Internal Edge Verification for AR Object Pose Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing augmented reality systems face challenges in accurately detecting the pose of objects with simple profiles, such as symmetrical objects, as they often rely on contour edges alone, which can lead to ambiguity in differentiating multiple potential poses without additional features like textures.
Innovation Solution
The method involves using internal edges, not part of the contour, for object detection and pose estimation, generating and matching feature points from real and synthetic images to enhance tracking accuracy and performance, allowing for higher precision in object tracking without relying on textures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If contour edges alone are used for pose estimation, then the method is simple, but it is impossible to differentiate multiple potential poses for symmetrical objects
Solution Approach 1:
The patent segments the object's edge information into two distinct types: contour edges (outer boundary) and internal edges (edges inside the object boundary). This segmentation allows the system to use both simple contour matching and more discriminative internal edge matching, resolving the contradiction by adding internal edge analysis without completely abandoning the simple contour-based approach.
2Measurement precision
If textures are used to differentiate poses, then pose differentiation is possible, but the method becomes more complex and requires additional features
Solution Approach 1:
The patent introduces internal edges as an intermediary feature between contour edges and textures. Internal edges provide geometric structure information that is more discriminative than contours alone but does not require the complexity of texture analysis. This intermediary approach enables pose differentiation while maintaining methodological simplicity.
3Measurement precision
If internal edges are used for detection, then tracking accuracy is improved, but additional processing steps are required
Solution Approach 1:
The patent merges the contour edge detection process with internal edge detection into a unified edge analysis framework. By combining both edge types in the same processing pipeline and using them together for pose estimation, the system achieves higher accuracy while minimizing additional processing overhead through integrated rather than separate operations.
Data Source
AI summary
A method for one or more processors to implement includes acquiring a synthetic image of an object from a second orientation different from a first orientation, using a three-dimensional model. The method further includes identifying, in the synthetic image, second edge points that are located on an edge of the object that is not a perimeter. The method further includes identifying matched edge points, which are first edge points and second edge points at substantially a same location on the object. The method further includes storing the matched edged points in a memory that can be accessed by an object-tracking device, so that the object-tracking device can identify the object in a real environment by identifying the matched edge points in images of the object.