Leaf Based Grape Variety Identification Using Computer Vision

Machine Learning has become more prominent over the years due to advancements in Artificial Intelligence. This paper evaluates the effectiveness of a convolutional neural network (CNN) for image classification on two grape leaf varieties from Anatolia, Turkey, using TensorFlow and Keras in Python. A dataset consisting of 100 images per leaf type was used to train CNN through 15 epochs. The model achieved an accuracy of 95 % and a precision of 93 %. These results demonstrate the potential for enhancing agricultural practices by assisting in the identification of grape leaf variants and effectively distinguishing between imported and local grape types.