Sensors
- Maxar WorldView-2: 0.46 m panchromatic, 1.84 m multispectral. Resolves individual pylon bases, transformer pad outlines and the gravel aprons of new substations. On-demand tasking with typical revisit of 1 to 3.5 days depending on latitude and collection angle.
- Planet SkySat: 0.5 m panchromatic collect. Useful for rapid change confirmation once a candidate corridor is flagged; less archive depth than WorldView but capable of near-daily tasking on request.
- Sentinel-2 MSI: 10 m multispectral, 5-day revisit at mid-latitudes with two satellites. Detects the spectral signature of freshly cleared vegetation strips and bare-soil corridors at medium scale; insufficient to distinguish pylon spacing or conductor gauge.
- Sentinel-1 SAR C-band: 10 m ground range detected, 6-day repeat (12-day single satellite). Backscatter change along a linear corridor persists through cloud cover and partial canopy, making it the primary tool for detecting cleared rights-of-way in humid or seasonally overcast regions.
- Planet PlanetScope: 3 m multispectral, near-daily global revisit. Monitors corridor development over time at a cadence that WorldView cannot match economically; used for temporal stacking rather than fine structural identification.
Why grid extension leaks development intent
Electricity infrastructure is not speculative. A utility does not clear a right-of-way, pour a transformer pad and string medium-voltage conductors on a hunch. The capital commitment, typically running into millions before a single building is connected, follows a signed agreement with a developer or a government infrastructure directive. That agreement often precedes formal planning approval by six to eighteen months, particularly in jurisdictions where grid connection is a prerequisite for a planning application rather than a consequence of it.
The implication for property intelligence is direct. A new substation footprint appearing on the urban fringe is a harder signal than a planning application, which can be withdrawn. It is more specific than a rezoning notice, which may cover thousands of hectares with no committed timeline. Grid extension is money already spent, pointing at a location.
What a cleared corridor gives away
The first visible signature is vegetation removal. A right-of-way for a medium-voltage overhead line is typically 10 to 30 metres wide, depending on voltage class and national standards. At 10 m resolution, Sentinel-2 picks up the spectral change from vegetated land to bare soil within one or two revisit cycles. The linear geometry distinguishes it from agricultural clearance or road construction, though not always unambiguously at that resolution.
Sentinel-1 SAR adds the cloud-independent dimension. C-band backscatter increases sharply over newly exposed soil and decreases over felled timber left on the ground. The combination of a bright linear feature in one polarisation and a geometric shadow in another is a reliable corridor indicator in forested or frequently cloudy terrain. Published change-detection studies using Sentinel-1 coherence change have demonstrated detection of linear infrastructure clearance at widths down to roughly 20 metres.
Once the corridor is confirmed, sub-metre optical imagery from WorldView-2 or SkySat resolves the structural details: pylon type (lattice versus monopole), approximate span length, and whether a transformer pad has been poured. Pad dimensions and the presence of security fencing around a substation compound allow a rough inference of the voltage class and therefore the scale of development the infrastructure is designed to serve.
Honest limits of the method
Dense tree canopy is the primary obstruction. Overhead lines routed through mature forest or under a closed canopy are invisible to optical sensors and only partially recoverable from SAR, where the signal from the canopy itself dominates. Underground cable routes leave no surface signature at all, though the associated substation pads and switching stations remain visible.
Resolution creates a classification ceiling. Distinguishing a 33 kV distribution line from an 11 kV spur, or a 132 kV sub-transmission corridor from a 400 kV transmission line, requires not just sub-metre imagery but also knowledge of local utility standards. Pylon height, conductor bundling and tower geometry are all voltage-class indicators, but they are difficult to read reliably from nadir-looking imagery alone. Oblique collection from WorldView-2 helps with tower height estimation but introduces positional uncertainty.
Timing ambiguity is real. Grid extension can precede development by two years or more, or it can accompany industrial expansion that has nothing to do with housing. The signal is most useful when combined with other layers: road access construction, earthwork signatures, or changes to adjacent land parcels.
From pixel to development signal: the analytic chain
The practical workflow begins with a baseline optical or SAR composite for the area of interest. Automated linear feature extraction, using methods such as Hough transform variants or morphological filtering on edge-detected imagery, identifies candidate line routes. These are then validated against the temporal archive: a feature that appears between two dates is a new installation rather than a pre-existing one.
