BEAM (Building & Establishment Automated Mapper)

BEAM is an AI tool that uses machine learning to automatically map rooftops of buildings in informal settlements for a variety of communal benefits.

The tool is simple and easy to use. It allows users to quickly detect and visualise building footprints and structures in a specific area by uploading aerial images of a given location. Through four easy steps, it predicts which buildings are present in an image by identifying and geo-referencing the pixels that are likely rooftops.

The version used by eThekwini and the City of Cape Town is designed to be run and hosted locally by the city, with the developments in the Central America project optimized for running on the cloud.

View the BEAM user manual PDF

Read about the participatory process in eThekwini


Buildings mapped using BEAM in Guatemala City and Santo Domingo


Who is it for?

  • Local and national governments
  • Urban planners and managers
  • Community organizations

BEAM Interface

Desktop view of BEAM tool processing aerial images

"BEAM is one of the tools which will enable us to better understand the trends in urban development in the city. It will also create better efficiencies and create a platform for sharing of data so that we can improve our delivery of basic services… to those excluded from the formal areas of the city." Sarah Watson, Human Settlements Unit, eThekwini Municipality

Want to use the tool in your region?

Contact unitac@un.org for more information about how to use the tool in your region.

Partners

  • The Human Settlements Department, eThekwini Municipality, South Africa
  • The City of Cape Town Geomatics Branch, South Africa
  • UN-Habitat Mexico
Features icon

Features and functionality

BEAM quickly generates shapefiles of rooftops to help municipalities to:

  • maintain up-to-date records and maps
  • better prioritize upgrading interventions
  • improve basic service delivery
  • Pixel level accuracy on high resolution aerial imagery: Over 90%
  • Pixel level accuracy on high resolution satellite imagery: Over 84%
  • Fast image processing: 1 image takes 70-90 seconds (depending on the computational power)
  • To effectively map building footprints BEAM requires high-resolution aerial or satellite imagery

Scalability

  • The tool’s open-source nature and user-friendly interface facilitates widespread adoption, empowering local governments to integrate it into their urban planning workflows and allowing it to be applied across different cities, contexts and use cases.
  • Access to training data, particularly imagery of the area of interest, is required to ensure accurate and context-specific results.
  • Initially deployed in South Africa, BEAM was subsequently scaled for implementation in eight Central American capitals namely Guatemala City, Tegucigalpa, Belize City, San Salvador, Santo Domingo, San José, Managua, Panama City. UNITAC is currently preparing its deployment in Namibia.