Top Highlights
- Innovative AI System: A research team from NYU, led by Rumi Chunara, developed an AI system that utilizes satellite imagery to track urban green spaces, achieving 89.4% accuracy—far surpassing traditional methods known for missing 37% of urban vegetation.
- Significant Urban Disparities: The analysis of Karachi illustrates stark inequalities in green space distribution, revealing an average of just 4.17 square meters of green space per person, well below the WHO’s recommended 9 square meters.
- Impact on Urban Health: The study demonstrates that neighborhoods with more vegetation experience lower surface temperatures, highlighting the cooling and public health benefits of urban green spaces, which are disproportionately absent in low-income areas.
- Call to Action for Urban Planning: Accurate mapping provided by this AI approach can assist urban planners in identifying neighborhoods needing green spaces, crucial for addressing climate change and health disparities in rapidly urbanizing areas.
New AI System Maps Urban Green Spaces, Revealing Environmental Divides
A research team led by Rumi Chunara, an NYU associate professor, has unveiled an innovative artificial intelligence (AI) system that uses satellite imagery to track urban green spaces. This new system outperforms previous methods and plays a vital role in creating healthier cities.
The researchers tested their approach in Karachi, Pakistan’s largest city. Karachi’s combination of dense urban areas and diverse vegetation made it the ideal test case. The team’s analysis, accepted for publication by the ACM Journal on Computing and Sustainable Societies, revealed a stark environmental divide. Some neighborhoods boast tree-lined streets, while others have very little vegetation.
Traditional satellite methods often miss up to 37% of urban greenery. As cities confront climate change and rapid urbanization, especially in Asia and Africa, accurate mapping of green spaces becomes increasingly important. Green spaces help reduce urban heat, filter air pollution, and provide essential areas for exercise and mental well-being.
However, the distribution of these benefits is often unequal. Low-income neighborhoods frequently lack vegetation, resulting in hotter and more polluted environments than their wealthier counterparts.
To develop their solution, the research team enhanced AI segmentation architectures like DeepLabV3+. They used high-resolution satellite imagery from Google Earth to train the system. By applying a technique called “green augmentation,” they improved the accuracy of vegetation detection by 13.4% compared to existing methods. The system achieved an impressive 89.4% accuracy with 90.6% reliability, in stark contrast to traditional methods, which only reached 63.3% accuracy with 64.0% reliability.
Chunara noted, “Our system learns to recognize subtle patterns that differentiate trees from grass, even in challenging urban environments. This data is essential for urban planners to identify areas lacking vegetation and develop new green spaces.”
Need of more Urban Green Spaces
The analysis of Karachi revealed the city averages just 4.17 square meters of green space per person, well below the World Health Organization’s recommended minimum of 9 square meters. The disparity between neighborhoods is striking; while some local councils have over 80 square meters per person, others feature less than 0.1 square meters.
Interestingly, areas with more paved roads—often indicative of economic development—tend to have more greenery. The study also found that neighborhoods with abundant vegetation exhibited lower surface temperatures, highlighting the importance of green spaces in urban cooling.
Singapore serves as an inspiring example, providing 9.9 square meters of green space per person despite its dense population. This stark contrast demonstrates the possibilities for cities that prioritize green space planning.
The researchers plan to make their methodology publicly available. However, applying it in other cities will require retraining the system using local satellite imagery.
Chunara’s work emphasizes the connection between technology, Urban Green Spaces and public health. Previous studies have utilized social media data to map systemic issues, such as racism and health disparities during the COVID-19 pandemic.
This study, funded by the National Science Foundation and National Institutes of Health, includes collaborators from The Aga Khan University, showcasing a strong interdisciplinary approach to tackling urban challenges.
Expand Your Tech Knowledge
Dive deeper into the world of Cryptocurrency and its impact on global finance.
Discover archived knowledge and digital history on the Internet Archive.
SciV1
