DroneVis: Versatile Computer Vision Library for Drones
Ahmed Heakl, Fatma Youssef, Victor Parque, Walid Gomaa
TL;DR
DroneVis presents a versatile Python library that automates a broad set of computer vision tasks on Parrot drones, integrating state-of-the-art models for detection, tracking, segmentation, depth, pose, face detection, crowd counting, action recognition, and more. It emphasizes usability through multiple user interfaces (GUI, CLI, gesture) and a friendly API, complemented by thorough documentation and high test coverage. The framework provides practical demonstrations on a Parrot drone, compares with related software, and offers data-driven recommendations for default models to balance accuracy and speed. The work enables researchers and practitioners to deploy diverse CV pipelines on drones with real-time capabilities and extensibility, paving the way for autonomous navigation and advanced mission planning. The long-term vision includes adding localization to support autonomous drone navigation and more sophisticated mission-level autonomy.
Abstract
This paper introduces DroneVis, a novel library designed to automate computer vision algorithms on Parrot drones. DroneVis offers a versatile set of features and provides a diverse range of computer vision tasks along with a variety of models to choose from. Implemented in Python, the library adheres to high-quality code standards, facilitating effortless customization and feature expansion according to user requirements. In addition, comprehensive documentation is provided, encompassing usage guidelines and illustrative use cases. Our documentation, code, and examples are available in https://github.com/ahmedheakl/drone-vis.