RaD-VIO: Rangefinder-aided Downward Visual-Inertial Odometry
This work addresses robust odometry for Micro Aerial Vehicles in scenarios where forward-facing methods fail, such as during initialization or unobservable motions, but it is incremental as it builds on existing downward-facing approaches.
The authors tackled the brittleness of forward-facing monocular visual-inertial odometry by developing RaD-VIO, a fast, dense, and direct algorithm for downward-facing cameras that uses a homography-based photometric cost and IMU regularization. They demonstrated superior performance over existing state-of-the-art downward-facing odometry algorithms for Micro Aerial Vehicles in various scenarios.
State-of-the-art forward facing monocular visual-inertial odometry algorithms are often brittle in practice, especially whilst dealing with initialisation and motion in directions that render the state unobservable. In such cases having a reliable complementary odometry algorithm enables robust and resilient flight. Using the common local planarity assumption, we present a fast, dense, and direct frame-to-frame visual-inertial odometry algorithm for downward facing cameras that minimises a joint cost function involving a homography based photometric cost and an IMU regularisation term. Via extensive evaluation in a variety of scenarios we demonstrate superior performance than existing state-of-the-art downward facing odometry algorithms for Micro Aerial Vehicles (MAVs).