Coastal obstacle and beach-defence emplacement monitoring
Anti-landing obstacles, wire barriers and beach-defence structures leave detectable signatures in very-high-resolution optical imagery at low tide and in X-band SAR. This page explains the physics, the sensors, and the honest limits of what space-based monitoring can and cannot see.
Sensors
- Maxar WorldView-3: 0.31 m panchromatic, 1.24 m multispectral. The resolution is sufficient to distinguish individual obstacle types (hedgehogs, tetrahedra, wire pickets) and count emplacements. Revisit at a given coastal point varies from same-day to roughly three days depending on latitude and tasking priority.
- Airbus Pléiades Neo: 0.30 m panchromatic, 1.2 m multispectral. Four-satellite constellation gives a revisit of around one day at mid-latitudes. Stereo pairs can produce 1 m digital surface models, useful for estimating obstacle height above the beach surface.
- ICEYE X-band SAR (spotlight mode): Approximately 0.25 m azimuth resolution in spotlight mode. Metal and reinforced-concrete obstacles produce radar cross-sections orders of magnitude above bare sand, making them detectable regardless of cloud cover or time of day. Coherent change detection between two passes can flag new installations within the scene.
- Planet SkySat: 0.50 m panchromatic. Lower resolution than WorldView-3 or Pléiades Neo, but the constellation of 21 satellites supports near-daily tasking of the same coastal strip, making it well-suited to change-detection workflows that prioritise revisit over fine detail.
What the intertidal zone reveals at low tide
Beach-defence obstacles are designed to impede amphibious landings, not to hide from satellites. Anti-tank hedgehogs, concrete tetrahedra, wooden or steel stakes, and wire entanglements are typically emplaced in the intertidal zone, where they are exposed for hours each tidal cycle. A WorldView-3 or Pléiades Neo image acquired within two hours of local low water can resolve individual structures down to sub-metre scale, allowing analysts to classify obstacle type, estimate spacing, and map the extent of a defended beach frontage.
Shadows matter. Obstacles of 0.5 to 1.5 m height cast measurable shadows even at high solar elevation, and the shadow geometry provides an independent height estimate when stereo imagery is unavailable. Multispectral bands add a further discriminant: concrete and galvanised steel have spectral signatures distinct from wet sand and seaweed, reducing false-positive rates in automated detection pipelines.
Why SAR sees what cloud and darkness hide
Optical sensors are blind at night and degraded by cloud. The Taiwan Strait and the Black Sea coast are both subject to seasonal overcast that can persist for days. X-band SAR, operating at roughly 9.6 GHz, penetrates cloud and works at any hour. More importantly, the radar cross-section of a metal hedgehog or a concrete block is dramatically higher than that of a flat sandy beach. In SAR imagery, a cluster of newly emplaced obstacles appears as a bright linear or gridded return against a low-backscatter background.
Coherent change detection (CCD) is the most sensitive technique available. By differencing two complex SAR acquisitions of the same scene, CCD highlights any change in the scattering geometry at centimetre scale. A new row of obstacles installed between two ICEYE passes separated by 24 hours will appear as a high-coherence-loss stripe precisely where the installation occurred. The method is well-documented in the published SAR literature and has been applied in open-source analyses of both the Taiwan Strait and Black Sea coastal zones.
One caveat: tidal state affects SAR returns significantly. An obstacle submerged by even 0.3 m of water will have its radar return attenuated by the water column. Acquisition scheduling must account for local tidal predictions to ensure the intertidal zone is exposed during the satellite pass.
Where the physics stops working
Submerged obstacles below mean low water are not detectable by passive optical sensors. Multispectral water-penetration is limited to roughly 10 to 20 m in clear tropical water using coastal and blue bands, but turbid coastal water, common in the Yellow Sea and parts of the Black Sea, reduces that to 2 to 5 m or less. Obstacles on the seabed in turbid conditions are effectively invisible to optical methods.
SAR has a different problem. Beneath the water surface, the X-band signal is absorbed within centimetres. Submerged obstacles generate no usable return. Detecting underwater beach obstacles requires acoustic methods or airborne lidar, neither of which is available from current commercial satellite constellations.
Resolution also sets a minimum detectable target. A lone wire picket of 5 cm diameter will not appear as a discrete object in any current commercial optical sensor. What is detectable is the pattern: a row of pickets at 2 m spacing creates a linear texture anomaly that automated detectors can flag, even when individual elements are below the resolution floor.
Change detection as the operational product
A single image is a snapshot. The operational value comes from comparing scenes over time. A baseline image of a beach acquired before any construction activity establishes the reference state. Subsequent tasking, ideally at the same tidal phase to control for intertidal exposure, is differenced against the baseline. New high-backscatter objects in SAR, or new shadow-casting structures in optical, trigger an alert.
