Process for recovering an unmanned vehicle
Inventors
Anderson, Michael L. • Wilson, Charles B. • Hyde, Michael A.
Assignees
United States Department of the Air Force
Publication Number
US-9911059-B1
Publication Date
2018-03-06
Expiration Date
2036-08-18
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
A process for recovering a vehicle includes obtaining a red green blue (RGB) image comprising a target on a recovery device. An input received from a user designates a target hue value and a target luminance value. The RGB image is converted to a hue value saturation (HSV) color model. The HSV color model is split into a hue value plane and a luminance value plane. A hue band pass filter and a luminance band pass filter are configured with appropriate thresholds. The configured hue band pass filter and the luminance band pass filter are applied to the hue value plane and the luminance value plane, respectively. The filtered hue value plane and the filtered luminance value planes are combined to yield a plurality of potential target pixel groupings. The most probable target is determined from the plurality of potential target pixels. The vehicle is directed to the target.
Core Innovation
The invention provides a process for recovering an unmanned vehicle by obtaining an RGB image that includes a target on a recovery device, converting the image into an HSV color model, and applying hue and luminance band pass filters to identify potential target pixel groupings. From these groupings, the most probable target is determined, and the vehicle is directed to the center of that target. The process involves user input to designate target hue and luminance values and utilizes image processing to enhance target recognition and vehicle guidance.
The problem addressed is the difficulty in recovering UAVs, especially when landing on a moving vehicle, which involves complex three-dimensional dynamics that can exceed operator skill. Current guidance methods such as GPS or IMU are insufficiently precise or fast for recovery at sea. Highly trained operators are required, and failed recoveries are costly and dangerous. Thus, there is a need for improved, automated recovery methods for UAVs that reduce operator skill requirements and increase reliability.
Claims Coverage
The patent includes several independent claims that describe a process for vehicle recovery using image processing and vehicle guidance based on filtered target detection.
Process for recovering a vehicle using filtered image processing and user input
Obtaining an RGB image from a camera coupled to the vehicle; receiving user input to designate a target hue and luminance value; converting the image to HSV color model and splitting it into hue and luminance planes; configuring and applying band pass filters on these planes; combining filtered planes to yield potential target pixel groupings; determining the most probable target; and directing the vehicle to the target's center.
Applying a culling filter to exclude small pixel groupings
Configuring a culling filter with a threshold minimum number of pixels and applying it to potential target groupings to exclude those below the threshold.
Directing the vehicle to the target using centroid calculation and autopilot navigation
Calculating the centroid of the most probable target, establishing bearing and azimuth in a vehicle-centric coordinate system, and instructing the vehicle's autopilot to maneuver toward the target.
Displaying centroid and bounding box for user confirmation
Displaying the centroid crosshair and bounding box of the most probable target on a display for user confirmation.
Configuring hue band pass filter thresholds and wrapping
Utilizing a hue band pass filter with thresholds approximately between 10% and 20%, including wrapping around hue values of 0 or 1 to capture red-pink and red-orange hues.
Configuring luminance band pass filter thresholds
Applying a luminance band pass filter with thresholds approximately between 10% and 30%, preferably around 20%.
User input via pixel selection and adjacent pixel averaging
Receiving user input by clicking a desired pixel on a display and applying adjacent pixel averaging filters to reduce mischaracterization of hue or luminance values.
These inventive features collectively provide a robust and efficient method for automated vehicle recovery by filtering image data based on user-designated color parameters, refining target detection, and guiding the vehicle using calculated coordinates and autopilot control.
Stated Advantages
Does not require highly skilled operators for vehicle recovery.
More reliable than operator-guided recovery, reducing crashes and damage.
Highly robust to varying lighting conditions during target tracking.
Capable of tracking moving targets especially in maritime environments.
Fast and efficient image processing allowing near real-time guidance.
User can change target tracking parameters in real time to improve accuracy.
Low-cost retrofitting of existing recovery devices with distinctive visual features improves target recognition.
Documented Applications
Military missions including surveillance, reconnaissance, target acquisition, data acquisition, communications relay, decoy, harassment, or supply flights.
Civilian applications such as firefighting, natural disaster reconnaissance, police observation of civil disturbances or crime scenes, scientific research (weather and volcano observation).
Recovery of unmanned aerial vehicles (UAVs) onto ships at sea using recovery nets or arresting cable devices.
Commercial fishing vessels using UAVs to scan for schools of fish and perform automated recovery at sea.
Interested in licensing this patent?