Naval vessel presence and port-activity monitoring
SAR backscatter and optical imagery reveal warship berth occupancy, submarine pen access, and underway movements even when AIS is silent. Cross-referencing transponder gaps against detected hull signatures is the core method.
Sensors
- Sentinel-1 SAR (IW mode): C-band (5.405 GHz), 10 m ground range resolution in Interferometric Wide Swath mode, 250 km swath. Revisit of 6 days at the equator for a single satellite, 3 days with both Sentinel-1A and 1B operational. Free and open; archive from 2014. Strong backscatter from metal hulls makes frigates and larger vessels reliably detectable, though vessels under roughly 20 m length approach the noise floor.
- Capella Space X-band SAR: X-band (9.65 GHz), spotlight mode delivers sub-0.5 m resolution products. Shorter wavelength than Sentinel-1 gives finer hull detail and better discrimination of vessel class by length and superstructure profile. Tasked commercially; latency from order to delivery can be under 12 hours depending on orbital geometry. Particularly useful for confirming submarine pen gate status and distinguishing destroyer from frigate by hull length.
- Planet Dove (PlanetScope): 3 m to 4.77 m native resolution multispectral optical, near-daily global revisit from a constellation of over 150 smallsats. Useful for port-activity context: pier crane positions, vehicle concentrations, fuel barge presence, and vessel colour or pennant markings visible at 3 m. Cloud cover is the hard constraint; optical is blind through overcast, which is where SAR takes over.
- Maxar WorldView-3: 30 cm panchromatic, 1.24 m multispectral. At this resolution, vessel class, pennant numbers, and deck equipment become legible. Revisit is irregular and tasking-dependent, typically 1 to 4.5 days depending on latitude and collection priority. Cost and tasking lead time make it best reserved for high-value confirmatory looks rather than routine monitoring.
Why metal hulls are loud in the radio spectrum
SAR systems transmit microwave pulses and measure the energy returned to the sensor. Steel and aluminium naval hulls are nearly perfect reflectors at C-band and X-band wavelengths. A frigate sitting at a pier returns a backscatter signal orders of magnitude brighter than the surrounding water surface, which scatters energy away from the sensor and appears dark. That contrast is the detection mechanism, and it does not depend on daylight, cloud cover, or the vessel's cooperation.
Hull geometry amplifies the effect further. Right-angle junctions between a ship's side and its deck, or between the hull and a pier, act as dihedral corner reflectors, bouncing energy directly back to the satellite. Larger vessels with more vertical metal surface area produce proportionally stronger returns. A nuclear-powered aircraft carrier is essentially unmissable in Sentinel-1 imagery. A coastal patrol boat under 20 m is a different problem: at 10 m resolution, it occupies perhaps two pixels and can be lost in sea clutter under moderate wave conditions.
Reading a port from orbit: berths, pens, and the absence of ships
Berth occupancy analysis compares a time series of SAR or optical scenes against a baseline map of pier geometry. When a vessel is present, a bright elongated return appears at a known berth. When it departs, the berth clears. The change is unambiguous. Accumulated over weeks, this produces an occupancy record: which hulls are in port, which are underway, and how often each berth turns over. High turnover at a fuel or ammunition pier tells a different story than a capital ship that has not moved in three months.
Submarine facilities add a layer of ambiguity. Covered pens, common in Soviet-era and some current Asian naval infrastructure, physically obscure the interior from optical sensors. SAR can sometimes detect a submarine's bow or stern protruding from a pen entrance, and the presence or absence of support vessels alongside the pen entrance is a reliable secondary indicator of activity. Gate position at a floating drydock is often legible in WorldView-3 imagery at 30 cm resolution. None of this is certain identification; it is probabilistic inference from multiple indicators read together.
AIS gaps are not silence, they are a signal
The Automatic Identification System broadcasts a vessel's identity, position, speed, and heading. Naval vessels are not required to transmit AIS, and many do not. Some commercial vessels operating in sensitive areas selectively disable their transponders. From space, the absence of an AIS return from a location where SAR detects a bright vessel-sized target is itself informative.
The cross-referencing method is straightforward in principle. A SAR detection at a given position and time, with no corresponding AIS transmission, flags a dark vessel. Vessel length estimated from the SAR return, combined with wake geometry and heading, narrows the class. A 150 m hull with no AIS in a strait used by tankers is a different concern from a 90 m hull with no AIS leaving a naval base at night. Neither is conclusive on its own. The honest limit is that SAR-derived length estimates carry an uncertainty of roughly plus or minus 10 to 15 percent depending on incidence angle and vessel orientation relative to the flight path, which can blur the boundary between adjacent ship classes.
Wake signatures and what they reveal about speed and heading
A vessel underway leaves a Kelvin wake, a V-shaped surface disturbance whose half-angle is approximately 19.5 degrees regardless of vessel speed, a consequence of deep-water wave physics first described by Lord Kelvin in 1887. SAR images this wake as a pair of bright lines diverging from the vessel's stern. The wake's length and contrast encode speed: a fast-moving vessel generates a more pronounced turbulent wake visible for tens of kilometres behind it in calm sea states.
