# AI driven shadow model detection in agropv farms

**Authors:** Sai Paavan Kumar Dornadula, Pascal Brunet, Dr. Susan Elias  
**arXiv:** [2304.07853](https://arxiv.org/abs/2304.07853)  
**Date:** 2023-04-16  
**Summary page:** https://www.summarizepaper.com/en/paper/2304.07853v1/

## Key points

- Agro-photovoltaic (APV) combines agriculture and solar photovoltaic projects within the same area.
- The APV market is expected to experience significant growth in the next few years, with a projected investment of $9 billion in 2030.
- Shadows impact plant growth, microclimate, and evapotranspiration in APV farms.
- A recent study used state-of-the-art CNN and GAN-based neural networks to detect shadows accurately in agro-PV farms.
- Partial shadowing from moving objects can be difficult to identify using current technology, and real-time monitoring of shadow patterns is necessary for effective management of APV farms.
- Future research should focus on developing more sophisticated neural network-based shadow detection algorithms that can account for partial shadowing from moving objects and integrate them with control systems for APV farms.
- AI-driven shadow model detection has tremendous potential to improve the efficiency and sustainability of APV farming practices while contributing to global efforts towards renewable energy production.

## Summary

Agro-photovoltaic (APV) is an innovative farming practice that combines agriculture and solar photovoltaic projects within the same area. This emerging market is expected to experience significant growth in the next few years, with a projected investment of $9 billion in 2030. However, identifying shadows is crucial to understanding the APV environment, as they impact plant growth, microclimate, and evapotranspiration. To address this issue, a recent study by Sai Paavan Kumar Dornadula, Pascal Brunet, and Dr. Susan Elias used state-of-the-art CNN and GAN-based neural networks to detect shadows in agro-PV farms. The results demonstrated the effectiveness of these technologies in detecting shadows accurately. However, challenges remain in shadow detection for APV farms. Partial shadowing from moving objects can be difficult to identify using current technology. Additionally, real-time monitoring of shadow patterns is necessary for effective management of APV farms. Future research should focus on developing more sophisticated neural network-based shadow detection algorithms that can account for partial shadowing from moving objects and integrate them with control systems for APV farms. Such advancements would increase productivity and profitability while supporting the environment, soil health, and farmers' livelihoods. Overall, AI-driven shadow model detection has tremendous potential to improve the efficiency and sustainability of APV farming practices while contributing to global efforts towards renewable energy production.