How trimodaldataset/multi-modal-drone-data works
This repository is a small multi-sensor data-collection application, not a drone-detection pipeline. Its runtime entry point starts separate audio, video, and RF recorders; the RF recorder signals readiness before the audio and video scripts begin. The scripts save modality-specific recordings. The README describes a broader synchronized drone/background dataset, but the shown code records background data and does not implement dataset labeling or synchronization. Relationships below follow the excerpts; broader dataset capabilities are documented, not shown as runtime components.
Main entry points: main.py Capture controller · iq_recording.py RF recorder · audio_recording.py Audio recorder · video_recording.py Video recorder
Components 10
Capture orchestration
- Capture controller
main.py - Ready signal file
iq_recording.py
Sensor capture
- Audio recorder
audio_recording.py - Video recorder
video_recording.py - RF recorder
iq_recording.py - USRP sensor
Recording outputs
- WAV recordings
audio_recording.py - MP4 recordings
video_recording.py - I/Q binary recording
iq_recording.py
Other components
- Operator
Connections 12
- Operator → Capture controller: starts capture
main.py - Capture controller → Audio recorder: launches process
main.py - Capture controller → Video recorder: launches process
main.py - Capture controller → RF recorder: launches process
main.py - RF recorder → USRP sensor: receives I/Q
iq_recording.py - RF recorder → I/Q binary recording: writes samples
iq_recording.py - RF recorder → Ready signal file: creates signal
iq_recording.py - Audio recorder → Ready signal file: waits for file
audio_recording.py - Video recorder → Ready signal file: waits for file
video_recording.py - Capture controller → Ready signal file: removes file
main.py - Audio recorder → WAV recordings: writes segments
audio_recording.py - Video recorder → MP4 recordings: writes segments
video_recording.py