Marker Position Correction Using Pre-calculated Relative Positions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Augmented reality systems face accuracy issues in determining the position and posture of terminal devices due to factors like large relative angles, small marker size in images, obstacles, and improper illumination, which deteriorate measurement accuracy.
Innovation Solution
An information processing device with a camera, processor, and memory that recognizes markers, calculates relative positions, and performs position correction using historical data and statistical methods to ensure accurate tracking even in deteriorated conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If marker-based tracking is used to determine position and orientation, then tracking functionality is provided, but measurement accuracy deteriorates under certain conditions
Solution Approach 1:
The system pre-calculates and stores multiple candidate relative positions between the terminal device and marker at different distances and angles before actual tracking. When the marker is detected, the system compares the actual relative position with these pre-calculated candidates to select the most accurate match, thereby maintaining measurement accuracy even when the terminal device is rotated or moved away from the marker.
Solution Approach 2:
The system changes the parameter of relative position representation from a single fixed value to multiple candidate values corresponding to different distances and orientations. By maintaining a set of pre-calculated relative positions with different parameters (distances, angles), the system can adapt to varying tracking conditions and select the appropriate parameter set that matches the current scenario.
2Area of stationary object
If the terminal device moves away from the marker or rotates at large angles, then tracking coverage area increases, but measurement accuracy deteriorates
Solution Approach 1:
The system pre-calculates relative positions at multiple distances and orientations before tracking begins. This allows the system to maintain accurate position determination even when the terminal device moves away from the marker or rotates at large angles, as the appropriate pre-calculated relative position is already available for selection.
Solution Approach 2:
The system extends the tracking capability from a single fixed relative position to multiple dimensions by pre-calculating and storing relative positions at different distances and angles. This multi-dimensional approach allows the terminal device to move freely within a larger coverage area while maintaining measurement accuracy through selective matching of pre-calculated positions.
3Measurement precision
If only current image data is used for position determination, then processing speed is fast, but accuracy deteriorates in deteriorated conditions
Solution Approach 1:
The system pre-calculates and stores multiple candidate relative positions before tracking begins. During actual tracking, it compares the detected marker position with these pre-stored candidates to determine the most accurate current position. This approach maintains high accuracy in deteriorated conditions while keeping processing relatively simple through direct comparison with pre-computed values.
Solution Approach 2:
The system uses feedback by comparing the actual detected marker position with multiple pre-calculated candidate positions. By selecting the candidate that best matches the current detection results, the system refines the position determination accuracy while managing processing complexity through efficient comparison operations.
Data Source
AI summary
An information processing device includes a memory that stores a known position of a marker, a camera, and a processor. The processor is configured to determine whether the image includes the marker, calculate a relative position of the marker with respect to the camera and store the calculated relative position in the memory if image includes the marker, and determine a position of the camera using the known position of the marker and two or more relative positions calculated for the marker stored in the memory.