Marker Calibration Using First-Person and Bird's-Eye View Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for accurately calibrating the placement of markers in Mixed Reality applications, particularly in determining the position and orientation of imaging devices, suffer from low precision when the relative positional relation between markers is unknown, often relying on manual measurements or inaccurate position and orientation data.
Innovation Solution
A method and device for accurately calculating the position and orientation of markers on an object by using a combination of first-person view and bird's-eye view image processing, involving marker detection units, position-and-orientation calculating units, and calibration information calculation, to obtain precise marker placement information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual measurement or inaccurate position data is used for marker calibration, then the calibration process is simple, but the measurement precision of marker placement is low
Solution Approach 1:
The patent combines first-person view imaging (from the imaging device itself) and bird's-eye view imaging (from an external camera) into a unified calibration system. The first-person view provides accurate marker detection while the bird's-eye view provides precise spatial positioning, and their integration resolves the contradiction by achieving high measurement precision without requiring complex manual measurement procedures
Solution Approach 2:
The patent introduces a calibration information calculation unit as an intermediary that processes data from both imaging views. This intermediary component coordinates the information from first-person and bird's-eye views, enabling accurate marker placement calibration while maintaining system simplicity through automated computation rather than manual measurement
2Measurement precision
If multiple markers with known relative positional relations are used, then the calibration accuracy improves, but the ease of operation decreases due to complex setup requirements
Solution Approach 1:
The patent enables the system to automatically determine marker relative positional relations through image processing and coordinate transformation algorithms. Instead of requiring manual setup of markers with pre-known positions, the system self-calibrates by detecting marker positions in both first-person and bird's-eye views and computing the spatial relationships automatically, thus improving ease of operation while maintaining calibration accuracy
Solution Approach 2:
The patent transitions from two-dimensional marker pattern recognition to three-dimensional spatial coordinate transformation by incorporating bird's-eye view imaging. This dimensional change allows the system to accurately determine marker positions and relative relations in 3D space without requiring complex pre-arranged marker configurations, thereby improving both calibration accuracy and operational simplicity
Data Source
AI summary
An information processing device for obtaining placement information regarding an imaging device of markers to be calibrated that are provided upon the imaging device, includes: a first image obtaining unit adapted to obtain a first image photographed with the imaging device; a second image obtaining unit adapted to obtain a second image obtained by photographing the imaging device from a bird's-eye view position; a first detecting unit adapted to detect information relating to image coordinates of reference markers placed in a scene, from the first image; a second detecting unit adapted to detect information relating to image coordinates of markers to be calibrated from the second image; and a calibration information calculating unit adapted to obtain the calibration information using information relating to image coordinates of the reference markers detected by the first detecting unit and the second detecting unit.