The Poles AI Pack includes any pole with either a light, or power lines leading to it.
As well as being available for visualization as raster AI Layers, they have been merged as a point object, where each point represents the location of a pole, and flags indicate the presence (and confidence) of both power and light attachments. The data included in the "Pole" feature class includes:
- Latitude/longitude of pole centroid
- Presence of Power (multiple power lines leading to the pole)
- Presence of Light (presence of a light on the pole)
3x confidence scores:
- Confidence that it is a pole containing at least power OR light
- Confidence in the presence of power
- Confidence in the presence of light
- Areas, giving approximate vertical area of the pole. This helps distinguish between larger transmission towers and groups of poles in substations, and ordinary ones
AI Feature API
(Gen 1 & 2 Data)
| Metadata for power and light presence attached to pole centroids.
| Not available in Gen 1 / 2 Data
Characteristics and recommended use
Pole locations can be used for a variety of use cases, from identifying locations for networking equipment to identifying risk of electrical hazards for property insurance or drone flight.
Some areas with underground power will still have a large number of light poles, which can provide benefit in analyzing the adequacy of street lighting in a neighborhood.
Pole centroids are typically accurate within about half a meter (RMSE). This accuracy is lower than our imagery GSD figures as poles are subject to parallax error, and the head of the pole may obscure the precise location with large cross-arms or protruding lights.
The most common causes of error are overhanging vegetation and extensive shadowing. Given poles are typically not built/removed on a regular basis, we recommend choosing dates where "leaf off" surveys are available, and the sharpest imagery with the best lighting in the area is chosen for export. It would even be possible to request multiple dates of results to merge.