Real-Time Object Tracking for Drones
Key Technologies: Deep Learning, OpenCV, NVIDIA Jetson, Docker, ROS2, C++, Python, Selenium
Vermeer needed a portable computer vision system that allowed for object detection and tracking aboard a drone. The algorithm needed to run at the edge in real-time to support future autonomous behaviors.
Adinkra developed and deployed an advanced real-time perception system, retrained for an aerial perspective and built on ROS2. We validated our system in the real-world, integrating the vision system into an intuitive UI for mission monitoring.
Vermeer developed over $1M in new business as a direct result of this product demonstration. They invited us to work together on this newly won work to continue expanding their robotics capabilities.