Paris lighting simulation

The project is still in its early stages, consider this more as a technical demo than as a physically accurate simulation. Contributions are welcome!

The aim of the project is to simulate and provide a visual representation of the lighting of Paris streets by performing a 2D illuminance calculation at street level. The position of street light sources and their characteristics is pulled from a public Paris Data database.

Lighting simulation of Place de l'Odéon.
Notice the illuminance reading on the top right.

Simulation of two light poles side to side.

Live demo | Place de l'Odéon (100m)
(Click anywhere to add a light source)

Features

• Simulate and visualise the lighting of any given area in Paris in the form of an interactive web map
• Read the calculated illuminance of a point on the street by hovering the pointer over it
• Upcoming: calculate the uniformity of the lighting of an area

Installation and usage

• Requirements:

  • Python 3
  • A console application and some basic command-line knowledge

• Installation:

  • Clone/download the repository
  • Install the python libraries in requirements.txt:
    pip install -r requirements.txt
  • Download the Paris Data public lighting database in CSV format here and copy it in the directory DB

• Usage:

  • In the python directory, use:
    python map.py -l "[location]"
    For example : python map.py -l "Place de l'Odéon"
  • The generated map is located at out/map.html

• Optional:

  • The option -d can be used to specify the radius of the simulation in meters (default is 100 meters)
  • The option --cellSize can be used to specify the size of the cells (default is 4 pixels). The larger they are, the faster the rendering is.

How does it work?

• Physics

  • Illuminance of a point by a lighting appliance is calculated with Lambert's cosine law, taking into account the height of the lighting appliance and its luminous flux.

• Simulation and visualisation

  • For each lighting appliance, the area that it can illuminate is determined by raycasting against buildings.
  • The screen is divided into a grid. For each cell its total illuminance is calculated by adding the illuminance of each lighting appliance that illuminate the cell (using the fact that illuminance is additive).

Limitations

• Physical accuracy:

  • Paris Data public lighting database only provides data on the luminous flux, so for now a uniform luminous intensity distribution is assumed along with a 75° beam cutoff angle. The solution to increase the accuracy of the simulation would be to use luminous intensity distribution diagrams. Those would need to be experimentally measured on real installations to reflect the effect of the reflector and the disposition of the light source.
  • For some lighting appliances, including most wall-mounted appliances the height data is missing in the Paris Data database. In this case a height of 10 meters is assumed, but it is not always accurate.

• Performance : For now the intended usage is to simulate lighting in a radius of about 100 meters around a given point. Simulation of larger areas or areas densely illuminated can be very slow at the moment.

What does it use?

• Mapping framework
  • Leaflet is used as the mapping framework.
  • The simulation/visualisation is written as a leaflet plugin in javascript. The rendering is based on the Leaflet IDW plugin.
• Databases
  • Paris Data public lighting database is used to get the coordinates, height and luminous flux of lighting appliances.
  • OSMnx is used to query Open Street Map for building footprints
• Data query and map generation
  • Python is used to query the databases and generate the map with Folium.

How is the project organized?

• The python part, located in the directory python handles extracting data from the databases and including it in the map generated with Folium.
• The generated map, in html format, relies on three javascript files present in the out directory:
  • leaflet-illuminance.js is the leaflet plugin and handles the leaflet integration along with the rendering.
  • illuUtils.js contains the core functions to simulate the lighting and calculate the illuminance of a point
  • extra.js includes the javascript dependencies listed earlier. It can be recreated by doing the following:
    • Download the javascript dependencies with npm install (requires Node.js/NPM)
    • Bundle them with browserify dependencies.js --s Extra --outfile out/extra.js

Possible improvements

• A feature to calculate the uniformity of the lighting of a given area
• Performance improvements
• Implement streaming the data from the databases instead of having it included it in the html file, so that the python generation isn't needed anymore
• Implement a non-linear gradient to accurately display high illuminance values (for now the max illuminance value can be set with --gradientMax, default is 80 lux)
• Obtain luminous intensity distribution diagrams for common light appliances to increase the simulation accuracy

Don't hesitate to contribute or to contact me!