The practical cadence depends on the threat assessment. For a monitored strait or landing corridor, weekly SAR passes combined with optical tasking on cloud-free days provides a reasonable detection latency of two to seven days for a new installation of meaningful scale. A rapid emplacement of a kilometre-long obstacle belt over 48 hours would likely be captured within one SAR revisit cycle given current ICEYE constellation density.
Open-source analysts have demonstrated this workflow using Sentinel-1 C-band SAR (10 m resolution in interferometric wide-swath mode) for large-scale pattern detection, then cuing commercial X-band spotlight collections for confirmation at higher resolution. The two-tier approach balances cost against detection probability.
Applying this to real threat corridors
The Taiwan Strait and the Black Sea coast of Ukraine are the two settings where open-source coastal obstacle analysis has been published and discussed in detail. Both share relevant characteristics: long, gently shelving beaches with predictable tidal cycles, active military engineering activity, and significant commercial satellite archive depth going back several years. That archive matters. Establishing when a particular beach was clear, and when structures first appeared, is often as analytically important as knowing the current state.
For defence planners, the key output is not a single image but a time-series product: a georeferenced record of obstacle extent, density and type, updated on a defined schedule, with change flags and confidence assessments. Satellize builds and maintains exactly this kind of persistent monitoring layer for clients, drawing on both open constellations and commercial tasking. The methodology is the same one applied in our Tonga crop-estimation programme, where the operational discipline is regular, tidal-phase-aware collection and automated change flagging, not one-off tasking.
The honest limit of any space-based coastal monitoring programme is that it tells you what is on the surface. Engineering judgement, bathymetric data and human intelligence remain necessary to assess what lies beneath.
Typical figures
| Best optical resolution (panchromatic) | 0.30 m (Pléiades Neo), 0.31 m (WorldView-3) |
| Best SAR resolution (spotlight mode) | ~0.25 m azimuth (ICEYE); ~1 m (Capella Space spotlight, comparable systems) |
| Typical revisit at a given coastal point | 1 day (SkySat constellation); 1–3 days (Pléiades Neo); same-day to 3 days (WorldView-3 tasked); sub-daily possible with ICEYE multi-satellite tasking |
| Detection latency (new installation to alert) | 2–7 days under routine monitoring; potentially under 24 h with priority tasking |
| Spectral bands (optical) | Panchromatic, coastal blue, blue, green, red, red-edge, NIR, SWIR (WorldView-3); panchromatic plus 6 multispectral bands (Pléiades Neo) |
| SAR frequency | X-band (~9.6 GHz); C-band (5.4 GHz, Sentinel-1) for wide-area cueing |
| Minimum detectable target (practical) | Individual obstacle structures ~0.5 m and above in optical; linear arrays of smaller elements detectable as texture patterns in SAR |
| Optical water penetration (clear water) | 10–20 m in coastal/blue bands; 2–5 m in turbid coastal water |
| Archive depth | Sentinel-1 from 2014; WorldView-3 from 2014; Pléiades from 2012; ICEYE from 2018 |
| Delivery formats | Georeferenced GeoTIFF, GeoJSON change layers, PDF intelligence summary, GIS-ready vector overlays |
Analytics Satellize can run
| Obstacle-belt extent map | Object-based image analysis (OBIA) on VHR optical imagery at low tide; shadow-length height estimation | GeoJSON polygon layer with obstacle type classification, density estimate and confidence score, updated per collection |
| New-installation change alert | Coherent change detection (CCD) on repeat X-band SAR spotlight pairs; differenced backscatter thresholding against baseline | Automated alert with georeferenced change mask and SAR chip, delivered within hours of satellite downlink |
| Time-series obstacle growth record | Multi-date optical and SAR stack, tidal-phase-normalised, with per-epoch extent polygons | Chronological GIS layer set with annotated timeline report; suitable for legal or intelligence archive |
| Obstacle type classification | Template matching and spectral signature analysis on sub-metre optical imagery; hedgehog, tetrahedron and wire-picket classes | Classified point layer with structure count, type label and image chip per cluster |
| Tidal-phase collection schedule | Tidal prediction model (published harmonic constituents) cross-referenced against satellite overpass windows to maximise intertidal exposure | Tasking calendar with predicted low-water windows and recommended sensor per pass |
| Wide-area cueing layer | Sentinel-1 C-band wide-swath backscatter anomaly detection across a full strait or coastal sector; flags areas for VHR follow-up | Weekly anomaly map (GeoTIFF + PDF) covering up to 250 km of coastline per Sentinel-1 swath |
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.