Searching for Slums with Machine Learning

Machine learning techniques are being increasingly used to detect features in satellite and aerial imagery. Artificial intelligence and machine learning can be used to train algorithms to search for familiar patterns in aerial imagery of the Earth. The reasons for searching satellite imagery can be varied and can be for commercial, environmental or social purposes.

One example of machine learning being used to identify common features in aerial imagery is OneSoil, which uses AI to detect where different types of crops are being grown around the world. Another example is Земляна проказа, which was created using machine learning to identify Ukraine's illegal amber mines. Another example, recently covered on Maps Mania, is Curio Canopy, which has used machine learning based techniques to identify tree canopy cover in European cities.

Another example is Dymaxion Labs Maps of Potential Slums and Informal Settlements. Dymaxion Labs used machine learning to search the satellite imagery of a number of South American cities in order to identify and find slums and informal settlements. The resulting maps are being used to help urban planners and local councils identify where vital utilities need to be directed.

To help identify the informal settlements Dymaxion Labs used the Random Forest machine learning technique. You can read more about the process on the Mapbox blog.