All the drone vocabulary, in plain terms.

LiDAR emits up to one million laser pulses per second and times the return echo of each one. Paired with on-board GNSS-RTK and IMU, every returned point is positioned in XYZ coordinates with typical 2-5 cm accuracy. The ability to record multiple returns per pulse (canopy, undergrowth, ground) makes LiDAR the only technology able to reconstruct the bare ground under forest cover without human intervention.

See also : DTM, CHM, GNSS-RTK, IMU

GNSS-RTK combines a classical satellite receiver with a differential correction broadcast by a reference station (or RTK network). Absolute position moves from several metres to 1-2 cm in real time. This is the technology that makes airborne photogrammetry and LiDAR usable for centimetric topographic work. Without RTK, absolute accuracy drops to 30-50 cm.

See also : GCP, LiDAR, Photogrammetry

The IMU combines accelerometers and gyroscopes to continuously measure drone orientation across three axes. This data, fused with GNSS-RTK, lets us know at every instant not only where the drone is but also where it's looking. Essential to correctly georeference each LiDAR point and each photogrammetric pixel.

See also : GNSS-RTK, LiDAR

A multispectral sensor captures 4 to 10 spectral bands simultaneously (red, green, blue, near infrared, red edge). It's the basis of every agricultural vigour index (NDVI, NDRE, SAVI). In agriculture, we mount it on a drone to map water and plant-health stress before it becomes visible to the naked eye.

See also : NDVI, NDRE

Thermal sensors measure infrared radiation emitted by surfaces and reconstruct a temperature map. Drone-mounted, they're used for thermographic inspection of photovoltaic panels, power towers, insulated roofs and HV connectors. Typical thermal resolution is 0.05 °C, enough to detect connection defects or faulty PV modules.

See also : Thermography

Aerial photogrammetry reconstructs a site's 3D geometry by cross-referencing hundreds of photos taken from the drone at different angles. Typical accuracy: 2-5 cm XY and 3-8 cm Z, provided a GCP-RTK protocol. Cheaper than LiDAR, it adds colour texture as a bonus, but does not penetrate vegetation cover.

See also : LiDAR, Orthophoto, GSD, GCP

GSD measures the real size that one pixel of the orthophoto represents on the ground, in cm. The smaller the GSD, the more detailed the orthophoto. At 2 cm/pixel, you read a road crack; at 10 cm/pixel, you can no longer distinguish mortar from joints. GSD depends on flight altitude, focal length and sensor. OHM WORKS standard: < 2 cm/pixel on precision missions.

See also : Photogrammetry, Orthophoto

A GCP target is a visible mark (painted cross, plastic checkerboard) laid out before the flight and georeferenced via GNSS-RTK. 5 to 8 well-distributed targets are enough to anchor an entire point cloud in the client's coordinate system with 2-3 cm absolute XY accuracy and 5 cm Z accuracy. Without GCPs, absolute accuracy drops to 30-50 cm.

See also : GNSS-RTK, Photogrammetry, LiDAR

The DTM (MNT in French) models the bare ground surface, without vegetation or buildings. It is extracted from the 'ground' class of the LiDAR point cloud (or directly from the photogrammetric cloud on open terrain). Typical Z accuracy: 2-5 cm with LiDAR, 3-8 cm with photogrammetry. Serves as the basis for road profiles, hydraulic studies, volumetric calculations and cadastral plans.

See also : DSM, CHM, LiDAR, Photogrammetry

The DSM (MNS in French) models the surface as seen from above, including vegetation canopy, buildings and all elevated elements. Unlike the DTM, it does not penetrate trees. For site monitoring or urban analysis, the DSM is often the priority deliverable. The DSM - DTM difference gives the CHM (canopy height model).

See also : DTM, CHM

The CHM (MNH in French) gives the real height of each tree to the cm. Computed by differencing the DSM (canopy) and DTM (bare ground), it feeds forest inventories, biomass calculations, detection of trees exceeding electrical clearance zones, and temporal tracking of plantation growth.

See also : DTM, DSM, LiDAR

An orthophoto is a mosaic of aerial images corrected to remove perspective distortion and georeferenced. Each pixel carries cm-accurate XY coordinates. Serves as a cartographic base for planning, cadastre, site monitoring. Typically delivered as GeoTIFF with 2-5 cm/pixel GSD.

See also : Photogrammetry, GSD, GeoTIFF

A digital twin reproduces an existing structure (telecom tower, building, HV line) as textured 3D with georeferenced dimensions and components. On a BTS tower, the twin lets you measure each antenna's azimuth and tilt without climbing. On a building, it feeds preventive maintenance and renovation work. Typical formats: OBJ, GLB, IFC.

See also : Photogrammetry, IFC, OBJ / FBX / GLB

A point cloud is the base data produced by a LiDAR sensor or dense photogrammetric reconstruction. Each point carries XYZ + intensity + colour (RGB) + classification (ground, vegetation, built…). Typical useful density: 10-50 pts/m² in LiDAR, 30-100 pts/m² in photogrammetry. Delivered as LAS/LAZ for integration in CloudCompare, ArcGIS, QGIS, Metashape.

