Thermal Spoke Convergence for Aerial Refueling Drogue Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Highly skilled human operators are required for complex docking operations like air-to-air refueling and spacecraft docking, which are challenging to automate due to reliance on human judgment and difficulty in certifying artificial intelligence-based solutions.
Innovation Solution
A vision processor uses thermal image data to determine the center of an object, such as a drogue, by projecting vectors based on image gradients and populating an accumulator map to estimate the object's location, enabling autonomous connector mating without significant cost or complexity increase.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If human operators are used to guide complex docking operations, then the operations can be performed with human judgment, but the operations cannot be extended to autonomous vehicles and require highly skilled operators
Solution Approach 1:
The patent replaces complex stereoscopic vision systems and AI-based solutions with a simplified optical gradient-based center detection algorithm. By using gradient calculation on thermal image data to project vectors and populate an accumulator map, the system achieves autonomous docking without requiring highly skilled human operators or complex certification processes
Solution Approach 2:
The patent creates a simplified computational model (accumulator map) that replicates the human operator's center-detection function. Instead of using complex AI to mimic human judgment, the system uses gradient-based vector projection that mathematically converges on the object center, providing predictable and certifiable autonomous operation
2Measurement precision
If complex stereoscopic vision systems are used to aid human operators, then precision maneuvers can be achieved, but the systems are difficult to extend to autonomous vehicles and challenging to certify
Solution Approach 1:
The patent extracts only the essential function needed for docking (center detection) from complex vision systems. By using thermal image gradient analysis to project vectors from edge pixels toward the center, the system achieves precise center detection without the complexity of stereoscopic vision or AI-based solutions
Solution Approach 2:
The patent replaces complex optical systems with a computational approach using gradient-based vector projection. The accumulator map method provides precise center detection through mathematical convergence of projected vectors, eliminating the need for complex hardware while improving certifiability for autonomous operation
Data Source
AI summary
This disclosure provides systems, methods, and devices for performing thermal image-based identification, such as identification of a location of an object. In some aspects, a device includes a vision processor. The vision processor is configured to obtain, based on thermal image data representing an image that depicts at least a portion of an object, a location associated with a point-of-convergence. The point-of-convergence associated with the multiple structures of the object. The vision processor is also configured to identify the object based on the location.