Deep Umbra: A Generative Approach for Sunlight Access Computation in Urban Spaces

Kazi Shahrukh Omar,Gustavo Moreira,Daniel Hodczak,Maryam Hosseini,Nicola Colaninno,Marcos Lage,Fabio Miranda

2024-02-27

Abstract:Sunlight and shadow play critical roles in how urban spaces are utilized, thrive, and grow. While access to sunlight is essential to the success of urban environments, shadows can provide shaded places to stay during the hot seasons, mitigate heat island effect, and increase pedestrian comfort levels. Properly quantifying sunlight access and shadows in large urban environments is key in tackling some of the important challenges facing cities today. In this paper, we propose Deep Umbra, a novel computational framework that enables the quantification of sunlight access and shadows at a global scale. Our framework is based on a conditional generative adversarial network that considers the physical form of cities to compute high-resolution spatial information of accumulated sunlight access for the different seasons of the year. We use data from seven different cities to train our model, and show, through an extensive set of experiments, its low overall RMSE (below 0.1) as well as its extensibility to cities that were not part of the training set. Additionally, we contribute a set of case studies and a comprehensive dataset with sunlight access information for more than 100 cities across six continents of the world. Deep Umbra is available at

Computer Vision and Pattern Recognition,Computers and Society

What problem does this paper attempt to address?

The paper primarily addresses the problem of calculating sunlight exposure and shadow coverage in urban environments, proposing a new computational framework—Deep Umbra, which is used to efficiently quantify sunlight availability and shadow distribution in urban spaces across different seasons on a global scale.

Specifically, Deep Umbra addresses the problem through the following ways:

1. **Background and Need**: With the acceleration of global urbanization, the distribution of sunlight exposure and shadows in urban spaces has a significant impact on residents' quality of life. Therefore, an effective method is needed to analyze and plan the sunlight availability in cities.

2. **Methodology**: Deep Umbra is based on Conditional Generative Adversarial Network (cGAN), utilizing the physical morphology information of cities (such as building heights) as well as geographical location (latitude) and seasonal factors to calculate high-resolution spatial sunlight exposure cumulative data.

3. **Dataset**: Researchers collected training data from 7 different cities and validated the model's performance in cities not involved in the training. Additionally, they contributed a global dataset containing sunlight exposure information for over 100 cities.

4. **Experimental Evaluation**: A series of experiments validated the model's effectiveness, including the model's generalization ability in unseen cities and its accuracy compared to other methods. The experimental results show that Deep Umbra has a lower overall Root Mean Square Error (RMSE) and can quickly process large-scale urban data.

5. **Case Study**: Practical cases demonstrated the application value of Deep Umbra in evaluating sunlight exposure in public spaces such as street networks and parks.

In summary, Deep Umbra aims to improve the efficiency and accuracy of sunlight exposure and shadow analysis through machine learning techniques to support sustainable urban planning and development.