Detection Of Grape Leaf Disease Using Transfer Learning Methods: VGG16 & VGG19

Abstract

Grapes are the most consumed fruit all around the world. For the healthy development of the grapefruit industry, it is calumniatory to control the escalation of grape leaf contamination in the crop. Due to the lack of knowledge in rural areas farmers fail to detect these diseases at the early stages, which results in feeble harvest quality. The proposed system uses the transfer learning VGG models which classify the three most common grape leaf diseases along with the healthy grape leaves. These models achieve a mean accuracy of 98% on the testing data, which indices the feasibility of the neural network approach compared to the manual detection of these diseases.