How Aerial LiDAR Works: From Flight Planning to Classified Point Cloud

How an Aerometrex aerial LiDAR survey moves from flight plan to classified point cloud, and what sets the accuracy of the data you receive.

Classified LiDAR point cloud of a treed suburb, with trees, buildings and ground shown in separate colours.

Key Facts

A LiDAR scanner on an aircraft maps the shape of the land by firing laser pulses at the ground and timing each echo. Survey sensors such as Aerometrex's can fire up to 2 million pulses per second. Aerometrex then lines up the flight strips (the band of ground scanned on each pass) so their heights agree, and sorts every point into classes such as ground, trees and buildings, following national standards from the Intergovernmental Committee on Surveying and Mapping (ICSM).

TL;DR

  • Each pulse's return time gives the distance to what it hit, worked out from the known speed of light.
  • Satellite positioning (GNSS) and a motion sensor place every point, tracking the aircraft's position and tilt.
  • Flight paths are aligned, so there are no anomalies or seams where two passes overlap.
  • Classified points let you use one layer at a time, such as ground, water, trees or buildings.
  • An accuracy figure needs its confidence level, the share of points that fall inside the stated range.

Decision Matrix

Aerial LiDAR from a planeDrone LiDAR
What area does it suit?Suburbs, catchments and whole regionsSingle sites and small areas
How are flight strips aligned?Aligned and checked in processing, with the figure stated in the quoteDepends on the operator's processing
What density can it reach?Set by flight height, speed and overlap to suit the jobCan be very dense, because it flies low
What stops a flight?Cloud, fog, smoke and rainWind, rain and airspace limits

The Verdict

If you are commissioning an aerial LiDAR survey, three factors govern the positional quality of the data: ground control, flight strip alignment and the positioning of the scanner relative to global navigation satellite system (GNSS) base stations. Ground control consists of points surveyed on the ground, which Aerometrex uses to test the finished survey and adjust its heights. Strip alignment makes overlapping flight strips agree on the position of the same surface, and Aerometrex aligns its strips to within 5 cm or better. Scanner positioning depends on the base stations used and their distance from the survey area, with one station within 20 km allowing Aerometrex to place the scanner to within 2 to 5 cm. Classification is a separate measure of quality, as a point assigned to the wrong class remains positionally correct, meaning the classification type determines what the data can be used for rather than how accurately each point is positioned.

How does aerial LiDAR work?

Aerial LiDAR works by timing laser pulses fired from a scanner mounted under an aircraft. Each pulse bounces off the ground, or a roof or tree on the way, and the scanner records how long it took to come back. Light travels at a known speed, so that time gives the distance. Aerometrex's sensors fire up to 2 million pulses per second.

LiDAR stands for Light Detection and Ranging. According to the US ocean agency NOAA, its two main types use different light: topographic LiDAR maps the land with a near-infrared laser, while bathymetric LiDAR uses green light that passes into water. Aerometrex flies topographic LiDAR and does not offer bathymetric survey. A single pulse can also return more than once, bouncing back from a treetop, then if gaps in the canopy are present it can also reach a branch, then the ground. Each return becomes its own point, which is how LiDAR can measure the ground under trees, where photos cannot see it.

Diagram of a laser pulse from a plane passing through a tree, with returns at the treetop, a branch and the ground
A single laser pulse can echo back from the treetop, a branch and the ground, which is how LiDAR maps ground hidden under trees.

Millions of LiDAR points together form a point cloud: a 3D map made of dots, each with its own position and height.

Aerometrex surveys typically range from 1 to 50 points per square metre, and the job sets the density:

  • Lower density suits ground models over large areas, such as a whole region.
  • Higher density suits detail, such as building edges, power lines or small drains.

Choosing the density the job needs saves you paying for points you will not use. Aerometrex flies custom LiDAR surveys and also sells data it has already flown, and both options sit on Aerometrex's aerial LiDAR page alongside a coverage map of the existing data.

What are the parts of an aerial LiDAR system?

An aerial LiDAR system has three main parts carried on the aircraft: a laser scanner, a GNSS receiver that tracks the aircraft's position, and an inertial measurement unit (IMU) that tracks its tilt and heading. Aerometrex flies a RIEGL VQ-780 II-S scanner, which RIEGL rates to measure each distance to within 2 cm in its own tests at 250 m range; the finished survey is less exact, as the accuracy section below explains. A colour camera can also fly on the same aircraft.

