Research on Multi environmental Parameter Prediction Model of Solar Greenhouse Based on the Whole Growth Period of Tomato

Based on the potential relationships and temporal characteristics among four types of environmental parameters in the tomato production environment of solar greenhouses, three multi-environment parameter prediction models were constructed using the environmental dataset of the whole tomato growth period as model inputs: (1)The 3DCNN-CLSTM model was constructed by integrating the 3D Convolutional Neural Network (3DCNN) and Convolutional Long Short-Term Memory (CLSTM) network; (2)The PSO-3DCNN-CLSTM model was constructed by introducing the Particle Swarm Optimization (PSO) algorithm to optimize the parameters of the fully connected layer and the learning rate of the 3DCNN-CLSTM model; (3)The PSO-Attention-3DCNN-CLSTM model was constructed by incorporating an Attention mechanism layer into the PSO-3DCNN-CLSTM model, enabling the model to focus on key features. Through analysis and comparison of the prediction performances of the three models, the results show that the PSO-Attention-3DCNN-CLSTM model has the highest prediction accuracy.