VLM output
A wooden utility pole with a crossarm, ceramic insulators, cutouts and overhead conductors is shown against a clear blue sky, then a quadcopter drone with a gimbal camera flies in and hovers beside the pole, with animated blue streamers overlaid to suggest data capture. The final frame is a close-up from below as the drone passes overhead, with no worker or vehicle visible.
- Pole InspectedYes
- Pole Count1
- Visible Structural DefectNo
- Pole MaterialWood
- Inspection MethodDrone
- Overhead Conductors PresentYes
- Vegetation EncroachmentLow
- Visibility ConditionsClear
Customer challenges
At 5,000-pole scale, pole-by-pole manual inspection cannot be completed within an inspection cycle
Patrol vehicles pass the poles every day, but that video is never analysed
Corrosion, missing fuse-cutout covers and equipment defects are small objects inside cluttered street scenes
Leaning poles, downed conductors and branch contact are reported by the public after an outage, not before
One pass sees only part of each pole; which faces were actually inspected is never recorded
Defect records lack image, pole number and coordinates, so replacement planning is guesswork
Customer benefits
1. 10x inspection efficiency from patrol driving that already happens
2. Annual labour cost for one operating zone reduced by about US$130K-225K
3. Vehicle cameras cover 85-90% of pole faces, and additional vehicle and drone types close the remainder
4. AI records results and notifies anomalies automatically, and the data feeds long-term trend analysis
5. Public-risk events are shared to the city platform under the PPP, so cross-agency cases are routed instead of disputed
Key features
Bird Nest & Pole-Top Transformer Corrosion Detection (Vehicle-Mounted Cameras)
Detects bird nests and corrosion on pole-top transformers from vehicle-mounted camera footage.
Distribution Equipment Anomaly & Fuse-Cutout Cover Detection
Detects anomalies in distribution equipment, including missing or damaged fuse-cutout covers.
Public-Risk Event Detection (Leaning or Collapsed Pole, Downed Conductor, Branch Contact)
Detects public-risk events such as leaning or collapsed poles, downed conductors, and branch contact.
Before & after
| Before | After — with Linker AI Nexa |
|---|---|
| Before 1. Inspection efficiency at manual pole-by-pole baseline, 1x | After — with Linker AI Nexa 1. 10x inspection efficiency |
| Before 2. Pole-face coverage per pass, partial and unrecorded | After — with Linker AI Nexa 2. 85-90% of pole faces per pass |
| Before 3. Annual inspection labour cost per operating zone at baseline | After — with Linker AI Nexa 3. About US$130K-225K per zone per year lower |
| Before 4. Share of defects found before failure, near 0% | After — with Linker AI Nexa 4. Majority found on the routine pass before failure (target) |
| Before 5. Record completeness, notes without evidence | After — with Linker AI Nexa 5. 100% of findings with image, pole ID, GPS and defect class |
| Before 6. Public-risk events routed to the responsible party manually, case by case | After — with Linker AI Nexa 6. Automatic assignment on 100% of detections |