Sensors
- TanDEM-X WorldDEM (DLR / Airbus): Bistatic X-band SAR interferometry yields a global DSM at 12 m posting with a relative vertical accuracy of approximately 2 m (1-sigma) over flat terrain. The WorldDEM4Ortho product refines this to roughly 0.4 m relative vertical accuracy at 5 m posting for targeted areas. Penetrates cloud but scatters from vegetation canopy, so heights include canopy top.
- Pleiades Neo (Airbus): 30 cm panchromatic, 1.2 m multispectral. Stereo and tri-stereo acquisitions produce DSMs with ground sampling distances of 0.5 m and vertical accuracy in the 0.3–0.5 m range (1-sigma) under good contrast conditions. Revisit at any point is theoretically daily given the four-satellite constellation.
- WorldView Legion (Maxar): 30 cm native GSD. Stereo collection supports DSM generation at 0.5 m posting with sub-metre vertical accuracy. Up to six satellites in the full constellation give revisit intervals of hours over mid-latitude targets, useful for monitoring construction progress near an aerodrome.
- SPOT 6/7 (Airbus): 1.5 m panchromatic stereo pairs produce DSMs at 1–2 m posting with vertical accuracy of roughly 1–2 m (CE90). Lower resolution than Pleiades Neo but wider swath (60 km) makes it cost-effective for mapping the full outer horizontal surface extent of a large aerodrome, which can extend 4 km beyond the runway end.
What ICAO Annex 14 actually requires you to measure
ICAO Annex 14, Volume I defines a family of imaginary surfaces around every certified aerodrome. The conical surface rises at a 5 per cent slope from the outer edge of the inner horizontal surface (itself a flat plane 45 m above aerodrome elevation, extending 4 km from the runway centreline for a Code 4 instrument runway) out to a radius of 6 km. The transitional surfaces slope at 14.3 per cent (1:7) inward from the strip edge up to the inner horizontal surface. The approach and take-off climb surfaces slope at gradients between 2 and 3.33 per cent depending on runway code. Any permanent structure whose top penetrates any of these surfaces is an obstacle by definition and requires notification to the aerodrome authority, lighting, or removal.
The practical problem for a telecoms planner is that these surfaces are three-dimensional solids defined relative to the aerodrome reference point and published runway thresholds, not to a flat map grid. A mast that appears to have adequate clearance when its height is read from a topographic map may penetrate the transitional surface when the sloped geometry is applied correctly. The calculation is straightforward trigonometry, but it requires accurate ground elevation and structure-top elevation at every candidate mast location within the obstacle limitation zone, which can extend 15 km or more along the approach path.
Why a DSM is the right starting point, and what it gets wrong
A digital surface model captures the first reflective surface: rooftops, canopy tops, mast heads. That is exactly what ICAO cares about for obstacle assessment. A 0.5 m stereo DSM from Pleiades Neo will resolve a telecommunications mast of 1 m diameter or wider, and the WorldDEM at 5 m posting will capture any structure taller than its vertical noise floor of roughly 2 m relative to its surroundings.
The complication is that a DSM cannot distinguish a permanent steel lattice tower from a temporary construction crane, a mature oak from a sapling, or a parked aircraft from a building. All three pairs look identical in height to the sensor. Bare-earth separation, using a digital terrain model derived from the same or a complementary dataset, removes the ground slope contribution, but it does not resolve the permanence question. Vegetation requires a separate classification step, typically using multispectral normalised difference vegetation index from the same Pleiades Neo collect or a coincident Sentinel-2 scene. Temporary structures require a change-detection pass against an archive DSM from a prior date. None of these steps are optional if the output is to support a regulatory submission.
TanDEM-X introduces a further subtlety: X-band radar partially penetrates sparse vegetation and reflects from within the canopy rather than the true top. For dense broadleaf canopy the height bias can reach 3–5 m relative to the optical stereo DSM. In practice, the conservative approach is to use the optical stereo DSM as the primary height source within 500 m of the candidate mast and TanDEM-X for the wider obstacle limitation zone where optical stereo coverage may be patchy.
Constructing the limitation surfaces from published aerodrome data
Every certified aerodrome publishes its reference point coordinates, aerodrome elevation, runway threshold coordinates and runway code in the national Aeronautical Information Publication. These are the only inputs needed to construct the ICAO surfaces geometrically. The construction is done in a projected coordinate system centred on the aerodrome reference point: each surface is a planar or conical polygon with a defined origin, slope and lateral extent tabulated in Annex 14 Table 4-1.
Once the surfaces are constructed as 3D solids in GIS, the DSM is queried at every grid cell within each surface footprint. The clearance at each cell is the surface elevation minus the DSM elevation. Negative clearance is a penetration. The output is a penetration raster and a vector layer of discrete obstacle objects, each attributed with its maximum penetration depth, surface type, horizontal distance from the nearest runway threshold, and confidence classification (permanent structure, vegetation, or unclassified).
Honest limits: what the method cannot guarantee
Vertical accuracy is the binding constraint. Pleiades Neo stereo achieves 0.3–0.5 m (1-sigma) under good contrast and low wind conditions, but accuracy degrades over low-texture surfaces such as flat roofs, and over targets smaller than roughly 1 m in plan. A mast with 0.4 m clearance to the conical surface is not safely compliant when the DSM carries a 0.5 m vertical uncertainty. Regulators typically require a safety buffer of at least 3 m above the DSM-derived obstacle height before a structure is declared non-penetrating, precisely because of this uncertainty.
