Ship-Breaking and Beaching Activity Monitoring at Scrapping Yards
Time-series optical and SAR imagery tracks vessel arrivals, beaching events, and progressive dismantlement at South Asian ship-recycling yards, where hull length and radar cross-section change measurably as cutting proceeds over weeks to months.
Sensors
- Maxar WorldView Legion: 30 cm panchromatic resolution at nadir; revisit up to 15 times per day over targeted sites. Resolves individual cutting crews, hull section removal, and scrap pile growth. Cloud-limited; useless during heavy monsoon.
- Airbus Pléiades Neo: 30 cm native resolution, 50 cm multispectral, constellation of four satellites giving same-day stereo and roughly daily revisit over a fixed site. Colour bands help distinguish active rust-cutting from painted idle hulls and from onshore scrap accumulation.
- Planet SkySat: 50 cm resolution, tasked revisit up to twice daily. Lower cost per collect than WorldView Legion; useful for daily change detection between the higher-resolution anchor images.
- Sentinel-1 SAR (C-band, 5.405 GHz): Interferometric Wide Swath mode at 10 m resolution; 6-day repeat at the equator, 1–3 days at higher latitudes. Cloud-penetrating, so it maintains temporal continuity through the June–September monsoon when optical sensors go dark. Detects large metal structures reliably; cannot resolve cutting detail at 10 m.
What a shrinking hull tells you
A Capesize bulk carrier is roughly 290 m long. By the time it leaves a breaking yard it is gone entirely. That disappearance is not instantaneous: cutting gangs work bow-to-stern or stern-to-bow, removing sections over eight to sixteen weeks depending on the yard's throughput. Each week the measurable length of the hull in a satellite image decreases by a predictable amount, typically five to twenty metres depending on vessel class and cutting pace. That rate of change is the primary analytic signal.
Radar cross-section tells a complementary story. A fully intact ship returns a strong, coherent SAR signal from its superstructure, hull plating, and machinery spaces. As those metal structures are removed, the RCS drops and its spatial extent contracts. Sentinel-1, imaging at C-band, cannot see a cutting crew, but it can detect a hull that was 200 m last week and is now 170 m. That 10 m pixel is coarse, but the change in the bright return's length is measurable across time steps.
The beaching event itself: what arrives and when
Most South Asian breaking yards, concentrated along the Alang coast in Gujarat, the Gadani beach in Pakistan, and Chittagong in Bangladesh, operate on a beaching method. A vessel is run aground at high tide under its own power or with tug assistance. The moment of beaching is often visible in optical imagery as a ship that was previously at anchor in the approach zone is now bow-on to the beach. AIS typically goes silent before or during the final approach, which is why satellite imagery rather than transponder data is the primary evidence source here.
Very-high-resolution optical imagery, particularly from WorldView Legion or Pléiades Neo, can distinguish an active cutting operation from a vessel in idle storage by the presence of cutting scaffolding, removed hull plates stacked on the beach, and the characteristic gap in the hull where sections have been lifted away. An idle vessel looks intact. An actively cut vessel does not. That distinction matters for compliance monitoring: some vessels are beached but held pending sale of the ship's parts, and the dismantlement clock has not yet started.
Monsoon continuity: why SAR is not optional
The Indian subcontinent monsoon runs from roughly June to September. During that period, persistent cloud cover can render optical sensors effectively blind over coastal Gujarat and Chittagong for days or weeks at a time. A vessel can arrive, beach, and have its superstructure removed entirely within a single cloudy fortnight.
Sentinel-1 C-band SAR penetrates cloud and rain without signal degradation. Its 6-day repeat orbit over South Asian latitudes, reducible to 1–3 days when ascending and descending passes are combined, provides the temporal backbone during optical blackout periods. The trade-off is resolution: at 10 m, SAR cannot confirm cutting crew presence or distinguish scrap pile growth from other beach material. The practical workflow pairs SAR-based hull-length measurement through the monsoon with high-resolution optical confirmation at the start and end of each cloud-free window.
Limits the data honestly cannot overcome
Sub-metre optical imagery is expensive to task repeatedly over multiple yards simultaneously. WorldView Legion and Pléiades Neo tasking costs scale with area and frequency; monitoring all active berths at Alang, which can hold sixty or more vessels at peak throughput, requires prioritisation. Yards that operate at night partially evade optical detection, though SAR is unaffected by darkness.
Hull-length measurement from SAR carries an uncertainty of roughly one to two pixels, meaning five to twenty metres at Sentinel-1 resolution. For a vessel being cut at ten metres per week, that uncertainty can obscure a week's worth of progress. Panchromatic optical measurement from WorldView or Pléiades is far more precise, resolving length changes to within a metre or two, but only when cloud-free imagery is available. Scrap pile volume estimation from optical imagery is possible in principle using shadow-length photogrammetry or stereo elevation models, but the irregular shape and mixed composition of scrap heaps make tonnage estimates unreliable without ground truth.
