Build Interactive Scenes from UAV Imagery
There are more and more UAVs in the air that have additional autonomy requiring less active control, and the choke point to greater use of these UAVs has become the human processing of the sensor data. With our Visualization Interface for Sensor Team Aggregation (VISTA), we take video streams from multiple physically distributed UAVs and fuse the data into a synthetic world-in-miniature. An operator can then move their eye-point around the world-in-miniature (WiM) arbitrarily to use the data independent of when or by which vehicle the data was collected. VISTA uses intelligent support tools to help the operator make sense of what is going on in the world and uses video selectively to show movement in the environment. The project is in collaboration with CMU-RI, who are providing state-of-the-art computer vision algorithms so that accurate UAV telemetry is not required for the construction of the world in miniature.
Computer vision algorithms fuse multiple UAV video streams into a single 3D world
An intuitive user interface allows the operator to view the data
AI support tools help the operator understand the world
Raw data can be accessed at any time for the operator to dig deeper into what is going on