Odometry
Benchmarked Two Odometry Algorithms:
- Robust Visual Thermal Inertial Odometry (ROVTIO)
- Multi-Spectral Odometry (MSO)
Benchmarks with real-world data show these algorithms to be robust and reliable in the face of perceptual degradations such as smoke.
A sample trajectory captured (and the corresponding ground-truth) in the presence of smoke is shown below.
Dense Depth Estimation from Thermal Cameras
Multiple approaches to dense-depth estimation from thermal cameras re being investigated. Tradeoffs are being studied in order to select one algorithm.
Outputs from MAST3R are shown in the image below.
The algorithms being investigated are:
- MoGE
- MAST3R
Hardware Design
Mounts are being designed for the two thermal cameras, while keeping in mind the FOVof the thermal cameras and protection from crashes.
Our design, shown below, places the cameras within the convex hull of the drone’s body, thus making it safer during crashes as it will not be first to strike the ground. The design also keeps the propellers outside of the FOV of the cameras.
Logistics and Misc.
Two members of the team have obtained Part-107 drone licenses in order to pilot drones.
Arranged for testing sites at two locations:
- NREC Drone Cage
- Fire Academy