Grapevine Leaf Disease Identification Using Transfer Learning

Abstract

The grapevine (Vitis vinifera L.) is an important fruit crop with global commercial significance. Grapevine diseases such as black rot, black measles, and leaf blight are commonly found. The importance of a fast and correct diagnosis in limiting the spread of disease and decreasing production losses cannot be overstated. The recent utilization of pre-trained deep learning models has opened doors for new diagnostic algorithms in the domain of plant disease identification. In this paper, we propose a disease identification method using a pre-trained deep learning model with a new dense classifier. The model was trained using 3423 grapevine leaf images from PlantVillage database that belong to four classes, namely, Healthy, Black Measles, Black Rot, and Leaf Blight. The proposed model achieved 98.53% classification accuracy on the validation dataset with a 70-30 train-test ratio.