RaD-VIO: Rangefinder-aided Downward Visual-Inertial Odometry

TL;DR

RaD-VIO introduces a fast, dense downward visual-inertial odometry method that leverages the local planarity assumption, providing more robust and accurate MAV navigation compared to existing algorithms.

Contribution

The paper presents a novel dense, direct frame-to-frame visual-inertial odometry algorithm for downward facing cameras using homography-based photometric cost and IMU regularization.

Findings

Outperforms existing downward facing odometry algorithms in various scenarios.

Demonstrates robustness and accuracy in MAV flight conditions.

Provides real-time capable odometry with improved resilience.

Abstract

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).