Narrower internal waves within the Kelvin envelope, called transverse waves, can also be detected in SAR under the right conditions, and their spacing is related to vessel speed. This allows a rough speed estimate independent of AIS. Combined with heading derived from the wake axis, an analyst can reconstruct a vessel's course and extrapolate its destination, though atmospheric and oceanographic noise degrade the estimate significantly in rough seas or high wind conditions above roughly 10 to 12 metres per second wind speed.
Honest limits of the method
Revisit is the first constraint. Sentinel-1's 3-day repeat at mid-latitudes means a vessel can depart, transit hundreds of kilometres, and arrive at a new port between passes. Commercial SAR constellations such as Capella reduce this gap but at tasking cost. No current open-access SAR constellation provides sub-daily global maritime coverage.
Resolution sets the second constraint. At Sentinel-1's 10 m, vessels shorter than about 30 to 40 m are unreliable detections. Corvettes are detectable; rigid inflatable boats are not. X-band commercial SAR at sub-0.5 m resolves hull detail but covers a much smaller footprint per pass. Optical sensors are faster to interpret and give colour and marking information that SAR cannot, but a single overcast day removes them entirely from the picture. The practical answer is sensor fusion: SAR for all-weather presence detection, high-resolution optical for confirmatory identification when conditions allow.
Satellize runs multi-sensor fusion pipelines on open Sentinel-1 data combined with commercial tasking on client licence, applying the same change-detection logic it uses in agricultural contexts, such as the Kingdom of Tonga crop-estimation programme, to military infrastructure monitoring. The analytic outputs are structured for integration into existing intelligence workflows rather than delivered as raw imagery.
What a useful monitoring product actually looks like
A port-activity monitoring product is not a single image. It is a structured time series: a database of berth states, vessel detections with estimated lengths and headings, AIS correlation flags, and confidence scores attached to each detection. Delivered as a GIS layer or structured JSON feed, it allows an analyst to query the record rather than stare at imagery.
Change alerts, triggered when a previously occupied berth clears or a new vessel appears at a monitored facility, are more operationally useful than periodic reports for most clients. Alert latency is bounded by satellite revisit and processing time, not by analyst availability. For high-priority facilities, stacking Sentinel-1 passes with commercial SAR tasking can reduce detection latency to under 24 hours on most days, with the caveat that orbital geometry occasionally produces longer gaps at specific latitudes.
Typical figures
| SAR spatial resolution (Sentinel-1 IW) | 10 m ground range, 20 m azimuth |
| SAR spatial resolution (Capella Space spotlight) | Sub-0.5 m (published product specification) |
| Optical resolution (Planet Dove) | 3.0 to 4.77 m multispectral |
| Optical resolution (WorldView-3) | 0.30 m panchromatic, 1.24 m multispectral |
| Revisit (Sentinel-1, dual satellite) | 3 days at mid-latitudes; 1 to 2 days at high latitudes |
| Revisit (Planet Dove optical) | Near-daily global; cloud-limited |
| Minimum detectable vessel (Sentinel-1) | Approximately 30 to 40 m hull length in calm sea states; smaller vessels unreliable |
| SAR frequency bands used | C-band 5.405 GHz (Sentinel-1); X-band ~9.65 GHz (Capella) |
| Archive depth (Sentinel-1) | From April 2014 (Sentinel-1A launch) |
| Alert latency (stacked SAR plus commercial tasking) | Under 24 hours for most facilities on most days; orbital geometry can extend this |
Analytics Satellize can run
| Berth occupancy time series | SAR bright-target detection against baseline pier geometry; change flagged when target appears or clears | GIS polygon layer with occupancy state per berth per pass, timestamped |
| Dark vessel detection and AIS gap flagging | SAR vessel detection cross-referenced against AIS position records; unmatched detections flagged by position, estimated length, and heading | Structured alert feed (JSON or GIS point layer) with confidence score and sensor source |
| Vessel length and class estimation | SAR-derived hull length measurement from bright-target extent; comparison against published vessel-class length ranges | Classification report with length estimate, uncertainty range, and candidate class list |
| Wake-derived speed and heading reconstruction | Kelvin wake axis extraction from SAR imagery; transverse wave spacing analysis for speed estimation | Tabular record of estimated speed (knots) and heading (degrees) per detected underway vessel |
| Submarine pen activity indicator | Multi-source fusion: SAR detection of protruding hull sections, optical observation of support vessel presence, gate-state reading from high-resolution optical | Activity-state assessment (active indicators present / no indicators / obscured) with supporting imagery chips |
| Port throughput and operational tempo trend | Aggregated berth-occupancy and vessel-movement records over rolling time windows; statistical comparison against historical baseline | Monthly trend report with annotated time-series charts and anomaly flags |
Who does the work
We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.