From pixels to compliance and market intelligence
The primary buyers of this analysis fall into two groups. The first is environmental and labour compliance: international conventions including the Hong Kong International Convention for the Safe and Environmentally Sound Recycling of Ships establish standards that are difficult to audit remotely but whose basic precondition, knowing which vessel went to which yard and when, is exactly what satellite time-series provides. Flag states, port state control authorities, and NGOs use imagery evidence to correlate a vessel's last AIS position with its appearance at a breaking beach.
The second group is commodity and financial intelligence. Steel scrap from ship-breaking is a significant feedstock for South Asian electric arc furnace mills. The rate of dismantlement across Alang, Gadani, and Chittagong, aggregated from satellite observations, is a leading indicator of domestic scrap supply. Vessel arrivals at breaking yards also signal the end of a ship's commercial life, which has implications for fleet capacity models in bulk shipping and tanker markets. Satellize runs time-series change-detection analytics on open and commercial constellations; the same change-detection pipeline developed for the Tonga crop-estimation programme applies directly to sequential hull-length measurement.
Building a monitoring cadence that matches the cutting pace
A practical monitoring programme for a single major yard combines a weekly Sentinel-1 SAR collect for hull-length tracking, a twice-weekly Planet SkySat collect for intermediate optical change detection, and a fortnightly WorldView Legion or Pléiades Neo collect for high-confidence measurement and visual confirmation. During monsoon, the SAR cadence holds; optical tasking is suspended until a clear-sky window is forecast.
Archive depth matters. Sentinel-1 data is freely available back to 2014 through the Copernicus Data Space, and Planet's archive extends to 2016 for many sites. That allows retrospective reconstruction of dismantlement histories for vessels whose documentation is disputed. For a sanctions-related investigation, knowing that a vessel arrived at Gadani on a specific date and was fifty per cent dismantled within six weeks is evidence that no AIS record can provide after the transponder went dark.
Typical figures
| Optical resolution (VHR tasked) | 30 cm panchromatic (WorldView Legion, Pléiades Neo); 50 cm (SkySat) |
| SAR resolution (Sentinel-1 IW mode) | 10 m ground range; 5 m range × 20 m azimuth in Stripmap |
| Revisit (optical, tasked) | Up to daily (SkySat); up to 15×/day over a point (WorldView Legion); ~1/day (Pléiades Neo constellation) |
| Revisit (Sentinel-1 SAR) | 6-day repeat per orbit; 1–3 days combining ascending and descending passes at South Asian latitudes |
| Minimum detectable hull-length change (SAR) | ~10–20 m (one to two Sentinel-1 pixels; uncertain at weekly intervals) |
| Minimum detectable hull-length change (optical VHR) | ~1–2 m under cloud-free conditions |
| Cloud penetration | SAR: full; optical: none. Monsoon blackout risk June–September over South Asian yards |
| Archive depth | Sentinel-1: 2014–present (Copernicus Data Space); Planet SkySat: 2016–present; WorldView: site-dependent tasking history |
| Coverage | Global; primary monitored clusters: Alang (India), Gadani (Pakistan), Chittagong (Bangladesh) |
| Delivery formats | GeoTIFF change layers, GeoJSON vessel footprint polygons, timestamped length-measurement CSV, PDF evidence report |
Analytics Satellize can run
| Vessel arrival and beaching event log | Sequential object detection on VHR optical and SAR time-series; comparison against prior-period anchor-zone imagery | Timestamped GeoJSON record of each beaching event with vessel position, estimated length, and first-detect date |
| Hull-length time series per vessel | Semi-automated bright-target extent measurement on SAR amplitude imagery; manual verification on VHR optical where cloud-free | CSV time series of measured hull length per collect, with uncertainty bounds; plotted dismantlement curve |
| Active-cutting vs. idle-storage classification | Visual and spectral interpretation of VHR optical: presence of hull gaps, removed plates, scaffolding, and scrap accumulation adjacent to hull | Per-vessel status label (beached-idle / active-cutting / dismantlement-complete) updated at each optical collect |
| Yard-level throughput and capacity utilisation | Aggregated vessel count and cutting-rate statistics across all monitored berths; normalised by yard length from baseline imagery | Monthly yard throughput report with berth occupancy rate and estimated scrap tonnage range |
| Retrospective dismantlement reconstruction | Archive SAR and optical time-series analysis; cross-referenced against IMO vessel identity and last-known AIS position | PDF evidence dossier with annotated imagery sequence, hull-length chart, and estimated arrival and completion dates |
| Monsoon-period continuity alert | Sentinel-1 SAR change detection; threshold trigger on RCS extent reduction exceeding one pixel between consecutive passes | Automated alert with SAR chip and hull-length delta when significant change detected during optical blackout |
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.