DeepShade: Enable Shade Simulation by Text-conditioned Image Generation

2025-07-16

AI Summary

Existing routing systems struggle to accurately estimate shadows from satellite imagery and suffer from a lack of high-quality, shadow-annotated training data. Method: We propose DeepShade—a text-conditioned, edge-aware diffusion model—that introduces (i) the first multi-region, multi-temporal satellite image–shadow paired dataset with solar-angle alignment; (ii) joint multimodal modeling of RGB and Canny edge maps via contrastive learning; and (iii) high-fidelity temporal shadow distribution generation using Blender 3D simulation. Contribution/Results: DeepShade is the first framework to embed fine-grained dynamic shadow generation directly into path planning, significantly improving shade ratio estimation accuracy. Evaluated on real-world satellite imagery from Tempe, Arizona, it enables high-resolution thermal comfort-aware route planning. The method provides a deployable technical solution for smart cities addressing extreme heat resilience.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Computer Vision: Diffusion Models for Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Large language models for search
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Heatwaves pose a significant threat to public health, especially as global warming intensifies. However, current routing systems (e.g., online maps) fail to incorporate shade information due to the difficulty of estimating shades directly from noisy satellite imagery and the limited availability of training data for generative models. In this paper, we address these challenges through two main contributions. First, we build an extensive dataset covering diverse longitude-latitude regions, varying levels of building density, and different urban layouts. Leveraging Blender-based 3D simulations alongside building outlines, we capture building shadows under various solar zenith angles throughout the year and at different times of day. These simulated shadows are aligned with satellite images, providing a rich resource for learning shade patterns. Second, we propose the DeepShade, a diffusion-based model designed to learn and synthesize shade variations over time. It emphasizes the nuance of edge features by jointly considering RGB with the Canny edge layer, and incorporates contrastive learning to capture the temporal change rules of shade. Then, by conditioning on textual descriptions of known conditions (e.g., time of day, solar angles), our framework provides improved performance in generating shade images. We demonstrate the utility of our approach by using our shade predictions to calculate shade ratios for real-world route planning in Tempe, Arizona. We believe this work will benefit society by providing a reference for urban planning in extreme heat weather and its potential practical applications in the environment.

Problem

• Estimating shade from noisy satellite imagery is challenging
• Limited training data hinders generative shade simulation models
• Current routing systems lack shade information for heatwave planning

Innovation

• Blender-based 3D simulations for shadow dataset
• Diffusion-based model with RGB-edge joint learning
• Text-conditioned shade generation for route planning