Motor Anomaly Event Detection by Sound Analysis
Application #4
Published in
Jun 18, 2019
Description
- Detect motor anomaly events by sound analysis
- Status: Off, Normal, Stuck, Pre-Stalling, Stalling
- Record audio with raspberry pi and microphone array
- Use deep learning based for clustering
About Dataset
- Every clips are 0.5 second, sampling rate is 48000 Hz
- Clean background (3400 clips)
- With noise environment (Music 2880 clips, Factory 2356 clips)
Listen some samples
- Clean / Normal
- Music / Stalling
Label audio clips
Visualize in time frequency domain
Deep learning model
Run with LED lights
