Osaka, Japan – Researchers at the University of Osaka have developed an innovative method using artificial intelligence (AI) and street view imagery to analyze and improve the visual appeal of urban vegetation. Published in Landscape Ecology, the study introduces a technique that captures detailed seasonal changes in plant species, helping planners create more vibrant and restorative green spaces.
A New Approach to Urban Greenery
The team combined deep learning and 3D reconstruction technology to create the Seasonal Species-Specific Plant View Index, which identifies 51 urban plant species with an average accuracy of 82.17%. This method overcomes limitations of traditional green view analysis by standardizing perspectives and eliminating distortions in street view images. Notably, it highlights seasonal features like cherry blossoms in spring and maple leaves in autumn, offering planners precise tools to enhance biodiversity and visual interest.
Boosting Well-Being and Sustainability
“Diversity in plant color and species enhances the ‘feel-good’ factor of urban green spaces,” said lead author Anqi Hu. The technology, tested in Suita City, Osaka, enables virtual park design and supports the revitalization of brownfield sites. By incorporating plants with varied colors, shapes, and growth patterns, cities can extend ecological, economic, and psychological benefits year-round.
Future Applications
Senior author Tomohiro Fukuda emphasized the method’s potential for 4D urban design, allowing planners to evaluate and optimize green spaces across seasons. The framework promises to transform urban planning by making greenery more dynamic and impactful for residents and visitors alike.
Add comment