A Novel Approach for Agricultural Crop Classification with Incremental Learning

Agriculture is a vital element that contributes significantly to feeding the world's growing population. Agricultural crop classification based on various parameters can assist farmers in forecasting which crop to yield in a given season. The proposed machine learning models in the literature make the classification with static data and do not handle changes in data occurring at a later time. To overcome this limitation, the authors in this research have developed an innovative approach with an incremental learning-based model which has the ability to remember the previously learned knowledge while handling newly arriving data. Authors have considered more than dozens of crop categories for the classification of crops with five machine learning classifiers and four ensemble learning classifiers to develop models with incremental learning. In addition, to prove the robustness of the proposed model, the authors have compared the newly developed models’ performance with state-of-the-art methods. The proposed approach showed the best performance with the Logistic Regression classifier.