Port Infrastructure Damage Assessment from Optical and SAR
Pre- and post-event SAR coherence analysis and optical image differencing reveal structural damage to quays, cranes, warehouses and breakwaters regardless of cloud cover, answering the operational question: can this port physically function?
Sensors
- Sentinel-1 SAR (C-band, ESA): 5 x 20 m ground resolution in Interferometric Wide Swath mode; 6-day exact repeat at mid-latitudes (12-day for a single satellite). Coherence differencing between pre- and post-event image pairs flags surfaces that have changed structurally, cloud-independent. Free and open archive from 2014.
- Maxar WorldView-3 optical: 0.31 m panchromatic, 1.24 m multispectral (8-band VNIR plus SWIR). Tasked on demand; revisit of 1 to 4.5 days depending on off-nadir angle accepted. Provides damage classification, extent polygons and visual confirmation that SAR coherence anomalies cannot alone supply.
- Airbus Pleiades Neo optical: 0.30 m panchromatic, 1.2 m multispectral. Constellation of four satellites gives same-day revisit over most locations. Stereo tasking allows 3D point-cloud generation of structural profiles, useful for estimating crane or warehouse collapse geometry.
- ICEYE SAR (X-band): Spotlight mode delivers approximately 0.5 m resolution; strip mode approximately 3 m. Revisit of under 24 hours over any target is achievable across the constellation. X-band is more sensitive to small metallic structures (crane jibs, container stacks) than Sentinel-1's C-band, at higher tasking cost.
What a coherence map actually measures
SAR coherence compares the phase relationship between two radar images acquired over the same area at different times. Where a surface is physically unchanged, the phase relationship stays stable and coherence is high. Where structure has shifted, collapsed or been buried under debris, coherence drops sharply. The technique does not require any pre-existing map of the port; the pre-event image is the baseline.
Sentinel-1's 6-day repeat cycle is the practical constraint. If a storm or strike occurs on day one and cloud persists for five days, the first post-event SAR acquisition still arrives on schedule. A coherence loss map can be generated within hours of that acquisition. The spatial resolution of 5 x 20 m in Interferometric Wide Swath mode means individual berths and warehouses are distinguishable, though a single crane arm may not be. That is where ICEYE's sub-metre Spotlight mode earns its place.
Optical imagery: classification, not just confirmation
Coherence loss tells you that something changed. Optical imagery at 0.3 to 0.5 m resolution tells you what changed and how badly. WorldView-3 and Pleiades Neo imagery, compared against archive imagery of the same port, support a four-class damage schema commonly used in post-disaster rapid mapping: no damage, minor damage, major damage and destroyed. That classification drives the operational decision: which berths can accept vessels immediately, which require engineering inspection before use, and which are out of service for weeks or months.
SWIR bands on WorldView-3 add a further diagnostic. Wet concrete, fire damage and exposed structural steel each have distinct SWIR reflectance signatures that are invisible in standard RGB imagery. A warehouse roof that looks superficially intact in natural colour may show clear stress or fire penetration in the SWIR, which matters when assessing whether stored cargo is compromised.
Cloud is the honest limit of optical methods. In tropical cyclone conditions, cloud cover can persist for 48 to 72 hours after storm passage. The correct workflow pairs SAR coherence as the first-pass damage indicator with optical tasking as soon as a cloud-free window opens. The two products are complementary, not interchangeable.
Breakwaters and quay walls: the structures that matter most
Port operability depends on structures that are rarely the most photogenic targets. A collapsed warehouse is visible and classifiable in any sub-metre image. A quay wall that has settled 40 cm, or a breakwater section that has lost its armour stone without visibly breaching, is far harder to detect remotely. SAR coherence is sensitive to both scenarios because even subsidence of a few centimetres alters the phase return from a rough masonry surface.
Pleiades Neo stereo pairs offer a partial solution for quay walls. By generating a post-event digital surface model and differencing it against a pre-event model (or a design-elevation dataset), vertical displacement of quay decks can be estimated to within roughly 0.5 to 1 m vertically, depending on stereo geometry and ground control. That is sufficient to flag a berth as structurally suspect, though it does not replace a bathymetric survey for confirming navigable depth alongside.
