IoT-Integrated Graph Convolutional Networks and LSTM Models for Smart Plant Growth Monitoring in Controlled Environment Agriculture

The agricultural sector experiences a decrease in crop yield due to abrupt changes in weather, labour constraints, and soil aridity. Local weather patterns are already becoming more unpredictable due to global warming, which only makes things worse. These problems necessitate SPGM employing cutting-edge technology. Critical to the IoT, M2M connections allow for real-time decision-making through the integration of IoT with human-machine communication, teamwork, and technological analytics. Model training, feature extraction, and preprocessing are all necessary for this investigation. They execute correlation-based CCM feature extraction after preprocessing to guarantee data quality. The last step is to use data from SPGM or nearby district metering stations to train a one-of-a-kind GCRN model that can predict plant demand. With an MAE of 0.052, an MSE of 0.085, and $\mathbf{R}^{\mathbf{3}}$ value of 0.716, the suggested GCRN model surpasses its competitors. These results show that SPGM can increase agricultural yields, better manage resources, and make more accurate predictions.