Find the mechanism before the formula
A LiDAR return becomes a mapped point only after its range and scan angle are combined with the platform position and attitude at the correct instant. If the scanner, GNSS and inertial clocks disagree, a perfectly measured range can be attached to the wrong place and the wrong mirror angle. The error then looks geometric even though its cause is temporal.
We will translate milliseconds into metres, separate constant offset from drift and scanner phase, calculate a phase-induced angular error, test a reversed flight line, and show why along-track smear, boresight and trajectory error need different residual evidence.
This guide belongs to the Gold Standard of Remote Sensing, the Gold Standard of Education and the How X Works hub.
Choose a reading route
Start with the question closest to the decision you need to make. Every route returns to the same worked evidence and transfer checks.
| Route | Chapters | Use it when |
|---|---|---|
| 1 | 1–3 | You need the sensor-time chain and the main timing error classes. |
| 2 | 4–5 | You want worked along-track and scanner-phase calculations. |
| 3 | 6–8 | You need drift, reversal and smear transfer checks. |
| 4 | 9–10 | You must estimate timing safely and distinguish nearby errors. |
| 5 | 11–12 | You want FAQs and a reproducible synchronisation audit. |
Chapter contents
Open the group you need, then use the chapter links. Teaching sections remain open on the page.
Build the model
Test the mechanism
1. Every point needs one defensible observation time
A georeferenced point combines laser range, mirror or scan angle, GNSS position, inertial attitude, lever arm and boresight. Those quantities are sampled by different devices. Processing must place them on a common time scale and interpolate the trajectory to the effective laser observation time.
Record whether a timestamp marks pulse emission, return detection, packet creation, scan-line start or another event. A constant labelled “latency” is not portable between systems unless its sign convention and reference event are known.
- Identify every clock and time scale.
- Define the event represented by each timestamp.
- Record offset sign and interpolation convention.
- Retain raw timing and synchronisation logs.
2. Offset, drift, phase and latency leave different patterns
A constant clock offset shifts every observation by the same time increment. Clock drift makes that offset change with elapsed time. Scanner phase error assigns the correct range to the wrong angular position in the scan cycle. GNSS or inertial latency attaches the return to a delayed trajectory state.
These terms can coexist and can be partly confounded. A strip adjustment that absorbs them into translations or rotations may improve overlap while hiding a timing defect that returns when speed, height or scan rate changes.
3. Motion turns time error into spatial error
For platform velocity vector v and a small time offset Δt, the first positional contribution is approximately Δx=vΔt. Along the flight direction, its magnitude is speed times time offset. Attitude change adds a range-dependent component because the beam is rotated using the wrong orientation.
The simple product is a diagnostic scale, not the full error budget. Terrain slope, scan angle, lever arm, attitude rate and interpolation all affect the residual seen at the ground. Preserve them rather than forcing every discrepancy into one horizontal shift.
| Error source | First diagnostic scale | Residual clue |
|---|---|---|
| constant time offset | speed × Δt | direction-linked along-track shift |
| clock drift | speed × changing Δt | residual grows with mission time |
| scanner phase | range × angular-rate × Δt | position varies through scan cycle |
| attitude latency | range × attitude-rate × Δt | height and cross-track dependence |
4. Worked example: one millisecond becomes seven centimetres
An aircraft travels at 70 m/s. If the LiDAR observation is paired with a trajectory sample one millisecond late, the first along-track displacement is 70×0.001=0.070 m. A quarter-millisecond offset gives 70×0.00025=0.0175 m.
The relationship is linear only for the positional term over a short interval. If the platform is turning or pitching, the wrong attitude can dominate. Repeat the calculation with recorded velocity and angular rate at each observation rather than applying one mission-wide correction blindly.
5. Worked example: scanner phase creates a range-dependent shift
Suppose the local mirror angular rate is 2 rad/s and the scanner timestamp is late by 0.5 ms. The angular mismatch is approximately 2×0.0005=0.001 rad. At 600 m range, the small-angle transverse scale is about 600×0.001=0.60 m.
This is an order-of-magnitude diagnostic. Oscillating mirrors do not have one constant angular rate through a full cycle, and the projection depends on scan geometry. Estimate phase with the instrument’s scan model or encoder data, then test whether residuals reverse or change magnitude across the scan.
6. Transfer check: let a small clock drift accumulate
A clock-rate error of 2 parts per million accumulates 2×10−6×1800=0.0036 s over thirty minutes if it is not corrected. At 70 m/s, the positional scale grows to 0.252 m by the end of that interval.
