From satellite
to seabed.
NAVBathy is Darkocean AI's platform for full-depth, AI-native bathymetry — built for hydrographers, surveyors, and coastal engineers who need seabed data they can trust and defend.
What NAVBathy Does
NAVBathy combines satellite-derived depth for the optically shallow zone with explainable multibeam (MBES) processing for the vessel-reachable deep — one continuous seabed picture instead of disconnected tools.
The shallow-water layer derives depth from Sentinel-2 imagery and is calibrated against open ground-truth sources such as NASA's ICESat-2. The deep-water layer ingests multibeam data and wraps established, proven tools — MB-System for ingest, CUBE for gridding — rather than reinventing them. Both layers, plus NAVTopo satellite topography and ocean forecasting, feed a single shared surface model that carries one IHO S-44 Ed 6.1.0 compliance assessment, the standard hydrographic offices use to certify survey data for navigation.
Automated cleaning is explainable and human-in-the-loop end to end: every decision is logged, and possible real seabed features are flagged for a human reviewer rather than silently discarded — because getting navigation-critical data wrong has real consequences.
Our Values
The principles that guide how NAVBathy is built
Explainable, Human-in-the-Loop
Every cleaning decision is logged and reviewable. The system flags possible real features for verification instead of silently deleting them — safety of navigation comes before automation convenience.
Open Data & Standards
Built on open scientific data — ESA Sentinel-2, NASA ICESat-2, and Copernicus — and evaluated against the IHO S-44 Ed 6.1.0 standard for hydrographic survey uncertainty.
One Shared Surface
Satellite-derived depth and multibeam processing both emit the same uncertainty-surface contract, so every source feeds a single S-44 compliance assessment rather than a patchwork of one-off outputs.
See NAVBathy in action
Whether you're evaluating the platform for a survey program or want to learn more about how it works, we'd love to hear from you.