Street lighting is an important aspect of modern life. The primary function being to extend the number of light hours to allow for activities to continue past sunset, especially in the darker winter months. In addition to this, street lights can be used to promote security in urban areas, as well as generally making roads and pathways safer to use. As such it is important that any issues with street lights are resolved as soon as possible to prevent a possible accident.
South Gloucestershire Council currently have their street lighting programmed to turn on 15 minutes before sunset and off 15 minutes after sunrise. While individuals can notify the council when a light stops functioning, the common approach to monitoring functionality is by having street teams run periodic manual checks – roughly every four weeks. These checks involve driving along stretches of road to check if street lights are showing normal behaviour, i.e. turning on and off at the times they are supposed to. The majority of street lights observed through this method function correctly, meaning that the costs of running such observations (man hours, vehicle maintenance and fuel) are inefficient.
The aim of the project is to monitor the functionality of street lights using UMBRELLA nodes with camera nodes attached to the top, pointing upwards towards the streetlight and sky.
The camera would be used to collect images of the streetlights at various times to train a machine learning model, that is then used to determine if a street light is turning on and off at the appropriate times.
This will then allow the street care team to be able to check the status of the street lights in real time, without having to travel to the street lights themselves.
Once the machine learning algorithm detects that a street light isn’t working as intended, an alert is sent to the street care team, meaning they can monitor street lights effectively and passively whilst undertaking other tasks.
- The reduction in man hours, vehicle maintenance and fuel will translate into cost reductions for the street care team.
- Less fuel burned unnecessarily means less pollution emitted to the environment.
- The ability to monitor the real time status of any connected street light without needing to visit it.
- To be automatically alerted when a street light isn’t operating normally.