In order to overcome the notable limitations of current methods for monitoring grain storage states, particularly in the early warning of potential risks and the analysis of the spatial distribution of grain temperatures within the granary, this study proposes a multi-model fusion approach based on a deep learning framework for grain storage state monitoring and risk alert. This approach combines two advanced three-dimensional deep learning models, a grain storage state classification model based on 3D DenseNet and a temperature field prediction model based on 3DCNN-LSTM. First, the grain storage state classification model based on 3D DenseNet efficiently extracts features from three-dimensional grain temperature data to achieve the accurate classification of storage states. Second, the temperature prediction model based on 3DCNN-LSTM incorporates historical grain temperature and absolute water potential data to precisely predict the dynamic changes in the granary’s temperature field. Finally, the grain temperature prediction results are input into the 3D DenseNet to provide early warnings for potential condensation and mildew risks within the grain pile. Comparative experiments with multiple baseline models show that the 3D DenseNet model achieves an accuracy of 97.38% in the grain storage state classification task, significantly outperforming other models. The 3DCNN-LSTM model shows high prediction accuracy in temperature forecasting, with MAE of 0.24 °C and RMSE of 0.28 °C. Furthermore, in potential risk alert experiments, the model effectively captures the temperature trend in the grain storage environment and provides early warnings, particularly for mildew and condensation risks, demonstrating the potential of this method for grain storage safety monitoring and risk alerting. This study provides a smart grain storage solution which contributes to ensuring food safety and enhancing the efficiency of grain storage management.

If paddy is not dried in time after harvesting, it very easily becomes moldy, which causes substantial losses. Inappropriate drying methods also affect the quality and safety of paddy when it enters the drying process. In order to select the optimal drying method, paddy with different moisture content was treated with sun drying, mechanical drying, and late harvesting, and the quality indexes of paddy were tested for safe moisture content. The effects of different drying methods on the quality of paddy were analyzed in terms of burst rate, taste value, appearance quality, and pasting characteristics. A paddy quality index system was constructed, and gray relational analysis (GRA) was used to evaluate comprehensively the paddy quality. The results showed that when the moisture content of paddy was 24.4%, the best quality of paddy was obtained after mechanical drying, when the gray relational degree was 0.996. Timely harvesting and mechanical drying can not only reduce the loss of paddy, but can also ensure the quality and safety of paddy.

Traditional post-harvest operation methods applied in rice fields lack advanced management knowledge and technology, which has led to post-harvest losses. We proposed the concept of Five Time (5T) management for the first time. 5T management divides the whole life cycle of rice into different growth time interval to complete process management. This paper mainly introduces the management of rice grain period, that is, the post-harvest management period, including the operation process management of harvesting, field stacking, drying, warehousing, and storing. In 2019, our research team formulated the 5T management method, which considers the entire post-harvest process, and carried out a pilot application of this method at the Jilin Rice Industry Alliance of Jilin Province. Moreover, to promote the 5T management method, our research team carried out follow-up experiments in rice production enterprises and found severe post-harvest rice losses. This paper combined a large number of literature and the basic theory research of rice post-harvest to analyze the traditional methods for post-harvest processing and the associated rice losses. By implementing the 5T management method, 4.33% of losses incurred during the T₁ harvesting period could be recovered. In the T₂ field period, drying rice within 48 h after harvesting could reduce losses by 2.5%. In the T₃ drying period, the loss rate could be reduced by 1.6% if traditional drying methods were replaced by mechanical drying and by 0.6% if cyclic drying was implemented to prevent over-drying. In the T₅ storage period, the loss rate of 7% could be reduced by adopting advanced grain storage technologies such as low-temperature storage. Overall, the rice loss rate could be reduced by 15.43%, which is equivalent to a yield of 32.68 million tonnes of rice. The important factors in each period are strictly controlled in the 5T management method to prevent the post-harvest losses caused by flawed concepts and improper management and to increase the amount of usable fertile land.

Against the background of increased population and resource depletion, managing food losses means conserving agricultural production resources and increasing farmer income. This paper mainly introduces the discovery and value of latent loss. In 2019, our experimental team formulated the 5T management method, which concerned the rice harvest period. Moreover, to promote the 5T management method, our team conducted relevant experiments about rice grain growth curve and found an accidental reduction in the dry matter weight of rice. To ensure the accuracy of the results of the latent rice loss in 2019, easy-to-use nuclear magnetic resonance is a non-destructive, rapid evaluation method, which is suitable for accurately determining high-moisture content and multi-variety rice. Overall, the experimental results in 2020 showed that if rice was harvested at the optimal time, the rice loss rate could be reduced by 3.5346%, which is equivalent to a yield of 235,051 tons of rice. The results are in agreement with evidence from field trials and suggest that the latent loss of dry matter caused by delayed harvest not only causes yield losses and economic losses but also increases postharvest grain loss and rice seed shattering loss. This significant factor, the optimal harvest time, in harvest period is strictly controlled to prevent the dry matter loss caused by innate knowledge and traditional management and to provide new possibilities for increasing the amount of available fertile land and generating income.

The development of smart farming comes with a lot of data problems. Studies have shown this is due to insufficient cognition of the structural relationship between data and events. Shili Theory is an attractive concept. To embed intelligent agricultural technology in events and the natural environment, especially to unify and standardize agricultural production data, firstly, this paper has defined the concept of Shili Theory which researches the natural regularity of the event by Shili Mirrored Structure. Secondly, this paper has proposed a Shili Mirrored Structure based on the technology development path (from the human brain memory mechanism to the information storage mechanism to intelligent technology). Finally, the structure has been applied to develop an intelligent system of agricultural production data management. In rice production of Jilin Province, it forms the event chain of the whole plant 5T (seed, seeding, paddy shoot, grain, product period operation) and grain period 5T (harvesting, field stacking, drying, warehousing, storing). The system application shows that this management structure can reduce data flow, improve data utilization, and enhance the correlation between data and events. It can realize the quality improvement of the agricultural production process, especially revealing the 8.83% significant latent loss in rice harvest.