Artificial intelligence
Public lighting census using drones
The cost of electricity for the public lighting park is charged based on an estimate of installed consumption. The problem is that the park is often changed and the estimate becomes out of date.
With this in mind, we developed an artificial intelligence system, capable of classifying each lighting point based on images obtained by a drone.
Main advantages
✔ Possibility of autonomous data collection by the drone.
✔ WEB interface for uploading data and managing censuses.
✔ Component recognition applying artificial intelligence .
✔ Recognition of the type and power of lamps through images .
✔ Georeferencing of components through 3D reconstruction .
Advanced technology
The system uses artificial intelligence and computer vision to identify, among other characteristics, the type and power of the lamps installed in the public lighting park.
The company in charge of the census creates a project, selecting the region that needs to be censused on a map, and sends this demand to the platform.
From there, the system processes the information and identifies, in a georeferenced way, the type and power of each lamp present in the monitored region.
The employee accepts the task and prepares the drone to map the area automatically or manually. The drone flight must be carried out at night to capture the characteristics of the lighting points.
After the flight, the images are transferred from the drone to a computer, and then sent to our storage server.
As a result, the system automatically generates a census report for the region, containing the main characteristics found by artificial intelligence. This report can be used to make decisions about investments in public lighting infrastructure, energy planning and management, and other applications.
Competence at high levels
Classification is standardized and reliable
(accuracy metrics are presented).
Artificial intelligence
is in process
continuous learning
(the more points registered, the better the quality of the system).
Georeferencing
is highly accurate
(standard deviation of 5m).
Inconsistencies
and updates are
easily identified
(types and powers of changed bulbs are reported).
The census is,
on average, 20X more
faster than the method
conventional
(developed in soil).
Robotictech in the media
Automated census of the Public Lighting Park with drone and Artificial Intelligence
In partnership with EDP in the FINDESLAB industrial entrepreneurship program, we developed the “Automated census of the public lighting park with drone and artificial intelligence”.
PARTNERS:
Industrial entrepreneurship program - 2021.
Robotictech was highlighted in the Findeslab article for presenting a low-cost solution for the energy company EDP.
Join our customers
