Hazard tree detection for power lines
Hazard tree detection finds the trees that could fall into a power line, including danger trees that stand outside the right-of-way. In GridGuardian the satellite screens every span for tall canopy beyond the clearance band, LiDAR or stereo imagery measures which trees are tall enough to reach a conductor, and the confirmed hazard trees go to crews as work packages in FieldGuard.
Why do trees outside the right-of-way matter?
Vegetation is the most common cause of power outages in the U.S., at more than twenty percent of incidents, according to the U.S. Department of Energy.[1] The same guide notes that in extreme events the most damage often comes from plants outside the utility right-of-way.[1]
The NERC transmission vegetation standard counts fall-ins from outside the right-of-way as their own outage category, apart from grow-ins.[2]
Clearance trimming inside the corridor does not touch these trees. Managing them takes two numbers per tree: its height and its distance to the nearest conductor.
Which carrier does what, from screening to proof
The open satellite looks at everything; the expensive tools go only where it points.
- Satellite · base resolution
Screen every span for tall canopy
Our own neural network estimates vegetation height in bands from Sentinel-2, calibrated with LiDAR. It flags spans where tall-band canopy stands in the environmental zone, beyond the clearance band.
Height bands, not metres. A 10 m pixel[3] cannot isolate a single tree.
- LiDAR or stereo satellite
Measure fall-in reach where flagged
On flagged spans, LiDAR (drone, aircraft or your own uploaded survey) or 0.3–0.5 m stereo imagery from the data marketplace gives the height of each tree and its distance to the conductor. A tree whose height exceeds that distance can reach the line.
The geometric test does not know lean, rot or root condition.
- Drone or vehicle
Look at the tree itself
Drone photos show dead crowns, broken tops and leaning stems on single trees. Along roads, the vehicle camera covers roadside trees in the same pass as the pole inventory.
Internal decay and root damage need an arborist on the ground.
- GridGuardian
Rank by risk class
Exposure times hazard puts every span into HC0–HC3. A tall tree beside a critical feeder ranks above the same tree beside a quiet rural spur.
- GridGuardian → FieldGuard
Work package and crew
Hazard trees are grouped per span into a work package with location, height and reason. The crew works offline in FieldGuard, confirms or rejects each tree, and records removal or topping with before and after photos.
- Satellite · next pass
Verify
The next usable pass compares the span with its own history.
Removing one tree may not change a 10 m pixel; then the photo evidence, or a targeted high-resolution or LiDAR re-check, is the proof.
What each signal tells you about a hazard tree
| Signal | Data source | What it tells you | Limit |
|---|---|---|---|
| Canopy height band in the environmental zone | Sentinel-2 + own neural network | Which spans have tall vegetation beyond the clearance band | Bands, not metres; 10 m pixels; cloud-free passes only |
| Tree height and distance to conductor | LiDAR (drone, aircraft, uploaded) | Exact fall-in reach per tree | Cost per km; a snapshot that ages as trees grow |
| Tree height from stereo | High-resolution satellite, 0.3–0.5 m | Single trees and heights without a flight | Bought per area with a minimum order; needs a clear day |
| Crown condition | Drone photo, vehicle camera | Dead, broken or leaning trees | Only what the camera sees; not decay or roots |
| Ground assessment | FieldGuard crew | Confirmation, action taken, evidence | Only on spans the crew visits |
More on choosing between carriers: satellite vs LiDAR vs drone.
Which platform links hazard tree detection to crew work orders and proof of completion?
With us you find all of it in one place. In GridGuardian the detection, the risk class, the work package and the evidence sit on the same span. FieldGuard takes the package to the crew, works offline, and returns located, time-stamped photos and a completion check.
Results export to SAP PM, Power BI, file or API.
- Detection and risk class (HC0–HC3) on the span and its poles, with your own identifiers
- Work package to the crew or contractor in FieldGuard
- Offline in the field; photos with location and time; signature
- Completion check in the office, then the next satellite pass
- Exports to SAP PM, Power BI, file or API
Hazard tree questions
What is a hazard tree near a power line?
A hazard tree, or danger tree, is a tree that could strike a conductor if it falls or drops a large limb. Many stand outside the right-of-way, so clearance trimming inside the corridor does not remove them.
What is the difference between a grow-in and a fall-in?
A grow-in is vegetation growing into the clearance distance of a line. A fall-in is a tree or part of a tree falling onto the line. NERC FAC-003-5 counts them as separate outage categories.
Can satellites detect hazard trees?
Partly. Through our own neural network, Sentinel-2 shows where tall canopy stands near every span, in height bands, which is enough to rank spans. For the exact height of a single tree, LiDAR or stereo high-resolution imagery is needed.
How is fall-in reach calculated?
The simple test compares each tree's height with its distance from the trunk to the nearest conductor. If the tree is taller than that distance, it can reach the line. It does not know lean or condition; the drone photo, the crew or an arborist checks those.
Do we need LiDAR on the whole network?
No. The satellite screens every span continuously, and LiDAR or stereo imagery is ordered only for the spans it flags. LiDAR processing is priced at $90 per mile or €50 per km; satellite screening starts at $9 per mile or €5 per km per year.
How do we prove that a hazard tree was removed?
The crew records it in FieldGuard with before and after photos, location and time, on the span. The next satellite pass checks the span against its own history; where one tree is too small for a 10 m pixel, the photos or a targeted re-measurement are the proof.
Try it on your own network
Send about 100 km of corridor and its network geometry. The base system is set up on it at no cost, and you can check the results in the field with FieldGuard.
- U.S. Department of Energy, Grid Deployment Office (prepared by Lawrence Berkeley National Laboratory): Vegetation Management. Resilience Investment Guide, September 2024. https://www.energy.gov/sites/default/files/2024-11/111524_Vegetation_Management.pdf
- NERC Reliability Standard FAC-003-5, Transmission Vegetation Management. https://www.nerc.com/pa/Stand/Reliability%20Standards/FAC-003-5.pdf
- ESA / Copernicus SentiWiki: Sentinel-2 mission overview (revisit and spatial resolution). https://sentiwiki.copernicus.eu/web/s2-mission