Modelling of energy demand prediction system in potato farming using deep learning method

Agriculture and energy are intricately connected, with agriculture being a significant energy consumer and supplier. In this comprehensive study, SPSS and Jupyter Notebook were used to model and predict the energy requirements of potato plants during cultivation. A system using deep learning methods, specifically the Convolutional Neural Network (CNN), was also developed to accurately predict the classification of potato plant growth phases using image data. The CNN model, developed with 100 epochs and 5 layers, used 1 125 image data of potato plants, categorising them into two classes: the vegetative phase, with an energy requirement of 4 195.80 MJ.ha-1, and the generative phase, with an energy requirement of 746.45 MJ.ha-1. The model's accuracy in reflecting the actual data, with a mean absolute error of 0.11, mean square error of 0.01, and root mean square of 0.13, indicates no significant issues. The test predicted categorization with 99% precision, underscoring the thoroughness and validity of this study and reassuring the audience about the accuracy of the results. The study findings not only validate the use of deep learning in agriculture but also inspire the development of applications to predict the energy demand for each growth phase using plant image data.