Each of the three parts fixes one thing about every point, and the aircraft that carries them sets the coverage:

PartWhat it doesWhy it matters
Laser scannerFires pulses and times each echoGives the distance to each point
GNSS receiverTracks the plane's position from satellitesPlaces each point on the map
IMU (motion sensor)Records the tilt and turn of the planeTells which way each pulse was aimed
Aircraft (carries the system)Flies the scanner, GNSS receiver and IMU along planned flight linesSets the coverage and the speed of capture

GNSS positioning is more exact with a base station nearby: a fixed receiver at a known spot on the ground, used to correct errors in the satellite fix. Geoscience Australia runs networks of about 100 base stations across Australia and the South Pacific.

Positioning accuracy depends on the base stations used and how far away they are. With one station within 20 km, Aerometrex can place the scanner to within 2 to 5 cm, while several stations within 70 km give 2 to 10 cm.

Combining aerial imagery captured with LiDAR can colour the point cloud and help pick out road markings and low plants. Capturing both on one flight means the imagery and the point cloud show the site on the same day.

How is aerial LiDAR data captured and processed?

Aerial LiDAR data is captured in a planned flight, or several for a large area, then processed in 8 steps into a classified point cloud. Aerometrex plans, flies, processes and checks every survey in-house, with positional accuracy resting on the check against ground control and on strip alignment, which makes overlapping strips agree on height, while classification, which labels what each point hit, is checked as a separate step.

Before any processing, our flight planners set the flight lines, speed and overlap to meet the density the job needs, and the aircraft flies at heights from 500 to 3,100 m. ICSM's 2025 national LiDAR specification asks for at least 10% side overlap, meaning neighbouring strips share at least a tenth of their width, and that overlap is essential because strip alignment works by matching the shared ground.

After the flight, Aerometrex processes the LiDAR data in 8 steps, from download to delivery:

  1. Download the data: the raw files come off the aircraft.
  2. Extract the raw LiDAR and process the GNSS data: GNSS and IMU data track the scanner's position and tilt throughout.
  3. Check the raw LiDAR: the team aligns the flight strips and checks point density.
  4. Classify the point cloud: the points are classified into ground, vegetation, buildings and other classes
  5. Check the classified point cloud: a second quality check looks for points in the wrong class.
  6. Check against ground control: the team tests and adjusts the data against points surveyed on the ground.
  7. Make the products: the point cloud becomes a digital terrain model (DTM, a bare-earth height map), contours and more.
  8. Report and deliver: you receive the data with a project report.

Strip alignment makes the separate flight strips agree on height. Each flight line scans a strip of ground, also called a swath, and neighbouring strips overlap. A cross line called a tie run is flown across many of them to tie them together. If the strips do not line up, a flat road can show a false step between two flight lines.

Two overlapping LiDAR flight strips over a flat road: misaligned strips show a false step, aligned strips show a flat road
Where two flight strips overlap, any mismatch shows up as a false step (for example as a step in a usually flat road). Aligning the strips removes it.

Aerometrex aligns the flight strips so overlapping strips agree to 5 cm or better, using an in-house method with RiProcess and BayesMap software and its own quality checks. This 5 cm figure describes strip alignment, meaning how closely overlapping strips match each other, rather than the overall accuracy of the survey, which is set out under the uses and limits of LiDAR mapping below.

The United States Geological Survey runs the same test on its own surveys: it checks how well strips agree where they overlap, and its specification accepts a height gap of 8 cm or less in areas of swath overlap.

Aerometrex offers three classification types built on ICSM standards, each sorting the points into more detail than the last, from ground models (Type 1) to fine detail such as bridge furniture and other man-made structures (Type 3). Every classified point carries a class number from the national scheme, adapted from the LAS format (the standard LiDAR file format) of the American Society for Photogrammetry and Remote Sensing (ASPRS): 2 is ground, 3 to 5 are low, medium and high vegetation, 6 is buildings and 9 is water. Our guide to how LiDAR point clouds are classified sets out what each type includes.

TypeWhat gets sorted and checkedTypical use
Type 1Ground, noise (stray points), water, bridges and culvertsGround models, contours and flood work. A cost-effective option
Type 2Type 1, plus tall trees, buildings and infrastructureBuilding heights, tree cover and planning
Type 3Type 2, plus bridge furniture and non-building man-made structuresSpecialised projects

The classification type decides which classes are labelled, while the level sets how much manual checking those labels receive. Levels follow the ICSM Australian Airborne LiDAR Acquisition Specification, Version 2.0.

Level 1: automated and semi-automated classification.

Level 2: surface improvement using automated and manual methods.

Level 3: manual correction with significant, highly supervised effort.

Grid of LiDAR point classes against Types 1 to 3, each cell showing whether the class is checked to Level 1, 2 or 3
Within each classification type, simpler classes are checked to a lower level and trickier ones to Level 3, so type and level are set separately.

How long does it take to get aerial LiDAR data?

