Deformable 3D Model for 6DOF Pose Estimation in Aerial Refueling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current aerial refueling techniques face challenges in accurately and efficiently conducting refueling operations due to limitations in camera-based information processing.
Innovation Solution
A system and method that utilize a camera to generate 2D images, determine 6DOF pose of a receiving aircraft using semantic keypoints and perspective-n-point optimization, and output estimated positions for autonomous control and visual guidance during refueling operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If camera-based techniques are used for aerial refueling, then information for refueling operations is provided, but accuracy and efficiency are insufficient
Solution Approach 1:
The patent replaces traditional mechanical vision systems with a deformable 3D model-based computational approach. Instead of using complex mechanical cameras and optical systems to track receiver aircraft, the invention uses image processing algorithms that deform a 3D model to match 2D image features, achieving high precision position estimation through software-based geometric transformation rather than mechanical measurement systems
Solution Approach 2:
The patent changes the parameter representation from fixed 2D image coordinates to 6DOF pose parameters (position and orientation) through deformable model transformation. By parameterizing the receiver aircraft model with 6 degrees of freedom and using image deformation to match observed features, the system transforms raw image data into precise spatial parameters that directly enable accurate and efficient refueling operations
2Adaptability or versatility
If a predefined point model of a generalized version of the second device is used, then 2D keypoints can be determined, but variations in actual device geometry are not accounted for
Solution Approach 1:
The patent applies dynamics by making the 3D model deformable rather than static. The deformable model can dynamically adjust its geometry to match the actual receiver aircraft configuration, allowing the system to adapt to variations in device geometry while maintaining accurate keypoint determination. The model's flexibility enables it to conform to different aircraft shapes and configurations observed in the 2D images
Solution Approach 2:
The patent creates a deformable 3D copy of the receiver aircraft that can be transformed and matched to the actual aircraft in the 2D image. This virtual model copy is then deformed and positioned to align with observed features, enabling accurate keypoint determination even when the actual aircraft geometry varies from the generalized model. The copying approach allows the system to work with a flexible representation that can adapt to real-world variations
Data Source
AI summary
Disclosed herein are methods, systems, and aircraft for performing image analysis for aiding refueling operations. A method includes a method includes receiving a two-dimensional (2D) image from a camera of a first device, determining 2D keypoints of a second device located within the 2D image based on a predefined point model of a generalized version of the second device, determining a 6 degree-of-freedom (6DOF) pose of the second device using the 2D keypoints, and outputting an estimated position of at least a component of the second device based on the 6DOF pose.