Pipeline right-of-way construction surveillance
Satellite optical and SAR imagery tracks every phase of pipeline construction, from initial clearing through reinstatement, flagging encroachment and reinstatement failures that ground inspectors routinely miss.
Sensors
- Planet SkySat: 50 cm panchromatic, approximately 72 cm multispectral; tasked daily over active sections. Resolves individual machinery, open trench width and spoil-heap position. Small swath (about 15 km) suits targeted corridor tasking rather than basin-wide sweeps.
- Airbus Pléiades Neo: 30 cm panchromatic native resolution; four-band plus red-edge multispectral at 1.2 m. Stereo pairs on a single pass enable spoil-volume estimates. Revisit typically 1 to 2 days over mid-latitudes when tasked.
- Sentinel-2: 10 m multispectral (visible and near-infrared), 20 m red-edge and SWIR; 5-day global revisit at the equator, improving to 2 to 3 days at higher latitudes. Provides corridor-wide NDVI baselines and bare-soil exposure indices at no acquisition cost. Cloud is the principal operational constraint.
- Sentinel-1: C-band SAR at 5.6 cm wavelength; Interferometric Wide Swath mode delivers 10 m resolution with 6- or 12-day repeat at mid-latitudes. Coherence differencing between two passes identifies active earthworks regardless of cloud or night. Backscatter change flags soil disturbance and standing water in open trenches.
Why a pipeline corridor is unusually legible from orbit
A major pipeline right-of-way (RoW) is typically 20 to 50 metres wide and can run for hundreds of kilometres. That geometry, a narrow linear feature crossing many land-cover types, is almost purpose-built for satellite detection. Vegetation is cleared in a single pass of machinery, producing an abrupt spectral boundary between the bare construction strip and adjacent land that registers clearly in 10 m Sentinel-2 imagery and with high precision in 50 cm SkySat frames.
The construction sequence itself is temporally structured. Clearing precedes topsoil strip, which precedes trenching, which precedes pipe-laying and backfill, which precedes reinstatement. Each phase has a distinct spectral and structural signature. Bare mineral soil has a very different SWIR reflectance from topsoil or growing vegetation. An open trench, depending on groundwater, may show standing water. Compacted backfill has higher SAR backscatter than loose spoil. This phase-to-phase contrast is the analytical foundation for satellite-based construction surveillance.
SAR coherence as the all-weather workhorse
Optical imagery fails whenever cloud cover persists, which is precisely when large earthmoving projects often accelerate to meet weather windows. Sentinel-1 C-band SAR fills that gap. Coherence, the statistical similarity of the radar phase between two passes, drops sharply wherever the ground surface has changed between acquisitions. Active earthworks, machinery movement, spoil redistribution and trench excavation all destroy coherence. A coherence map derived from two consecutive 6-day Sentinel-1 passes will show the active construction front as a low-coherence strip against a stable, high-coherence background.
The practical detection limit depends on the degree of surface change. A single pass of a tracked excavator over bare soil is sufficient to decorrelate C-band returns. Reinstatement quality is also detectable: ground that has been graded, seeded and left to settle will gradually recover coherence as vegetation re-establishes, providing an objective measure of ecological reinstatement progress over months. The method cannot, however, distinguish the cause of decorrelation without optical corroboration. Heavy rainfall on bare soil produces similar decorrelation to active machinery, so cross-checking with Sentinel-2 or commercial optical imagery is standard practice.
NDVI and bare-soil indices for reinstatement quality
Reinstatement, the restoration of topsoil, drainage and vegetation after backfill, is a contractual obligation on virtually every pipeline project and a frequent source of dispute. Satellite-derived NDVI (Normalised Difference Vegetation Index) from Sentinel-2 bands B8 and B4 provides a repeatable, auditable record of vegetation recovery along the corridor. A pre-construction NDVI baseline is computed from archive imagery; subsequent acquisitions track how quickly the RoW returns to background NDVI levels.
The Bare Soil Index (BSI), computed from SWIR, red and blue bands, complements NDVI by directly measuring exposed mineral soil fraction. Sections where BSI remains elevated three or six months after nominal reinstatement indicate either failed seeding, erosion of replaced topsoil, or compaction preventing germination. At 10 m resolution, Sentinel-2 can flag sections of a few tens of metres where reinstatement is lagging, though it cannot diagnose the cause. Pléiades Neo or SkySat imagery at 30 to 50 cm is then used to characterise the specific failure mode before a ground inspection is dispatched.
Telling authorised construction from third-party encroachment
This is the operationally difficult problem. A pipeline operator's construction contractor and an unauthorised third party both produce bare soil, machinery signatures and vegetation loss when viewed from orbit. The distinction requires context: the authorised construction front moves predictably along the corridor in a known direction and at a known pace, documented in the project schedule. Any disturbance that appears outside the active construction zone, or that runs perpendicular to the RoW, or that appears on a completed and reinstated section, is anomalous.
