Building and solar-panel detection, with contacts
Computer-vision project to detect buildings and solar panels within a date range, then retrieve the building owner’s contact details.
96%
detection accuracy
150k
buildings scanned
70%
owner contacts recovered
±30 d
configurable time window
The goal
For a given area and date range, identify visible buildings and solar panels, then attach each detection to the contact details of the person who owns the building.
The deliverable is not a pretty map: it is a list of actionable targets for prospecting or rollout tracking.
Visual detection
A computer-vision model locates buildings and panels on imagery available in the time window. Each detection carries confidence and geolocation.
The date constraint targets new installs, roof changes, or a time-bounded campaign — not a frozen inventory.
Panel detected
96%- Building and panel detection
- Filtering by date range
- Confidence score per detection
- Target geolocation
From detection to contact
Once the building is identified, the pipeline enriches the record with the owner’s contact details — so sales or partnerships can engage without manual research.
Vision + owner data = a complete solar lead, not just a pin on a satellite photo.
Outcome
A territorial prospecting tool that combines computer vision and contact enrichment, time-bounded, to prioritize the right buildings — and the right people.