UChicago Robotics Capstone December 2024
| Authors | Duncan Calvert, Joon Park, Zach Farahany, Mohammad Ayan Raheel |
| Package |  |
| Meta | License - MIT |
SkySearch is an open source Python package that allows simple, flexible, and robust DJI drone swarm capabilities tailored for search and rescue. Specifically, it provides a command center UI layer, flight controls, AI modeling extensibility, and speech-to-text for drone commands. SkySearch allows emergency personnel to search based on a single photo of a lost person as well as a textual description. This project was initialized as part of the University of Chicago's M.S. in Applied Data Science Capstone 2024.
- Ground Control Station (GCS): The GCS serves as the central hub for mission planning, monitoring, and flight control. It includes a user-friendly interface for defining search areas, setting flight parameters, and receiving real-time updates from the drone swarm.
- Drone Swarm: The swarm comprises of multiple UAVs equipped with high-resolution cameras, GPS, and onboard processing units. The drones communicate with each other and the GCS to coordinate their search patterns and share information.
- Image Recognition/Text-to-Image Matching: This module is responsible for generating images based on a textual description of the missing person or accepting uploaded photos. The module then compares these reference images against the real-time drone video feeds. Advanced computer vision techniques, such as convolutional neural networks (CNNs), YOLO, and FaceNet are employed for face detection and recognition.
Check back soon! SkySearch will be made available via PyPI and Streamlit Cloud in the next few months.
SkySearch is published under the MIT License
The SkySearch UI is meant to be a simple and intutitve experience for users of all skill levels.
- Users can cycle through the various drone swarm functions via the Settings menu on the left.
- To connect a drone, please follow the steps outlined in the Drone on Private python file