With about 8000 instances on the planet, the airport is one of the few architectural typologies which we can exhaustively catalog, compare, quantify, and map.

The future of the Aerosphere - the ecology of the planet’s air transit spaces - is not only about a handful of leading global city nodes but also the thousands of remote airstrips or intermediate terminals which make the entire fibrous network of air travel. We propose using machine vision, including comprehensive scanning of satellite photography, to more completely understand, compellingly map, and perhaps learn from the deep structures, family resemblances, and typological invariants of the complete family of every airport ever built: in short, a new way to automatically classify building type.

Here we present the results of such research, considering both the morphological characteristics of the terminals themselves and the similarities among the landscapes in which the airstrips are embedded and which they in turn mark.