摘要
It has been established that precision agriculture has evolved so quickly that blockchain has been embraced as a disruptive technology. Smart farms at first focused on the application of blockchain to increase operational effectiveness, but the shift towards an ‘Internet of Smart Farms’ (IoSF) is prepared for the improvement of crop yield optimization. However, precision agriculture has several challenges related to the secure sharing of data, data utilization in terms of efficiency and security, and record management in terms of data integrity especially when IoT systems are integrated. To overcome these difficulties, the present work presents a broad methodological framework that guarantees the safe, fast, and transparent processing of data in precision agriculture. The proposed framework comprises multiple layers—there are four layers as follows: data layer, artificial intelligence (AI) layer, security layer, and blockchain layer. There are measurements which are obtained from IoT sensors which include temperature, soil moisture, humidity, crop health and weather conditions. Outlier removal, normalization, feature selection, and extraction are performed to improve data quality, and the feature selections are chosen by using the binary slime mould algorithm (SMA). For prediction and analytical tasks, bidirectional long short-term memory (Bi-LSTM), and gated recurrent unit (GRU) deep learning are used for classification, anomaly detection and yield prediction. The implementation of blockchain is to guarantee the decentralization of the records, making the transactions safe and unchangeable in the system. For yield prediction tasks, the Bi-LSTM and GRU models yield an accuracy of 95.8% and 94.6%, respectively, and for F1, it was 0.96 for Bi-LSTM and 0.94 for GRU. Anomaly detection achieves a precision of 0.93 and recall of 0.92, significantly outperforming conventional machine learning models. The blockchain layer ensures 100% data integrity and reduces the risk of data tampering by 97% compared to traditional centralized systems.