Aerometrex delivers aerial LiDAR data in anywhere from 24 hours to about 4 weeks, depending on whether the area has already been flown. Off-the-shelf LiDAR can be ready in as little as 24 hours, while a new capture can take about 4 weeks from flight to delivery, with weather, area size and classification type moving that timeline most.

Off-the-shelf data is fast because Aerometrex has already done the flying and processing, which makes the coverage map the first place to check: your area may already be flown.

A new LiDAR capture takes longer because each step has to run in order, and these factors move the delivery date:

  • Weather: the national LiDAR specification discourages capture through cloud, fog, smoke or rain, so Aerometrex flies in clear conditions where possible.
  • Area size: more flight lines mean more flying and more processing.
  • Class type: Type 3 needs more manual checking than Type 1, so it takes longer.

Sharing your deadline at the start lets Aerometrex set the flight plan, classification type and outputs to meet it. Councils planning a yearly program can see how Aerometrex works with urban planning and local government teams.

What are the uses and limits of LiDAR mapping?

LiDAR mapping is used wherever people need accurate ground and object heights: engineers design on it, councils plan with it and forestry teams measure trees with it. It has three limits: cloud, smoke and rain stop a flight, thick bush reduces the points that reach the ground, and every survey shows the land as it was on the day of the flight.

Engineers, councils, forestry and asset teams use LiDAR mapping for five main jobs:

  • Ground models and contours for engineering and drainage design.
  • Flood and catchment studies to see where water runs, pools and puts buildings at risk.
  • Tree height and canopy cover for councils and forestry.
  • Building heights and outlines for planning and risk work.
  • Power lines and other assets for asset management and engineering projects.

The national LiDAR specification requires capture without cloud, fog, smoke or rain. The same specification names dense vegetation as the greatest challenge for processing, and where no pulses reach the ground, the ground model fills the gap by estimating between points.

Straight from processing, before any adjustment to ground control, Aerometrex LiDAR can achieve a vertical (height) accuracy of plus or minus 20 cm at 68% confidence, meaning that at surveyed check points, about 68 in 100 checked points fall within 20 cm of the true height. When ground control is used, the figure is plus or minus 10 cm at 68% confidence, also called 1 sigma, meaning about 68 in 100 checked points fall within 10 cm. These are minimum accuracies, and the final vertical accuracy of a survey may be higher.

Two dot charts of 100 points, each with 68 inside the band: plus or minus 20 cm without ground control, 10 cm with it
A confidence level is the share of checked points that land inside the stated range: at 68% confidence, about 68 in 100 fall within plus or minus 20 cm straight from processing, or within plus or minus 10 cm with ground control.

Aerometrex LiDAR delivers a horizontal accuracy (position on the map) of plus or minus 30 cm at 68% confidence, though actual specifications vary by project. Our guide on how accurate aerial LiDAR data is gives the full detail.

Frequently Asked Questions

What is LiDAR in simple terms?

LiDAR measures distance with laser light, a bit like timing an echo. A LiDAR scanner fires pulses and times how long each one takes to bounce back, and millions of these measurements build a 3D map called a point cloud.

Can aerial LiDAR see through trees?

Aerial LiDAR can partly see through trees. Pulses pass through gaps in the leaves and some reach the ground, although in very thick bush fewer pulses get through, which leaves the ground with fewer points.

What is a classified point cloud?

A classified point cloud is a point cloud where every point carries a label, such as ground, water, tree or building. That lets you build a ground model from the data, or measure only the trees or only the buildings.

How accurate is aerial LiDAR?

Aerial LiDAR accuracy depends on the sensor, the satellite positioning (GNSS) signal, how well the strips line up and the ground control. Aerometrex delivers plus or minus 20 cm vertical at 68% confidence without ground control, or plus or minus 10 cm at 68% confidence with ground control (points surveyed on the ground). Figures from different suppliers only compare when each one states its confidence level and whether ground control was used.

Why do flight strips need to be aligned?

Each flight line is captured on its own, so small errors can make neighbouring strips disagree. Strip alignment matches the ground the strips share where they overlap, so that heights agree across the whole survey.

Is aerial LiDAR better than drone LiDAR?

Aerial LiDAR suits suburbs and regions, which a plane covers in a few flights, while a drone can suit one small site. Either way, the real test is how each option aligns its strips and checks its accuracy.

Other pages that may be of interest

blog
Understanding the Classification of LiDAR Point Clouds for Precise Data Analysis
Understanding the Classification of LiDAR Point Clouds for Precise Data Analysis
Read more
webinar
LiDAR-derived Community Resilience Toolset: A Multi-Hazard Approach
LiDAR-derived Community Resilience Toolset: A Multi-Hazard Approach
Read more
blog
LiDAR-derived Vegetation Data and Analysis
LiDAR-derived Vegetation Data and Analysis
Read more
X