Naval mine-laying pattern detection from SAR and optical
SAR and optical imagery can reveal mine-laying activity through vessel track geometry, speed profiles, and ship-type confirmation, even when transponders are off. The method identifies suspicious behaviour; it cannot confirm mine presence or density.
Sensors
- Sentinel-1 SAR IW (C-band, ESA): 10-metre ground range resolution in Interferometric Wide swath mode, 250 km swath, 6-day repeat at the equator and roughly 2-3 days at higher latitudes. Provides all-weather, day-night vessel detection and wake imaging. Free archive from 2014.
- Capella Space X-band SAR: Spotlight modes down to approximately 0.5-metre resolution, with tasking latency of hours rather than days. X-band is more sensitive to small metallic objects and fine wake structure than C-band, useful for confirming vessel class at close range.
- Planet Dove optical: 3-5 metre resolution, daily global revisit in the constellation aggregate. Confirms vessel type, deck configuration and hull markings in clear conditions. Cloud and night are hard limits; Dove cannot image through either.
- Maxar WorldView-3 optical: 30 cm panchromatic resolution, the sharpest routinely available commercial optical. Resolves hull details sufficient to distinguish a dedicated minelayer from a general auxiliary vessel. Tasked on demand; archive depth depends on prior collection over the area of interest.
What a lawnmower track gives away
Mine-laying from surface vessels follows a constrained operational logic. A ship must move slowly, typically under 5 knots, to deploy mines safely and accurately. It must cover a defined area systematically, which produces the characteristic parallel-pass geometry that analysts call a lawnmower pattern. In open water, that geometry is distinctive. In a SAR image, the vessel itself appears as a bright point target; its wake, if conditions allow, can persist for minutes and appear as a faint linear signature trailing behind. A sequence of passes produces a set of near-parallel wakes spaced at intervals consistent with the mine spacing the operator intends.
This is not a theoretical detection method. During the period following Russia's February 2022 invasion of Ukraine, open-source analysts including Bellingcat and various naval-intelligence commentators applied exactly this methodology to Sentinel-1 imagery of the north-western Black Sea. Russian naval auxiliaries were observed making slow transits in shallow coastal waters off Odesa and the Crimean coast, with track geometries inconsistent with normal transit or patrol behaviour. The analysis was publicly documented and widely cited in the defence-press record.
Why SAR is the primary sensor, and what it cannot do
SAR works day and night and through cloud. In a contested littoral environment where optical coverage may be denied by weather for days at a time, this matters enormously. Sentinel-1's 6-day revisit is the baseline; commercial X-band SAR from Capella can be tasked within hours of a request, providing the temporal density needed to reconstruct a vessel's track across multiple passes.
The honest limits are significant. SAR detects the vessel and, in favourable sea-state conditions, its wake. It does not image the mines themselves. A mine is a small object, typically under a metre in its largest dimension, sitting on or below the surface. No unclassified SAR system has a published detection capability for individual moored or bottom mines from orbit. What SAR provides is behavioural evidence: the ship was here, moving at this speed, on this heading, repeatedly. The inference that mine-laying occurred is probabilistic, not confirmatory. Sea clutter at low wind speeds can obscure wakes; high sea states produce false wake-like features. Analysts must account for both.
Optical confirmation of vessel type
Knowing that a vessel was behaving suspiciously is more actionable when you can also confirm what type of vessel it was. Dedicated minelayers are a distinct ship class. Russia operates the Alexandrit-class (Project 12700) and older Natya-class vessels, among others, with identifiable hull forms, deck equipment and length-to-beam ratios. At WorldView-3's 30 cm resolution, these characteristics are resolvable. Planet Dove at 3-5 metres can confirm vessel presence and approximate size; it cannot reliably resolve deck equipment.
The fusion workflow is straightforward in principle. SAR identifies the behavioural anomaly and fixes the vessel's position. A tasked optical collect, if weather permits, confirms or refutes the vessel-type hypothesis. Where optical is unavailable, analysts fall back on AIS history, known vessel movements from public maritime databases such as MarineTraffic, and any available signals intelligence in the open record. The method is weaker without optical confirmation, but it is not blind.
Speed, depth, and the geometry of suspicion
Coastal water depth is a critical contextual variable. Contact mines are typically moored in water between 5 and 30 metres deep; bottom-influence mines can be laid in somewhat deeper water. Bathymetric charts, many of which are publicly available through hydrographic offices, allow an analyst to overlay a vessel's track against depth contours. A slow back-and-forth transit that sits consistently within the 10-30 metre depth band is far more suspicious than the same track in 100 metres of water.
