Implementing Convolutional Neural Networks and AdaBoost for Varied Turmeric Discrimination in India

This study looks at how well a Convolutional Neural Network (CNN) combined with AdaBoost can classify three main types of turmeric—Rajendra Sonia, Prabha, and Suvarna—in India. Using a carefully picked set of 11,000 high-quality turmeric root pictures, the suggested model did well. It had an excellent overall success rate of 98.2%. The model's success shows how good it is at telling small differences among turmeric types found in different farming areas. Table I shows measurements for performance evaluation. These include precision, recall, and F1-score which all work together well to show the model is good at reducing errors while identifying facts in the best way possible. The confusion matrix gives a big picture of how well the model does in classifying things. It shows correct and wrong guesses for each turmeric type group. This study helps to improve precision farming by offering a reliable system for the automatic classification of turmeric varieties. The shown correctness and good results support how important this model can be for improving farming ways, helping farmers make choices, and making sure the turmeric they sell is of nice quality. This research uses advanced computer methods to make it easier to classify different types of turmeric. It also shows a great example of using modern technology in farming, which could lead to better ways of sorting crops and higher farm output.