Substation and transformer pad detection uses object-based image analysis. The spectral and textural signature of a concrete or gravel pad, combined with the geometric regularity of the compound boundary and the presence of metallic structures, separates these features from other bare-ground objects. At 0.5 m resolution, WorldView-2 imagery supports reliable detection of pads larger than roughly 20 by 20 metres.
The output is a change layer: new line corridors as polylines with a confidence score, new substation footprints as polygons with an estimated commission date range. Overlaid on cadastral data, this identifies which land parcels sit within a plausible connection zone for the new infrastructure, and therefore which are most likely to see development pressure in the near term.
Where this fits in a property intelligence programme
Grid extension detection is most powerful as a screening tool. It narrows a large regional area of interest to a set of specific corridors and nodes where capital has already been deployed. A property fund assessing greenfield exposure across a 500 km² region cannot commission sub-metre tasking of the whole area economically. Sentinel-1 and Sentinel-2 change detection over the full extent, followed by targeted WorldView-2 or SkySat tasking of flagged locations, is the cost-rational approach.
Satellize runs this kind of tiered detection workflow on open and commercial constellations, applying the same analytic logic it developed for agricultural change detection in programmes such as the Kingdom of Tonga crop-estimation work, adapted to linear infrastructure rather than crop boundaries. The output integrates with standard GIS environments as a vector layer or a structured alert feed keyed to parcel identifiers.
For investors in markets where planning registers are slow, incomplete or not publicly accessible, this method provides a degree of transparency that documentary research alone cannot. Grid extension is a physical fact recorded in imagery, not a bureaucratic entry that may or may not reflect current intent.
Typical figures
| Finest spatial resolution (optical) | 0.46 m panchromatic (WorldView-2, SkySat) |
| SAR resolution for corridor detection | 10 m ground range (Sentinel-1 IW mode) |
| Revisit for change monitoring | 5 to 6 days (Sentinel-1/2 combined); 1 to 3.5 days on-demand (WorldView-2) |
| Minimum detectable corridor width (SAR) | Approximately 20 m in published coherence-change studies; narrower routes unreliable |
| Minimum detectable substation pad (optical) | Approximately 20 × 20 m at 0.5 m resolution with object-based analysis |
| Spectral bands used | Panchromatic (0.45–0.90 µm); NIR and SWIR for vegetation change; C-band SAR (5.405 GHz, Sentinel-1) |
| Cloud limitation | Optical blocked by cloud; SAR operates through cloud and partial canopy; combined approach mitigates but does not eliminate |
| Archive depth | Sentinel-1 from 2014, Sentinel-2 from 2015, WorldView-2 from 2009 (tasked archive varies by region) |
| Delivery format | GeoJSON or Shapefile vector layers; GeoTIFF change rasters; structured alert feed with parcel cross-reference |
| Latency from imagery acquisition to alert | Typically 24 to 72 hours for automated detection; 3 to 5 days for analyst-reviewed output |
Analytics Satellize can run
| New corridor detection layer | Temporal SAR coherence change detection combined with morphological linear feature extraction on Sentinel-1 and Sentinel-2 imagery | Polyline GeoJSON of new rights-of-way with confidence score and estimated clearance date range |
| Substation and transformer pad identification | Object-based image analysis on sub-metre optical imagery; spectral, textural and geometric classifiers | Polygon layer of new compound footprints with estimated commission window and voltage-class inference |
| Parcel exposure report | Spatial join of new infrastructure polygons and polylines against client-supplied or open cadastral data | Ranked list of land parcels within defined connection-zone buffers, delivered as spreadsheet or GIS layer |
| Temporal change stack | Multi-date PlanetScope or Sentinel-2 compositing to reconstruct construction timeline of a corridor | Annotated image series with dated milestones (clearance, pad pour, energisation indicators) |
| Regional screening report | Automated Sentinel-1 and Sentinel-2 change detection across a defined area of interest, followed by tiered optical tasking of flagged sites | PDF and GIS package summarising all new grid infrastructure detected within a specified date window and bounding box |
| Recurring monitoring alert | Scheduled re-run of change detection pipeline on each new Sentinel acquisition over a persistent area of interest | Email or API alert triggered when a new corridor or substation pad crosses a detection threshold, with supporting imagery chip |
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.