AI Base-Mapping of Lusaka Speeds Up Zambia's National Land Titling Programme in Informal Settlements

Zambia's National Land Titling Programme had issued only around 142,000 titled parcels nationwide, leaving an estimated 80% of land untitled, a gap sharpest in Lusaka, where about 70% of residents live in informal settlements with no reliable base map since the 1980s.

In 2021, UK Aid-funded Ordnance Survey (OS) partnered with Zambia's land ministries, the Zambia Survey Department, the International Growth Centre and the Commonwealth Association of Architects to rebuild a digital base map of a 420 km² area covering Lusaka.

OS trained a machine-learning image-classification system to identify and vectorise roofed structures, natural and sealed surfaces, roads, grass, trees and water from aerial imagery.

OS reports the map was produced in a tenth of the time that previous manual techniques would have taken, though no independent accuracy figure for the classification itself has been published, and the claim is a production-time comparison stated by the implementing agency.

Zambia's Ministry of Lands has used the base map for land audits, titling in informal settlements, and housing, climate-adaptation and utility planning; OS and IGC describe the technique as replicable elsewhere, though no other country has yet rolled it out at the same scale.