Geospatial jargon decoded: GIS, GNSS, LiDAR, DEM and the rest

A mapping proposal can read like alphabet soup: “We will acquire LiDAR, produce a DSM and DTM, georeference the orthomosaic to WGS 84 and serve it through WMS.” To a specialist, that is one clear sentence. To everyone else, it is a wall. Fortunately, geospatial jargon is built from a small set of ideas, and once you know them, the wall comes down.

This article groups the terms by the question they answer: where, with what, what data and what tools.

Where: positioning and coordinates

• GIS (geographic information system). Software and methods for storing, analysing and displaying spatial data.
• Geospatial. Anything connected with location on the Earth.
• GNSS. Global navigation satellite systems, the general term for satellite positioning, including GPS (United States), GLONASS, Galileo and BeiDou. GPS is one of them.
• RTK and PPK. Methods that correct satellite positioning to centimetre accuracy, in real time or after the fact.
• Coordinate reference system (CRS). The agreement that turns numbers into places.
• Datum. The reference model of the Earth’s shape that coordinates are measured against. WGS 84 is the global standard.
• Projection. A method of flattening the curved Earth onto a flat map, which always distorts something.
• EPSG code. A short number that identifies a coordinate system, for example 4326 for WGS 84 latitude and longitude.
• Georeferencing. Giving an image or map real-world coordinates.
• Geocoding. Turning an address into coordinates. Reverse geocoding does the opposite.
• Ground control point (GCP). A surveyed mark on the ground used to tie images or models to exact coordinates.

Data types

• Vector data. Points, lines and polygons with attributes, such as roads, buildings and boundaries.
• Raster data. Grids of cells with values, such as satellite images and elevation.
• Shapefile, GeoJSON, GeoPackage. Common vector file formats. GeoTIFF is the standard raster one.
• Attribute table. The data attached to each feature, like a spreadsheet.
• Layer. One set of data in a map, which can be switched on and off.
• Basemap. The background map, such as streets or satellite imagery, that your data sits on.
• Feature. One thing in a vector layer, such as a single building.
• Metadata. A description of a dataset: source, date, accuracy and licence.

Capturing: sensors and platforms

• Remote sensing. Gathering information about the Earth from a distance, usually from satellites, aircraft or drones.
• Optical imagery. Pictures taken in visible and near-infrared light.
• SAR (synthetic aperture radar). Radar imaging that works day and night, through cloud.
• LiDAR. A laser scanner that measures distance by timing light pulses, producing accurate 3D point clouds, including ground under trees.
• UAV or drone. An unmanned aircraft for mapping and inspection.
• Multispectral and hyperspectral. Sensors recording a few, or hundreds of, narrow colour bands.
• Thermal imaging. Measuring surface temperature.
• Resolution. Spatial resolution is the size of one pixel on the ground (GSD). Temporal resolution is how often the place is imaged. Spectral resolution is how many bands.
• Revisit time. How often a satellite images the same spot.
• Swath. The width of the strip a sensor sees.

Surfaces and heights

• DEM (digital elevation model). A general term for a raster of heights.
• DSM (digital surface model). Heights of everything on top of the ground: roofs, treetops, vehicles.
• DTM or DEM (terrain). Heights of the bare ground, with buildings and vegetation removed.
• nDSM or CHM. The difference between surface and terrain: the height of objects above the ground, such as tree height.
• Contour. A line joining points of equal height.
• Slope and aspect. The steepness of the ground, and the direction it faces.
• Point cloud. A set of 3D points, from a laser scan or photogrammetry.
• Mesh. A surface of triangles built from points.
• Orthomosaic (orthophoto). A map-accurate image stitched from many photographs, corrected for terrain and tilt.
• Photogrammetry. Measuring from photographs, including building 3D models from overlapping images.

Analysis and products

• Buffer. An area drawn at a set distance around a feature.
• Overlay. Combining layers to see where they overlap.
• Spatial join. Linking data by location, such as counting wells in each district.
• Interpolation. Estimating values between measured points, such as a rainfall surface from rain gauges.
• Classification. Assigning categories to pixels or objects, such as land cover types.
• Change detection. Comparing two dates to find what changed.
• NDVI. A vegetation index made from red and near-infrared light.
• Land use and land cover (LULC). What is on the ground, and what it is used for.
• Geostatistics. Statistics that account for the fact that nearby places are more alike.
• Spatial analysis. Any method that uses location to answer a question.
• GeoAI. Applying machine learning to geospatial data.
• Digital twin. A 3D model linked to live data about a real place.

Sharing maps on the web

• Web map. An interactive map in a browser.
• Tiles. A map cut into small squares that load as you zoom and pan. Vector tiles carry the data, and raster tiles carry pictures.
• WMS, WFS, WMTS. Standard services for serving map images, features and tiles over the web.
• API and SDK for mapping. Tools for putting maps into websites and apps, such as MapLibre.
• STAC. A standard way of describing and searching catalogues of satellite and aerial imagery.
• Cloud-optimised GeoTIFF (COG). A raster format designed so that software can read just the part it needs over the internet.
• Open data. Data published for anyone to use, under an open licence.
• OGC. The Open Geospatial Consortium, which publishes many of these standards.

Accuracy words that are easy to mix up

• Accuracy. How close a measurement is to the truth.
• Precision. How consistent repeated measurements are, which can be high even when accuracy is poor.
• Absolute and relative accuracy. How well positions match the real world, versus how well they match each other.
• RMSE. A common way of summarising the typical size of errors.
• Resolution is not accuracy. A sharp image can still be in the wrong place.

Decoding the opening sentence

Back to the proposal: “We will acquire LiDAR, produce a DSM and DTM, georeference the orthomosaic to WGS 84 and serve it through WMS.” In plain words: we will scan the area with a laser, make one height map including buildings and trees and another of bare ground, line up the stitched aerial image with real-world coordinates in the standard global system and publish it as a map service that other software can use.

Frequently asked questions

Is GPS the same as GNSS?

GPS is one satellite system. GNSS is the umbrella term for all of them. Modern receivers use several at once, for better accuracy.

What is the difference between a DSM and a DTM?

A DSM includes everything on the surface, such as buildings and trees. A DTM is the bare ground. Subtract one from the other, and you get object heights.

Why do I need a ground control point?

To tie a model or image to exact real-world coordinates, and to check its accuracy. Without them, a result can be well shaped yet misplaced.

Vector or raster?

Use vector for things with edges and attributes, such as roads and parcels. Use raster for continuous surfaces and images.

Spatial vocabulary looks forbidding, but it describes a very human question: what is where, and how sure are we? Learn the handful of core ideas, and you can read a map proposal with the confidence of an insider.