UAV Pixel Segmentation for GPS-Free Landing Space Positioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for positioning unmanned aerial vehicles (UAVs) using geofiducials are limited by visibility issues at varying altitudes, complexity in detecting and interpreting geofiducial content, and the requirement for sub-meter accuracy in geofiducial placement.
Innovation Solution
The use of machine learning models to process camera images and detect landing spaces by applying pixel labels, allowing the UAV to identify and navigate to unoccupied landing spaces efficiently, even in challenging environmental conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If geofiducials are used for positioning, then positioning accuracy is improved, but visibility and detection reliability deteriorate at varying altitudes
Solution Approach 1:
The patent changes the scale and resolution parameters of the visual features by using pixel-level segmentation instead of fixed-size geofiducials. This allows the system to maintain detection reliability across varying altitudes by adapting to different image resolutions and scales dynamically.
Solution Approach 2:
The patent segments the image into individual pixels and applies labels to each pixel, transforming the detection task from recognizing a single geofiducial object to analyzing multiple pixel-level features. This segmentation approach maintains positioning accuracy while improving detection reliability across different altitudes.
2Measurement precision
If geofiducials are used for positioning, then positioning capability is improved, but system complexity increases due to detection and interpretation requirements
Solution Approach 1:
The patent replaces the complex mechanical/visual interpretation system required for geofiducial detection with a machine learning-based pixel classification system. This substitution simplifies the detection process by automatically learning features from images without requiring manual geofiducial interpretation.
Solution Approach 2:
The system uses the UAV's own camera and onboard processing capabilities to perform positioning, eliminating the need for external geofiducial infrastructure. The pixel-labeling model automatically adapts to different environments without requiring pre-placed markers.
3Measurement precision
If geofiducials are used for positioning, then position determination is improved, but preparation time and effort increase due to sub-meter accuracy placement requirements
Solution Approach 1:
The system eliminates the need for manual geofiducial placement by using the UAV's camera to automatically identify and label pixels in the environment. The pixel-labeling model processes natural features without requiring human intervention for marker placement, thereby reducing preparation time while maintaining positioning accuracy.
Data Source
AI summary
In some embodiments, an unmanned aerial vehicle (UAV) is provided. The UAV comprises one or more processors; a camera; one or more propulsion devices; and a computer-readable medium having instructions stored thereon that, in response to execution by the one or more processors, cause the UAV to perform actions comprising: receiving at least one image captured by the camera; generating labels for pixels of the at least one image by providing the at least one image as input to a machine learning model; identifying one or more landing spaces in the at least one image based on the labels; determining a relative position of the UAV with respect to the one or more landing spaces; and transmitting signals to the one or more propulsion devices based on the relative position of the UAV with respect to the one or more landing spaces.