera-based tracking systems are inefficient, requiring many frame captures and being unable to effectively establish initial correspondences between 3D points of a tracked object and its 2D projections without prior knowledge of the object's pose, leading to prolonged startup times and potential blind spots in tracking.
Innovation Solution
A method that selects and utilizes LED groups to establish one-to-one mappings between fiducial markers on a tracked object and their projections on the image plane by acquiring mesh data, pose data, and camera position, initializing uncovered pose sets, deriving LED groups, projecting these groups, and selecting the most visible ones to cover all poses, thereby reducing the number of frames needed for correspondence determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional methods are used to solve the initial correspondence problem, then correspondence determination can be achieved, but the process requires many frame captures and results in prolonged startup times
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing correspondence relationships between 3D fiducial markers and 2D image plane projections for multiple possible poses before actual tracking begins. When tracking starts, the system can immediately query these pre-computed correspondences rather than calculating them in real-time, dramatically reducing startup time while maintaining accuracy.
Solution Approach 2:
The system performs preliminary computation of projection matrices and correspondence mappings for a range of predetermined poses. This advance preparation creates a lookup table of correspondences that can be quickly accessed during operation, eliminating the need for time-consuming real-time calculations when the object first enters the tracking volume.
2Reliability
If conventional tracking methods are used, then continuous tracking can be maintained, but blind spots occur when the tracked object leaves and re-enters the camera's view
Solution Approach 1:
The patent eliminates blind spots by having pre-computed correspondence data ready for all possible poses before tracking begins. When the tracked object leaves and re-enters the camera view, the system can immediately resume tracking using the pre-prepared correspondence information without requiring re-initialization, thus maintaining continuous tracking reliability.
3Ease of operation
If initial correspondence problem is solved without prior pose knowledge, then tracking can start, but the process is inefficient and requires multiple frame captures
Solution Approach 1:
The system resolves the initialization efficiency problem by performing all correspondence calculations in advance during system setup. The pre-computed correspondence tables are stored and readily available when tracking needs to start, allowing immediate initialization without requiring multiple frame captures or iterative calculations, thus dramatically improving initialization speed while maintaining ease of operation.
Data Source
Figure 1
Figure 2~3a-2
Figure 3b-1~3c-2
AI summary
Amethod for selecting LED groups to be used for solving correspondence problem is presented, which uses pose data representing possible poses of tracked object under working conditions thus LED groups are selected to satisfy requirements.