VESPER / ACOUSTIC IDENTITY CONCEPT
SOFAR AI / MARITIME ACOUSTIC INTELLIGENCE
Ocean
is not silent.
We turn underwater sound
into maritime intelligence.
01 / ACOUSTIC IDENTITY
Identity
in the noise.
Every vessel leaves an acoustic signature.
We are building the intelligence to tell them apart.
02 / THE RAW SIGNAL
It starts with
a trace.
Propulsion. Machinery. Movement. A hydrophone captures the patterns a vessel leaves in the water.
The fingerprint is a visual metaphor. This spectral replay comes from a hydrophone recording.
From presence to identityAbout this signal
A recorded hydrophone sample, not a live feed or VESPER identification result. Channel 0; 4,096-sample Hann FFT; 160 logarithmic frequency bins. Line height is normalized spectral magnitude, not calibrated sound pressure. Playback uses the same channel, converted to mono 16-bit PCM.
Adapted from PORTIA v2 by Anaida Fernández García, Alberto Belmonte Hernández, Jose Miguel Higon, Rafael Company, Francisco Jose Perez Carrasco, Federico Alvarez Garcia and Álvaro Gutiérrez. CC BY 4.0.
03 / TWO DIFFERENT QUESTIONS
Something is there.
But what makes it distinct?
Ocean Sentinel
Acoustic classification and maritime context turn sound into a vessel detection.
Explore the system ↗Research into recognizing individual vessels across recordings and conditions.
IN DEVELOPMENT / SOFAR AI04 / THE ACOUSTIC FOUNDATION
From sound
to a vessel detection.
Ocean Sentinel transforms recordings into spectral features, scores vessel presence and checks the result against the local conditions.
- 01Audio → spectrogram
- 02Spectral features → classification
- 03Detection → maritime context
05 / EVIDENCE, THEN ACTION
Evidence in view.
The operator in control.
A detection is a starting point. The acoustic evidence, confidence and maritime context stay together for an operator to inspect.
Detection today. Research into individual acoustic identity next.