Project to monitor and analyze loud sounds (ex. neighborhood roosters)
- automatically and accurately distinguish target sounds from background noise
- log data to create statistics about event frequency
- use data and statistics to generate trends and charts
- Raspberry Pi 3B+ (64-bit, more RAM is better)
- USB microphone
- Remote file storage or metrics aggregator endpoint (TIG stack?)
- python3, tensorflow
- Record sample file(s) containing target sound (audio.mp3)
- Label (labels.json) examples of target sound in sample file(s)
- Train the model and generate signatures
- Setup systemd service to monitor and log on boot
- v1: output positive identifications to daily log file (
wc -l
for count) - v2: create connector to push events to logging and analytics service
- v1: output positive identifications to daily log file (
- This project would not have been possible without the previous work and detailed instructions of Fabio Manganiello
├── datasets
│ └── sound-detect
│ ├── audio
│ │ └── train_sample_1
│ │ ├── audio.mp3
│ │ └── labels.json
│ ├── addl-samples
│ │ ├── 04-13-2022-23
│ │ ├── 04-13-2022-23.mp3
│ │ └── labels.json