Revisit and cloud cover matter for construction monitoring. Pleiades Neo can task a site daily, but persistent cloud over tropical or maritime aerodromes can produce gaps of weeks in the optical archive. TanDEM-X is cloud-immune but its standard global product is a snapshot from the 2011–2015 acquisition campaign and does not reflect recent construction. Commissioning a new TanDEM-X bistatic acquisition over a specific site is possible through the DLR science service, but lead times and cost must be factored into the project schedule.
Finally, the method produces a height compliance assessment, not a flight-safety determination. Formal obstacle data for aeronautical charts requires survey-grade GNSS or total-station measurement of any identified penetrating object. The satellite DSM is the screening tool that tells the surveyor where to look.
From DSM to a submission-ready obstacle register
The analytic workflow produces three deliverables. First, a penetration raster covering the full obstacle limitation zone, colour-coded by clearance margin and surface type, suitable for inclusion in a planning application. Second, a vector obstacle register listing every discrete object whose DSM-derived height exceeds the local surface elevation, with attributes conforming to the ICAO Annex 15 obstacle data format (WGS-84 position, elevation, height, accuracy class, and type code). Third, a site-specific compliance report for the proposed mast location, stating the clearance to each applicable surface and the confidence interval on that clearance given the DSM accuracy.
Satellize runs this workflow on commercial tasking paired with open Sentinel-2 multispectral data for vegetation classification. The Tonga crop-estimation programme demonstrated that combining open and commercial imagery in a single analytic pipeline is operationally straightforward; the same architecture applies here. For aerodromes in jurisdictions where the national AIP is published in ARINC 424 or AIXM format, the surface construction step can be automated from the published data feed, reducing the risk of manual digitising error.
Typical figures
| Primary DSM spatial resolution (Pleiades Neo stereo) | 0.5 m posting; native GSD 0.3 m panchromatic |
| Primary DSM vertical accuracy (Pleiades Neo stereo) | 0.3–0.5 m (1-sigma) over structured terrain with good contrast |
| Wide-area DSM resolution (TanDEM-X WorldDEM) | 5–12 m posting; relative vertical accuracy ~0.4–2 m (1-sigma) |
| Revisit for optical tasking (Pleiades Neo) | Daily at mid-latitudes; cloud-dependent for usable acquisitions |
| Minimum detectable structure height above surroundings | ~1 m (optical stereo at 0.5 m posting); ~2 m (TanDEM-X at 5 m posting) |
| Obstacle limitation zone coverage per collect | Pleiades Neo: 14 × 14 km swath; SPOT 6/7: 60 × 60 km swath |
| Archive depth (WorldDEM) | Global baseline from 2011–2015 TanDEM-X campaign; site-specific updates on request |
| Delivery formats | GeoTIFF DSM, Shapefile / GeoPackage obstacle register, PDF compliance report, AIXM 5.1 obstacle dataset on request |
| Applicable ICAO standard | Annex 14, Volume I (Aerodromes), Table 4-1 obstacle limitation surfaces |
Analytics Satellize can run
| Obstacle limitation surface 3D geometry | Parametric surface construction from published AIP runway threshold coordinates and Annex 14 Table 4-1 slope/extent tables, implemented in a projected GIS environment | GeoPackage layer set: approach, take-off, transitional, inner horizontal, conical and outer horizontal surfaces as 3D polygon solids |
| DSM-derived obstacle height raster | Stereo photogrammetry (Pleiades Neo or WorldView Legion) or SAR interferometry (TanDEM-X); DTM subtraction for above-ground height normalisation | GeoTIFF normalised surface model at 0.5 m resolution covering the full obstacle limitation zone |
| Vegetation mask for DSM correction | NDVI thresholding on Pleiades Neo multispectral or Sentinel-2 Band 8 / Band 4 ratio; object-based classification to separate woody canopy from built structures | Binary GeoTIFF vegetation mask with confidence score per object, used to flag DSM heights requiring field verification |
| Penetration analysis and clearance map | Cell-by-cell subtraction of DSM from surface elevation model; negative values flagged as penetrations; uncertainty propagated from DSM accuracy specification | Colour-coded penetration raster and vector layer of penetrating objects with clearance margin and ±1-sigma confidence interval per object |
| Discrete obstacle register | Object segmentation of penetrating cells; attribution with WGS-84 centroid, maximum elevation, height above ground, surface type, penetration depth and permanence classification | ICAO Annex 15 / AIXM 5.1-compatible obstacle data table; CSV and GeoPackage formats |
| Proposed mast compliance certificate | Point-in-surface query at mast design coordinates using proposed tip elevation; clearance computed to each applicable surface with uncertainty budget | PDF compliance report stating clearance (metres) to each ICAO surface, DSM accuracy class, and recommended survey-grade verification threshold |
| Construction-phase change detection | Bi-temporal DSM differencing between archive and newly tasked stereo collect; height-change objects above a 1 m threshold extracted and cross-referenced against the penetration model | Monthly GeoTIFF change layer and alert table of new or heightened structures within the obstacle limitation zone |
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.