Plant growth and agricultural output form the foundation of a country's sustainable development and directly affect livelihoods. Recent breakthroughs in frontier AI have opened new possibilities for scientific agriculture, especially regarding the impact of illumination on crop development. This paper highlights the importance and inherent complexity of plant shade simulation, as shading critically influences photosynthesis and yield.
To advance research and deliver societal benefits, we make two main contributions. First, we build a comprehensive plant growth and shade dataset covering four species—soybean, tomato, sugar beet, and strawberry. The dataset provides top‑down viewpoints and places an auxiliary light source on a circular trajectory, capturing dynamic shadows across multiple growth stages and diverse observation complexities. Second, we propose a generative shade simulation method based on diffusion models, enabling realistic shadow generation for unseen plants and supporting downstream robotic tasks such as perception, lighting control, and view planning. The model incorporates temporal conditioning, allowing flexible shade simulation at different growth stages.
Quantitative metrics and qualitative comparisons are used to evaluate the model, and results show superior shadow realism and temporal consistency compared to baselines. This work lays a foundation for plant‑aware shade modeling and has significant implications for agricultural robotics and intelligent farming systems.
Review