Conductors
Consistently high-accuracy wire extraction in internal validation, including corridors and sensors the model had not previously seen.
National Drones utility inspection service
Capture a powerline LiDAR corridor. Get back a classified, review-ready network.
National Drones combines corridor capture, LiDAR processing, expert QA, SmartData hosting, and powerline analytics in one managed delivery for utility and vegetation teams.
Available as a scoped service
Utilities, network owners, and vegetation programs
96%+
pole and tower precision in internal validation
DJI L3
and corridor-mapping LiDAR supported
< 1 day
demonstrated review-to-retrained-model loop
Powerline intelligence, built in
Trained on real networks, verified by humans
The classifier is trained and validated against expert-reviewed ground truth from live Australian corridors: distribution and transmission, open easements and heavy canopy, across multiple LiDAR sensors.
From processed L3 corridor to classified network
National Drones can process a supported DJI L3 mission into a georeferenced, hosted point cloud, then classify the same corridor for network and vegetation review.
01 Processed corridor
Georeferenced RGB point cloud ready for hosted review and downstream processing.
02 Classified network
Conductors, structures, vegetation, and ground separated while retaining their corridor context.
Poles and towers
96%+ precision in internal validation, from timber distribution poles to lattice transmission towers.
Sensor-flexible
Validated across corridor-mapping payloads and DJI L3 drone LiDAR without changing the downstream processing workflow.
Human-verified
Training and validation use expert-reviewed ground truth from live Australian distribution and transmission corridors.
Structures isolated for QA
Toggle classes independently to inspect timber poles, lattice towers, and mixed structure types before downstream analysis.
Complex structures retained
Review structure classification through changing terrain, dense vegetation, and overlapping network geometry.
Corridor-scale consistency
Move from an individual structure to the wider corridor without breaking the relationship between assets and terrain.
It gets better with your network
Start with a cold evaluation, then use one expert-reviewed section of your own corridor to tune the classifier. We have demonstrated the full loop from review to retrained model in under a day.
Operational outcomes
From point cloud to answers
Classification is the foundation. The same network can then support conductor geometry, span analysis, vegetation screening, contours, clearance envelopes, and evidence reporting for asset and vegetation teams.
01
Classified conductors, poles, towers, vegetation, and ground
02
Conductor geometry, pole clustering, and ordered spans
03
Catenary fitting and survey-day clearance context
04
Vegetation-encroachment and fuel screening
05
Worst-case thermal sag and wind blowout envelopes
06
Contours, QA notes, evidence reports, and review-ready hosted views
Worst-case clearance envelopes
Screen beyond the survey-day wire position
The worst-case clearance workflow reconstructs the observed conductor state, then screens how thermal sag and design wind assumptions can change the available margin. Results carry P50 and conservative P05 bands, uncertainty drivers, assumptions, and excluded spans.
1.6 km
demonstration corridor
14
measured spans from 391-557 m
0
screened clearance breaches in the demonstration
7.9 m
maximum hidden margin quantified
Survey-day position versus worst case
Corridor review
Move from corridor risk to span evidence
Compare survey-day conductors with modelled hot-and-wind envelopes, then select individual spans to review the clearance evidence and assumptions behind each screening result.
Clearance screening case study
A single LiDAR pass was used to separate phases, reconstruct observed conductor state, and screen 14 spans under a 100 C conductor-temperature scenario plus design wind assumptions. The demonstration reported no screened breaches and quantified up to 7.9 m of margin that was not visible from the survey-day wire position alone.
View the clearance screening case study (PDF)2-page PDF with assumptions, methodology, and span-level results.
Have a corridor to evaluate?
See the baseline on your own data before deciding how to proceed.
Scope a corridor evaluationScreening and prioritisation only
Clearance-envelope outputs are produced under stated assumptions. They are not a certified engineering assessment and do not replace detailed design tools or engineering sign-off.
Delivery workflow
One managed corridor workflow
Bring an existing corridor capture or engage National Drones to plan and fly it. We handle the path from source data to hosted evidence and a clear handover.
- 01
Scope the corridor, network type, capture standard, and decisions the inspection must support.
- 02
Capture the corridor or ingest an existing suitable LiDAR dataset, including DJI L3 where appropriate.
- 03
Process and classify conductors, structures, vegetation, and ground, then complete expert QA.
- 04
Build conductor geometry, spans, vegetation screening, contours, and clearance-envelope outputs where scoped.
- 05
Deliver review-ready SmartData views, evidence exports, assumptions, exclusions, and recommended follow-up.
Questions utilities ask
Clear boundaries around data inputs, model tuning, evaluation, and engineering follow-up.
Do we need to supply a pre-classified point cloud?
No. A cold evaluation can start from a suitable corridor LiDAR capture so the baseline classifier can be assessed before any network-specific tuning. Existing usable classes can also be retained where appropriate.
What does a cold evaluation show?
It runs representative data through the baseline workflow with no network-specific tuning. You can review conductor and structure extraction, exclusions, QA notes, and downstream analytics before deciding whether to tune the model or expand the corridor.
Can the classifier be tuned to our network?
Yes. A small expert-reviewed section can be used to tune the classifier for local structure types, vegetation, and capture characteristics. SmartData has demonstrated the review-to-retrained-model loop in under a day.
Is the clearance envelope a certified engineering assessment?
No. It is a screening and prioritisation layer under stated assumptions. It identifies spans that deserve vegetation work, field review, or certified engineering follow-up; it does not replace detailed design tools or engineering sign-off.
Put a representative corridor through the full workflow
We can begin with a cold evaluation on your own data, with no network-specific tuning, so you can see the baseline before committing to a wider program.