ForVis dataset enables evaluation of UAV navigation in forest canopies
ForVis: An In-Field Dataset and Benchmark for VIO Using Under-Canopy UAV Flights in Forests
Robotics
Summary
Flying drones in forests is tricky because trees, changing light, and vibrations make it hard for drones to know exactly where they are. The authors created ForVis, a collection of drone flights recorded under forests and open fields to test how well different navigation systems work in these tough conditions. They tried seven popular systems and found the choice of camera sensor mattered more than which software was used. ForVis helps people measure how fast and accurate drone navigation is when flying in forests.
What this means in practice
- •For drone developers: Test and improve UAV navigation software in realistic forest flight conditions using ForVis dataset and benchmarks.
- •For robotics engineers: Compare the effects of different camera sensors on visual-inertial SLAM accuracy in natural, cluttered outdoor environments.
Authors
Arman Kiani, Masoud Ataei, Elvis Gyaase, Jeffrey Eiyike, Aaron Weiskittel, Prabuddha Chakraborty, Vikas Dhiman
Abstract
Visual-inertial Simultaneous Localization and Mapping (VI-SLAM) for UAVs remains difficult to evaluate in real forest environments, where motion, illumination changes, repetitive vegetation, and vibration can all affect estimation. We present ForVis, an in-field dataset and benchmark for evaluating VI-SLAM during UAV flight in forest environments. The dataset contains twelve flights across open meadow, above-canopy, and under-canopy conditions in each environment. In total, it provides 563.8s of flight over 1096.8m of trajectory, recorded simultaneously with an Intel RealSense D435i and an OAK-D Pro Wide together with inertial and flight-controller data. We benchmark seven open-source VI-SLAM systems over 504 runs. The results show that sensor choice has a larger effect on trajectory error than the spread between algorithms: all seven methods achieve lower median error on the OAK-D Pro than on the D435i. ForVis is intended to support evaluation of speed, accuracy and robustness for VI-SLAM in challenging forest flight.