AI system accurately maps urban green spaces

Described in the ACM Journal on Computing and Sustainable Societies, the system used a technique known as 'green augmentation’ to better detect different types of vegetation from Google Earth satellite imagery. This involved enhancing their training data to include varied versions of green vegetation under different lighting and seasonal conditions.

According to the NYU team, tracking urban greenery from space is difficult, but increasingly vital for city planning in the face of climate change. Traditional satellite analysis has been found to miss up to 37 per cent of vegetation. By contrast, the new AI system achieved 89.4 per cent accuracy with 90.6 reliability.

"Previous methods relied on simple light wavelength measurements," said Chunara, director of the NYU Centre for Health Data Science and a member of NYU Tandon’s Visualisation Imaging and Data Analysis Centre (VIDA). 

"Our system learns to recognise more subtle patterns that distinguish trees from grass, even in challenging urban environments. This type of data is necessary for urban planners to identify neighbourhoods that lack vegetation so they can develop new green spaces that will deliver the most benefits possible. Without accurate mapping, cities cannot address disparities effectively."

Testing the new system across Karachi, Pakistan’s most populous city, it was found that it averaged just 4.17 square metres of green space per person - less than half the World Health Organisation's (WHO’s) recommended minimum of nine square metres per capita. The disparity within neighbourhoods was significant: while some outlying union councils have over 80 square metres per person, five union councils have less than 0.1 square metres per capita.

The study revealed that areas with more paved roads – typically a marker of economic development – tend to have more trees and grass. More significantly, in eight different union councils studied, areas with more vegetation showed markedly lower surface temperatures, highlighting the cooling role that green spaces can play in urban environments.

While the NYU team have made their methodology public, applying it to other cities would require retraining the system using local satellite imagery.