Soil Wetness Classification in Agriculture Using Machine Learning Models
S M Abdullah Al Shuaeb
Department of Computer Science and Technology, Tangail Polytechnic Institute, Tangail, Directorate of Technical Education, Bangladesh.
Md. Rakib Hassan *
Department Computer Science and Mathematics, Faculty Agricultural Engineering and Technology, Bangladesh Agricultural University, Bangladesh.
Machbah Uddin
Department Computer Science and Mathematics, Faculty Agricultural Engineering and Technology, Bangladesh Agricultural University, Bangladesh.
*Author to whom correspondence should be addressed.
Abstract
Soil wetness is the most important factor for a plant to survive. If the soil is completely dry for a long time, the plants will perish. Many plants will also die if the soil is submerged in the water for a long period of time. Without water, plants will not be able to take nutrients from the soil. Besides, different plants have different soil wetness or moisture requirements. To ensure proper plant growth, soil wetness levels should be monitored and maintained continuously. But most of the time, it is not possible to continuously monitor the water level manually. Therefore, in this work, we have proposed an image-based soil wetness classifier using different machine learning algorithms. We have classified the soil in six different wetness levels and have used five machine learning algorithms for classifying the soil wetness levels as an artificial neural network, convolutional neural network, decision tree, k-nearest neighbor, and support vector machine. We have compared these algorithms and found that the convolutional neural network achieves the highest accuracy which is 97.7%. Our proposed method can be used by the stakeholders to increase crop production by ensuring proper soil water levels for continuous plant growth.
Keywords: ANN (Artificial Neural Network), KNN (K-Nearest Neighbor), DT (Decision Tree), SVM (Support Vector Machine), CNN (Convolutional Neural Network), CM (Confusion Matrix), Soil Wetness (SW), Machine Learning (ML)