Capabilities
Discussion of some of the current and possible capabilities of the F-11.
Last updated
Discussion of some of the current and possible capabilities of the F-11.
Last updated
Current drones, such as the Flyby F-11, rely on GPS information to hold their position and battle against flight drift. This becomes increasingly more complex in environments that deny GPS, including many restricted civilian areas and modern battlefields, or areas with severely limited GPS capabilities. Many groups and organizations have become increasingly interested in navigation through environments with limited GPS availability.
Flyby Robotics has partnered with Palantir to showcase the system capabilities of the F-11 in harsh, GPS-denied environments. Palantir's Visual Navigation (VNAV) builds on the F-11 to navigate through such environements, combining the power of their geospatial algorithms, computer vision, machine learning, and the F-11's onboard Nvidia Jetson to keep the drone's flight on track.
Machine learning is at the cutting edge of many industries, and this revolution is largely enabled by the efficiency and accuracy of newer computer vision models. Object detection models such as YOLOv7 are at the backbone of many advancements. Mounted with an Nvidia Jetson GPU, our F-11 drones are capable of pushing the limits of current AI stacks and enable the next generation of aerial workloads.
Overhead identification and analysis is an active area of development, and Stephan Sturges has shown that WALDO is producing incredible results. Any such framework can be easily installed on an F-11, quickly enabling aerial object detection.
Paired with the Gremsy Vio, object detection can be further extended for thermal object identification. As HOT has shown, this has proved to have promising results.
A common use case for drones today is photogrammetry, and the F-11 supports these missions. With our high quality camera and gimbal integrations, as well as our upcoming RTK + geotagging features, mapping missions are simple and effective.
Inspection of oil and gas pipelines is notoriously dangerous, and drone technologies are substantially reducing this risk. Armed optical gas imaging cameras, drones can help with gas leak detection and repair. The same technology can be used for gas leaks and emissions, as can be seen in the following image of Turkmenistan.
ExxonMobil is using multiple aerial surveillance technologies to pinpoint and measure leaks, including LiDAR and gas sensors.
Paired with a thermal camera, our drones can be used for search and rescue missions that may be too challenging or dangerous for humans. With our Gremsy Vio integration, drone operators can explore harsh environments effectively and safely. Paired with the power of the onboard Jetson, these missions can be enhanced with computer vision and machine learning to automatically identify and track those in need.
Researchers have published a case study of using UAV that are coupled with machine learning for search and rescue missions that had a 94.73% accuracy, while being portable, cost-effective, and fast.
Visual-Based Person Detection for Search-and-Rescue with UAS: Humans vs. Machine Learning Algorithm.
Flying overhead, our F-11 drones can easily patrol areas and look for hazards. These tasks can be automated with machine learning, and can even cover an enormous area with drone swarm technologies. Items such as weapons or narcotics can also be automatically identified and can notify operators without risking any dangerous confrontation.
Firearm detection using computer vision is easily accomplished with YOLOv7, as was demonstrated here.
This method can also assist beach lifeguards cover a larger area with more precision. With an average of 10 fatal beach drownings every day, even a small amount of drone patrols can identify situations quicker and help first responders save lives.
Drowning-Detector is a computer vision tool for identifying individuals who are at risk of drowning, built on top of OpenCV and PyTorch and uses the You Only Look Once (YOLO) object detection algorithm.
Researchers at the Sri Lanka Institute of Information Technology published a paper for another drowning detection system in the IEEE International Conference on Advancements in Computing (ICAC).
Our drones can be used for field surveying and livestock tracking which can help increase agriculture efficiency and lower energy costs. Computer vision applications can be used to automatically track crop yields and identify problems in their infancy.
Viso published a comprehensive analysis of computer vision in agriculture, in which they highlight some applications of computer vision to agriculture. A Flyby F-11, together with the onboard Nvidia Jetson Orin NX, is fully capable of deploying these models in the sky.
Researchers have more applications such as
Drones can operate together synchronously to form swarms, which seems to be at the forefront of many military objectives. Drone swarms can be either centralized or decentralized, both of which may have different benefits.
Drone swarms can be used for many applications, including agriculture and crop cultivation, fighting wildfires, border patrol, surveillance, and military.
A 2023 paper published in Nature utilize real-time particle swarm optimization techniques for drone swarms to detect and track occluded targets.
Another paper surveys UAV swarm and asses that swarm technology is a fundamental future agenda.
As supporting swarm technology is one of our primary interests to enable, we are eager to work with teams building towards this goal.
Paired with the Sony LR1, which boasts an enormous effective 61MP full-frame sensor, our system enables for stunning photography with exceptional ease of use. Our integration supports all major functionality that is required for such a workload.
