A novel deep learning model to predict the soil nutrient levels (N, P, and K) in cabbage cultivation

Abstract

Cabbage (Brassica oleracea) is a green cruciferous vegetable. Major nutrients (nitrogen, phosphorus, and potassium) are frequently applied to the soil due to low fertility levels. However, optimizing required fertilizer levels are extremely important to avoid any overuse and underuse. Therefore, it is important to develop a comprehensive methodology for evaluating the major nutrients in the soil. In this research, a deep learning model was introduced to predict the nitrogen, phosphorus, and potassium content of the soil by analyzing the growing characteristics of the plants, such as plant height, the number of leaves, and the average leaf area of the plant. To achieve this, the growing characteristics of the cabbage plants were recorded weekly along with the respective soil nitrogen, phosphorus, and potassium content of the nearby soil. After the data was trained using the Levenberg–Marquardt algorithm and tested with different transfer functions such as logarithmic sigmoid, pure linear, and tangent sigmoid, better predictions were obtained through the model. According to the Pearson correlation values, pure linear and tangent sigmoid showed higher values, ranging from 0.99 for training, testing, validation, and all data points from the model, indicating a strong relationship between the actual and predicted values. According to the Mean Square Error values, the tangent sigmoid transfer function outperformed the others, giving a value of 1.0813, indicating better predictions of the soil nitrogen, phosphorus, and potassium content from the model.