🎵 SINGLE FILE DETECTION
Click to select audio
— or —
📡 3-MIC PRECISE LOCALIZATION
MIC 1 (origin)
No file
MIC 2 (0.2m,0)
No file
MIC 3 (0.1,0.17m)
No file
🚁🚁 MULTI-DRONE DETECTION
MIC 1
No file
MIC 2
No file
MIC 3
No file
🔊 NOISE ROBUSTNESS — SNR SWEEP
🛤️ PATH TRACKING SIMULATION
⚠️ Reliable zone ≤ 2.5m.
📂 MANUAL FILE + TIME WINDOW
📊 BATCH SCAN — DETECTION TIMELINE
Max 150 windows per scan (3000 when scanning the entire file) · scan_duration capped at 600s server-side
DETECTION LOG
No events yet…
Session Config
🔮 Simulated mode uses the exact same pipeline
as live deployment:
generate_synthetic_drone → localize_precise
→ PathTracker. Same thresholds, same cap-hit guard, same outlier rejection.
Noise level, spread and pattern control the acoustic environment.
🎙️ Live mic mode captures from your microphone,
runs the same
detect_from_single_audio + localize_precise
pipeline, and streams results via SocketIO. Requires PyAudio on the server.
Session: 0s | Avg conf: —
ACTIVE TRACKS (0)
No confirmed tracks yet
LAST DETECTION
Waiting…
GROUND-TRUTH vs ESTIMATE
Waiting for labelled segment…
SEGMENT LOG
No segments yet…
Repository Config
📁 DUNAKESZI RECORDING
Show: — Session: —
Array: — Drones: —
Time: — → —
Description: —
Phase: — Speed: — r=—
File: —
🗄️ Repository mode streams real recorded drone audio
directly from the online dataset ZIP via HTTP range-requests (
remotezip) —
no files are downloaded or stored on the server.
Ground-truth positions are overlaid when the segment has a label.
If the repository is unreachable, synthetic data is generated automatically using
the real BPF drone profiles so the dashboard stays live.
Session: 0s | Avg conf: —
ACTIVE TRACKS (0)
No confirmed tracks yet
LAST SEGMENT
Waiting…
LOCALISATION ERROR DIST.
No errors yet
🕘 DETECTION HISTORY
Click an entry to replay it in its original panel.
Loading…
⚙️ SYSTEM STATUS
API REFERENCE
GET /api/v2/status Health check
GET /api/v2/version Version + fixes
POST /api/v2/detect Single file detection
POST /api/v2/detect-3mic 3-mic localization
POST /api/v2/detect-multi Multi-drone detection
POST /api/v2/noise-test SNR sweep
POST /api/v2/path-simulate Path tracking
POST /api/v2/realtime/start Start RT session (mode: simulated|real|repository)
POST /api/v2/realtime/stop Stop RT session
GET /api/v2/realtime/status Session status
GET /api/v2/realtime/audio-devices List mic devices
WS realtime_frame {frame,detections,tracks,sim_positions,repo_label,mode}
WS realtime_stats {total_frames,detected_frames,detection_rate,avg_error_m}
WS realtime_status {running,mode,error,loading,final_stats}