Speed thresholds matter too. Normal transit speed for a vessel of minelayer size is typically 10-15 knots. Sustained speeds below 5 knots, particularly when combined with the lawnmower geometry and the correct depth band, form a convergent indicator set. No single indicator is conclusive. The argument is cumulative: wrong speed, wrong pattern, wrong depth, known vessel class. Four independent signals pointing the same direction is a different analytical product from one.
Archive depth and the retrospective question
One underappreciated capability of open SAR archives is the ability to ask retrospective questions. Sentinel-1 data is freely available from late 2014 onward. If a contested area becomes strategically significant, analysts can pull the full archive and reconstruct vessel behaviour over months or years, not just the period after the crisis began. This is how the Black Sea analysis was able to establish baseline patterns of normal vessel movement before the invasion, making the post-February 2022 anomalies statistically distinguishable rather than merely asserted.
Satellize's analytics infrastructure runs on open constellations including Sentinel-1 and adds commercial tasking on client licence, which means a government client can commission a retrospective track analysis alongside a current-operations monitoring feed from a single contract. The Overhead column has covered the methodological principles of this kind of archive-dive publicly.
The archive is not unlimited in utility. Sentinel-1 does not collect everywhere at maximum frequency; coverage gaps exist, particularly over ocean areas far from European ground stations during the early mission years. Analysts should verify actual acquisition density over their area of interest before drawing conclusions from apparent absence of evidence.
What this method produces, and what it does not
The output of a well-executed mine-laying pattern analysis is a track reconstruction with associated confidence assessments: high confidence where SAR, optical, AIS and bathymetry all converge; lower confidence where only one or two indicators are present. It can support a legal or political case that mine-laying activity occurred in a given area during a given period. It cannot tell a naval commander where the mines are, how many there are, or what type they are.
Mine clearance requires dedicated sonar, remotely operated vehicles, and in-water survey. Satellite data can define the search area and prioritise sectors by probability of mine presence, but it hands off to those systems at the water's edge. Analysts who overstate what the imagery shows do a disservice to the commanders who rely on it. The honest product is a probability map and a documented evidentiary record, not a minefield chart.
Typical figures
| SAR spatial resolution (Sentinel-1 IW) | 10 m ground range x 10 m azimuth (detected); 5 x 20 m single-look complex |
| SAR spatial resolution (Capella X-band spotlight) | ~0.5 m (published spotlight mode) |
| Optical resolution (WorldView-3) | 0.30 m panchromatic, 1.24 m multispectral |
| Optical resolution (Planet Dove) | 3-5 m multispectral |
| Sentinel-1 revisit (mid-latitudes) | 6 days single satellite; 2-3 days with constellation overlap |
| Minimum detectable vessel (SAR) | ~10 m length in low sea-state; larger targets more reliable |
| Sentinel-1 archive depth | From late 2014; free via Copernicus Data Space |
| Frequency bands used | C-band (5.4 GHz, Sentinel-1); X-band (~9.6 GHz, Capella) |
| Swath width (Sentinel-1 IW) | 250 km |
| Typical analysis latency (tasked commercial SAR) | 2-6 hours from tasking to image delivery (Capella published figures) |
Analytics Satellize can run
| Vessel track reconstruction | Multi-pass SAR vessel detection with speed and heading derivation across sequential acquisitions | GIS layer of reconstructed vessel tracks with timestamps, speed estimates, and heading vectors |
| Lawnmower-pattern anomaly flag | Geometric analysis of track parallelism and spacing against baseline transit-behaviour profiles | Scored anomaly report per vessel event, with confidence tier (high / medium / low) |
| Bathymetric overlay and depth-band assessment | Cross-referencing vessel track with public hydrographic depth contours to assess mine-laying depth compatibility | Annotated chart showing track segments within operationally relevant depth bands |
| Vessel-type confirmation | Optical imagery analysis (WorldView-3 or Planet) matched against known hull-form signatures of minelayer classes | Vessel identification report with imagery extract and confidence assessment |
| Probability-of-mining area map | Convergent-indicator scoring combining track geometry, speed, depth band, and vessel type | Gridded probability layer (GeoTIFF or shapefile) for handoff to in-water survey planning |
| Retrospective archive analysis | Sentinel-1 archive mining to establish pre-crisis baseline and identify anomaly onset date | Timeline report with baseline statistics and annotated anomaly events |
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.