Machine Learning-based Data Fusion Technique for Irrigation Management in Multi-Crop Framework

The development of smart agriculture depends on the irrigation system which is the main factor in crop growth. Therefore vast research is taking place in the smart irrigation system of smart agriculture that involves IoT, machine learning methods, hybrid techniques, etc. To obtain the maximum yield from the soil, a multi-cropping system is introduced that grows various crops in the same agriculture field that are selected based on the soil fertility and crop genome. Many sensors are deployed to enhance the irrigation system which increases the amount of data that needs to be communicated. Data fusion is a technique that reduces the dimension of the data and combines all modes of data, is an efficient method to resolve this issue. Therefore, this paper proposes the data fusion technique based XAI approach and ensemble transfer learning technique for decision-making process to manage the irrigation in the smart irrigation system. The LIME-based XAI model is used to increase the fusion level. Ensemble transfer learning increases the prediction accuracy of the system with the previous iteration dataset with the ensemble groups of training and testing.