Conflict damage: the same physics, a different context
Storm damage and conflict damage are analytically similar. Both produce coherence loss, both produce classifiable optical damage. The difference is that conflict damage is often localised to specific infrastructure nodes (crane machinery rooms, fuel storage, lock gates) rather than distributed across a port in the pattern of a weather event. That spatial signature can itself be informative.
There is an honest limit here. SAR coherence and optical classification identify that a structure has been damaged; they do not reliably distinguish whether the cause was a direct strike, a near-miss pressure wave or a secondary fire. For insurance, legal or attribution purposes, satellite imagery provides strong evidence of damage extent and timing but is not a substitute for on-site engineering survey. The archive depth of Sentinel-1 (operational since 2014) and commercial optical constellations means a pre-event baseline almost always exists, which is the foundation of any credible before-and-after assessment.
Turning image products into an operational answer
The output that matters to a port authority, insurer or government is not a GeoTIFF of coherence values. It is a berth-by-berth operability assessment: which structures are serviceable, which are degraded, and which are out of action. That requires the image analytics to be referenced against a port infrastructure layer (berth numbers, crane positions, warehouse footprints, breakwater sections) and expressed in terms a harbour master can act on.
Satellize structures damage assessments as GIS layers with associated tabular reports, keyed to port infrastructure identifiers where a client provides them. The analytic workflow draws on published change-detection methods, including coherence differencing per the ESA Sentinel-1 technical documentation and damage classification schemas consistent with Copernicus Emergency Management Service rapid-mapping outputs. The Copernicus EMS has published dozens of port and coastal infrastructure activations that validate the method class against ground-truth surveys.
Latency is the variable that clients most often underestimate. SAR acquisition is cloud-independent but still subject to satellite overpass timing. For a port at 40 degrees latitude, a Sentinel-1 acquisition will occur within 6 days of any event; with ICEYE tasking added, that window shrinks to under 24 hours. Optical cloud-free acquisition is unschedulable. A realistic planning assumption for a complete dual-sensor damage assessment is 24 to 96 hours after event, depending on cloud and tasking priority.
Typical figures
| SAR spatial resolution (Sentinel-1 IW) | 5 x 20 m (range x azimuth) |
| SAR spatial resolution (ICEYE Spotlight) | ~0.5 m |
| Optical resolution (WorldView-3 / Pleiades Neo) | 0.31 m / 0.30 m panchromatic |
| Sentinel-1 revisit (mid-latitudes, two satellites) | 6 days exact repeat; 1 to 3 days with cross-track overlap |
| ICEYE revisit | Under 24 hours to any target (constellation-dependent) |
| Optical revisit (Pleiades Neo constellation) | Same-day, subject to cloud and tasking priority |
| Minimum structurally detectable change (SAR coherence) | Surface displacement of order centimetres; collapsed structures reliably detected |
| Spectral coverage (WorldView-3) | 8-band VNIR (400–1040 nm) plus 8-band SWIR (1195–2365 nm) |
| SAR archive depth (Sentinel-1) | From April 2014 (Sentinel-1A launch) |
| Typical assessment latency | 24 to 96 hours post-event (SAR first pass; optical when cloud-free) |
Analytics Satellize can run
| SAR coherence loss map | Interferometric coherence differencing between pre- and post-event Sentinel-1 or ICEYE image pairs | GeoTIFF coherence difference layer with thresholded damage-probability overlay, cloud-independent |
| Optical damage classification | Object-based image analysis and visual interpretation of sub-metre pre/post image pairs, four-class schema (none / minor / major / destroyed) | Polygon GIS layer with damage class attributes, keyed to port infrastructure identifiers |
| Berth operability assessment | Fusion of SAR coherence and optical classification, referenced to client-supplied or open-source port infrastructure layer | Tabular report listing each berth or structure with status, confidence level and supporting image evidence |
| Quay-deck vertical displacement estimate | Pleiades Neo stereo-pair digital surface model differencing against pre-event baseline | Raster displacement map with flagged berths exceeding defined settlement threshold |
| SWIR anomaly detection for fire and material damage | WorldView-3 SWIR band analysis for exposed steel, fire char and wet concrete signatures invisible in VNIR | Classified raster layer highlighting SWIR anomalies overlaid on natural-colour imagery |
| Damage timeline reconstruction | Multi-date coherence and optical archive analysis to establish damage onset window and progression | Annotated image time series with event chronology, suitable for insurance or legal reporting |
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.