See also : LiDAR, Photogrammetry, LAS / LAZ

NDVI ranges from -1 to +1: negative values = water or bare soil, 0 to 0.3 = sparsely vegetated, 0.4 to 0.8 = vigorous vegetation. Computed pixel by pixel from the Red and NIR bands of a multispectral sensor, it detects water or plant-health stress 2 to 4 weeks before symptoms become visible to the naked eye.

See also : Multispectral sensor, NDRE

NDRE uses the Red Edge band instead of the Red band. More sensitive than NDVI on already-dense or mature crops (where NDVI saturates), it is preferred for plantations closing canopy and for early stress detection on orchard crops.

See also : NDVI, Multispectral sensor

Drone volumetrics compute a stockpile's volume by comparing the photogrammetric 3D mesh against a reference plane. Typical tolerance: ± 1 % with a GCP-RTK + double-grid protocol, vs ± 2-3 % for an interpolated ground survey. Lets quarry and mining operators move from quarterly to monthly inventory.

See also : Photogrammetry, GCP

Aerial thermography reveals temperature gaps invisible to the eye: hotspots on HV connectors, defective photovoltaic modules, roof thermal bridges, overheating technical bays. Essential for predictive audit of power networks and solar installations. Typical thermal resolution: 0.05 °C.

See also : Thermal

On a centimetric orthophoto, object-detection models (specialised convolutional neural networks) automatically identify each palm tree, each tower, each crack or each defective PV module. Typical error margin on oil palm: < 2 % on spaced rows, 3-5 % on dense plantation. Each detection is delivered with a confidence score and geolocated.

See also : Orthophoto

LAS is the ASPRS-standard binary format for LiDAR point clouds. It carries XYZ, intensity, classification, RGB, return number and all LiDAR attributes. LAZ is its compressed version (lossless, ~80 % ratio). Natively read by CloudCompare, QGIS, ArcGIS, Metashape and most modern GIS tools.

See also : Point cloud, LiDAR

GeoTIFF is a TIFF augmented with a georeferencing header (projection system, corner coordinates, resolution). It's the universal format to deliver an orthophoto, DTM, CHM or NDVI map. Compatible with all GIS tools (QGIS, ArcGIS, GlobalMapper) and modern web viewers.

See also : Orthophoto, DTM

The Shapefile is the historic vector format to carry geolocated features: one point per inventoried palm tree, one polygon per parcel, one line per HV cable. Aging against GeoPackage and GeoJSON, it remains the de-facto standard in most enterprise GIS.

See also : GeoJSON

GeoJSON is a JSON-based text format to carry vector features (points, lines, polygons) with their attributes and projection system. Natively readable by browsers and all modern web libraries (Leaflet, MapLibre, Mapbox GL). Preferred when deliverables must integrate into a web platform or CMMS.

See also : SHP

IFC is the open standard BIM format. It carries 3D geometry, semantics (wall, slab, antenna, equipment), properties and object relationships. Lets you integrate a drone digital twin into an existing BIM workflow without depending on a proprietary editor.

See also : Digital twin

OBJ is the most universal 3D format (readable by all 3D software). FBX (Autodesk) additionally carries animations. GLB is the binary glTF format, optimised for web delivery and in-browser 3D visualisation (used for the Place de l'Amazone 3D model on the home page). All three carry geometry + textures.

See also : Digital twin

EPSG is the registry maintained by OGP that assigns a unique identifier to each projection system. Common codes: EPSG:4326 (WGS84 worldwide geographic), EPSG:32631 (UTM Zone 31N for Benin and West Africa), EPSG:2154 (Lambert 93 for mainland France). We always specify the EPSG code so no projection ambiguity remains in the deliverable.

See also : UTM Zone 31N

UTM Zone 31N is the standard metric projection for the zone covering Benin, Togo, southern Nigeria and western Cameroon. Coordinates in metres, ideal for measuring distances and surfaces directly. OHM WORKS delivers in UTM 31N (EPSG:32631) by default, but adapts to specific client datums (notably mining ones).

See also : EPSG

ANAC Bénin is the Beninese authority in charge of civil drone flight regulation and authorisation. Every professional operator must be registered, pilots certified, and missions declared upfront. OHM WORKS holds full ANAC authorisations for flights across the entire Beninese territory.

See also : DGAC France, SORA, NOTAM

DGAC is the French authority regulating civil aviation, including professional drone flights (Open, Specific, Certified categories). OHM WORKS pilots are DGAC France certified, validating mastery of the harmonised European procedures (EASA) partially applicable in Benin by equivalence.

See also : ANAC Bénin, SORA

SORA is the EASA methodology to assess the operational risk of a drone mission in the Specific category (flight over people, urban area, beyond visual line of sight, etc.). It structures the ground-risk and air-risk analysis, and the definition of mitigation measures. OHM WORKS produces a complete SORA analysis whenever a mission justifies it.

See also : ANAC Bénin, NOTAM

A NOTAM is an official notice broadcast to airmen to signal a specific aerial activity in a given airspace at a precise time. For drone flights in regulated zones or near an airport, the operator publishes a NOTAM via ANAC (Benin) or DGAC (France) to inform aircraft and helicopter pilots.

See also : ANAC Bénin, DGAC France, SORA