Marker-Based Pose Estimation Using Orientation Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current information processing technologies require increased computational resources to accurately estimate the position and posture of devices with multiple markers, leading to inefficiencies in virtual reality applications.
Innovation Solution
An information processing apparatus and method that acquires images of devices with markers, extracts marker coordinates, and uses sensor data to derive position and posture information by selecting and discarding candidate markers based on their orientation and similarity to three-dimensional model coordinates, reducing unnecessary calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of marker images is increased to improve estimation accuracy, then measurement precision is improved, but the amount of calculation increases
Solution Approach 1:
The system performs preliminary actions by using sensor data to predict the device posture before marker matching, and pre-filters candidate markers based on this prediction. This preliminary estimation allows the system to narrow down the search space of candidate markers, maintaining high estimation accuracy while significantly reducing the computational burden of comparing all possible marker combinations.
Solution Approach 2:
The marker matching process is segmented into multiple stages: first, candidate markers are selected based on sensor data prediction; second, these candidates are filtered based on image intensity characteristics; third, the remaining candidates are used for final position and posture estimation. This segmentation allows the system to process only relevant markers at each stage, reducing overall calculation while maintaining precision.
2Measurement precision
If comprehensive marker matching is performed to ensure accuracy, then measurement precision is improved, but processing time increases
Solution Approach 1:
The system performs preliminary posture prediction using sensor data before conducting detailed marker matching. This preliminary action creates an expected marker distribution that guides the subsequent matching process, allowing the system to quickly identify and focus on relevant candidate markers rather than exhaustively checking all possible combinations, thus reducing processing time while maintaining accuracy.
Solution Approach 2:
The system employs periodic filtering of candidate markers based on image intensity characteristics during the matching process. By periodically eliminating candidates that do not meet intensity criteria, the system maintains high estimation accuracy while significantly reducing the number of calculations required, thereby decreasing processing time.
Data Source
AI summary
Provided is an information processing apparatus including a captured image acquisition unit that acquires an image captured of a device including markers, an extraction unit that extracts a marker image coordinate in the captured image, a position and posture derivation unit that derives position information and posture information of the device from the extracted marker image coordinate and a three-dimensional coordinate of a candidate marker in a three-dimensional model of the device by performing a predetermined calculation, and a sensor data acquisition unit. The position and posture derivation unit places the three-dimensional model in a virtual three-dimensional space, and discards or selects the candidate marker to be used for the calculation, on the basis of a difference between an orientation of a surface on which the candidate marker is provided among surfaces of the three-dimensional model and an orientation of a screen surface of the captured image.