A single constant offset cannot remove this pattern. Fit offset and drift only when the mission has enough temporal span and independent geometry to separate them. Repeated synchronisation events should reset or constrain the model, and their timestamps belong in the audit record.
7. Opposite headings reveal the sign of a timing lag
Fly the same corridor in opposite directions. A positive time lag moves points forward along each platform heading, so the displacement changes sign in a fixed map coordinate when the heading reverses. A fixed lever-arm translation does not follow that same rule.
Use stable planar surfaces, roof ridges or surveyed targets and compare residuals by flight direction, speed and scan side. If the sign follows heading and the magnitude follows speed, a timing explanation gains support. It is evidence, not proof, because trajectory errors can also be time correlated.
8. Along-track smear is a time window, not only a time offset
If a reported sample represents an effective averaging or batching window τ while the platform moves, the along-track support is approximately speed×τ. At 70 m/s, an effective 100 μs window spans 0.007 m, or seven millimetres.
Do not call this the laser pulse length unless that is the actual mechanism. Packet batching, return aggregation, rolling acquisition and interpolation can each create an effective window. A constant timestamp correction recentres the window but does not remove its finite extent.
9. Estimate timing from raw observations and independent geometry
A defensible adjustment uses raw ranges, scan angles, trajectory states and overlap or control observations. Include time offset as an explicit parameter with the system calibration, then inspect correlations with boresight, lever arm and trajectory corrections. Strong correlation means the mission geometry cannot identify them independently.
Use cross-lines, height variation and speed variation where practical. Hold out some planes or targets from estimation and test them afterwards. A correction that improves only the fitted overlap can be overfitting the terrain or compensating for an unrelated calibration error.
10. Limits and misconceptions
- A good range measurement can still map to the wrong point when time is wrong.
- A constant offset cannot represent clock drift.
- Scanner phase is not identical to platform-trajectory latency.
- Strip adjustment can hide a timing defect without removing its cause.
- The sign of a latency depends on timestamp and software conventions.
- Low-texture or nearly parallel overlap geometry can make timing unidentifiable.
Vegetation motion, water, changing surfaces and poor return correspondence can imitate smear. Inertial bias and boresight can imitate phase-dependent residuals. Treat the error budget as a coupled model and keep uncertainty on the timing parameters.
11. Frequently asked questions
Why can a microsecond-scale error matter?
The spatial effect is multiplied by platform speed, angular rate and range. High-speed motion or long range can turn a small time error into a measurable displacement.
Can network time protocol alone guarantee point accuracy?
No. Clock synchronisation is necessary, but the processor must also know what physical event each timestamp denotes and apply sensor-specific latency and interpolation correctly.
Should time offset be corrected before strip adjustment?
Estimate the physical timing and system calibration together where possible, then use strip residuals as evidence. Pure point-cloud alignment can absorb the symptom into parameters that do not transfer to another mission.
12. A reproducible LiDAR timing checklist
| Stage | Evidence to retain |
|---|---|
| Clocks | time scales, sync events and drift specifications |
| Events | emission, return, encoder and packet timestamp definitions |
| Motion | position, attitude, speed and angular rate |
| Adjustment | offset, drift, phase and parameter correlations |
| Transfer | opposite headings, speeds, ranges and held-out targets |
| Delivery | corrected timestamps, uncertainty and residual plots |
The gold standard is to convert time into distance and angle before adjusting the point cloud, then test the correction under changed speed, heading and scan position. That is how a timing explanation becomes a transferable calibration rather than a neat overlap fit.
Reading routes · Previous chapter · Contents · Continue reading
Continue through the Gold Standard ecosystem
Return to the Gold Standard of Remote Sensing, the Gold Standard of Education parent, or the How X Works hub.
Read next in this subject
- LiDAR Strip Adjustment, Boresight Calibration and Overlap Residual Networks
- LiDAR Scan-Angle Normalisation, Footprint Stretch and Range-Walk Bias
Authoritative reading
- U.S. Geological Survey, Lidar Base Specification 2025 revision A
- ISO/TS 19159-2, calibration and validation of lidar sensors and data
- Voges, Wieghardt and Wagner, timestamp offset determination between scanner and motor
- Pöppl, Mandlburger and Pfeifer, GNSS/IMU/LiDAR integration for airborne laser scanning