Scholarlink

<p>A&#xa0;digital twin-based environmental monitoring system for plant factories enables real-time detection and mitigation of environmental anomalies to maintain optimal growth conditions. This study proposes a reconfigurable digital twin framework that facilitates dynamic reconfiguration of digital models and algorithms when scenarios or monitoring strategies change. A key challenge in developing plant factory digital twins lies in the accurate and dynamic modelling of crop twins, particularly for densely growing crops. To address this, we introduce a 3D point cloud reconstruction method that integrates morphological growth models with imaging techniques, thereby improving reconstruction accuracy under occlusion. Furthermore, we develop an environmental anomaly prediction algorithm repository and a self-adaptive algorithm selection mechanism. A case study demonstrates the successful transformation from a mushroom plant factory system into a tea greenhouse monitoring system. Results show that our approach reused or adapted 61.1% of digital twin components, significantly reduced reconstruction workload, and achieved over 88% accuracy in predicting crop health anomalies. These results collectively validate the adaptability, modularity, and practical effectiveness of the proposed framework.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A reconfigurable digital twin framework for plant factories: crop modeling, environmental anomaly detection, and system reconfiguration

  • 期刊论文

  • 发表日期：

    2025年10月10日

查看全文
  • Hongxia Cai,
  • Jiaying Jiang,
  • Cheng Wang,
  • Shuzhen Yang,
  • Wanhe Du

摘要

A digital twin-based environmental monitoring system for plant factories enables real-time detection and mitigation of environmental anomalies to maintain optimal growth conditions. This study proposes a reconfigurable digital twin framework that facilitates dynamic reconfiguration of digital models and algorithms when scenarios or monitoring strategies change. A key challenge in developing plant factory digital twins lies in the accurate and dynamic modelling of crop twins, particularly for densely growing crops. To address this, we introduce a 3D point cloud reconstruction method that integrates morphological growth models with imaging techniques, thereby improving reconstruction accuracy under occlusion. Furthermore, we develop an environmental anomaly prediction algorithm repository and a self-adaptive algorithm selection mechanism. A case study demonstrates the successful transformation from a mushroom plant factory system into a tea greenhouse monitoring system. Results show that our approach reused or adapted 61.1% of digital twin components, significantly reduced reconstruction workload, and achieved over 88% accuracy in predicting crop health anomalies. These results collectively validate the adaptability, modularity, and practical effectiveness of the proposed framework.