Change detection algorithms applied to daily SkySat tasking can flag new disturbance patches within 24 hours of acquisition. The alert includes the location, approximate area, and a comparison chip showing the before and after state. The operator's field team then makes the encroachment determination. Satellite data does not replace that judgement; it reduces the area that needs to be physically inspected from hundreds of kilometres to a handful of flagged locations. On long corridors crossing remote or politically sensitive terrain, that triage function is the primary value.
Archive depth and the pre-construction baseline problem
One underappreciated advantage of satellite surveillance is retrospective analysis. Sentinel-2 archive runs continuously from 2015, Sentinel-1 from 2014, and Landsat from 1972. Before a pipeline project breaks ground, a multi-year baseline of land cover, seasonal NDVI cycles, flood extent and existing disturbance can be established from open archive data. This baseline matters for two reasons: it documents the pre-project condition for environmental liability purposes, and it identifies pre-existing disturbances that would otherwise be misattributed to construction activity.
Commercial archive from Planet and Airbus extends daily or near-daily coverage back several years over many regions. For projects in areas with persistent seasonal cloud, SAR archive from Sentinel-1 provides the only cloud-free historical record. Satellize incorporates this kind of pre-project baseline work into its analytics pipeline; the Tonga crop-estimation programme used a similar multi-season archive approach to separate genuine change from seasonal noise.
Honest limits of the method
Resolution sets a floor on what is detectable. At 10 m, Sentinel-2 cannot resolve an individual trench or a single vehicle. It measures aggregate land-cover change over pixels, which means a narrow disturbance of less than about 20 m width will be diluted and may not register as a clear anomaly. Commercial imagery at 30 to 50 cm resolves individual machines and open trench edges, but daily tasking over a 500 km corridor is expensive and the swath width of SkySat (roughly 15 km) means multiple passes are needed to cover wide corridors.
SAR coherence requires two passes separated by at least one repeat cycle, so the minimum latency for a coherence-based alert is 6 days on Sentinel-1, or 12 days if the 6-day track is not available for a given geometry. That is adequate for slow-moving encroachment but insufficient for rapid, targeted sabotage. Atmospheric water vapour and soil moisture changes introduce noise into both SAR and optical indices, particularly in tropical climates. No satellite method replaces physical inspection; it redirects it.
Typical figures
| Best optical resolution (tasked) | 30 cm (Pléiades Neo panchromatic); 50 cm (SkySat panchromatic) |
| Corridor-wide optical resolution | 10 m multispectral (Sentinel-2 visible/NIR) |
| SAR resolution | 10 m (Sentinel-1 IW mode, range × azimuth) |
| Optical revisit (commercial, tasked) | Daily to 2 days (SkySat, Pléiades Neo) over active sections |
| SAR coherence repeat cycle | 6 days (Sentinel-1 at mid-latitudes); 12 days where single-track geometry applies |
| Minimum detectable disturbance (optical) | Approximately 2 × 2 m at 50 cm resolution; approximately 20 m width at 10 m resolution |
| Key spectral bands | Red, NIR, SWIR (Sentinel-2 B4, B8, B11, B12) for NDVI and BSI; C-band 5.6 cm for SAR |
| Archive depth | Sentinel-1 from 2014; Sentinel-2 from 2015; Landsat from 1972 |
| Alert latency (change detection) | 12 to 24 hours after image acquisition for optical; 6 to 12 days for SAR coherence |
| Delivery formats | GeoTIFF change layers, GeoJSON alert polygons, corridor KML, PDF inspection reports |
Analytics Satellize can run
| Construction phase progression map | Supervised land-cover classification of multispectral imagery; change detection between sequential acquisitions | Weekly GeoTIFF layer showing cleared, trenched, backfilled and reinstated segments along the corridor |
| Active earthworks alert | SAR coherence differencing (Sentinel-1 repeat-pass pairs); backscatter change thresholding | Automated GeoJSON alert polygon with location, area and acquisition date, delivered within 24 hours of SAR pass processing |
| Reinstatement quality index | Time-series NDVI and Bare Soil Index computed from Sentinel-2; comparison against pre-construction baseline and seasonal norms | Monthly corridor report with colour-coded reinstatement compliance score per 100 m segment; flagged non-compliant sections |
| Third-party encroachment flag | Anomaly detection on daily optical change chips; spatial filtering against authorised construction zone boundary | Priority-ranked alert list with before/after image chips and coordinates, formatted for field-inspection dispatch |
| Pre-construction baseline dossier | Multi-year archive analysis of Sentinel-1, Sentinel-2 and Landsat; seasonal decomposition of NDVI time series | PDF and GIS package documenting pre-project land cover, existing disturbances and seasonal vegetation cycles |
| Spoil-heap volume estimate | Stereo photogrammetry from Pléiades Neo same-pass stereo pairs; DSM differencing against pre-construction elevation | Georeferenced volume estimate (m³) per spoil location with uncertainty bounds; updated on each tasked stereo acquisition |
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.