Grape and Apple Plant Diseases Detection Using Enhance DenseNet121 Based Convolutional Neural Network

Plant diseases have a severe influence on food production safety. Farmers invest a lot of time and resources in disease control, and they identify problems with their poor naked-eye method, which results in unhealthy farming. Therefore, an appropriate system is required to detect plant disease in an early age. Recent advances in deep learning have improved the performance in recognizing plant leaf diseases. In this paper, we present an improved—densnet121 model for identifying grape and apple leaf diseases. Instead of building a model from scratch, in this work we have employed a transfer learning technique based on two stages. The dataset in our experiments includes 18,754 images containing 8 different apple and grape diseased leaf image classes. In addition, multiple Transfer Learning architectures such as VGG16, inceptionV3, and Xception were evaluated and compared with our proposed model. The proposed model achieved the highest training overall accuracy and loss of 99.87% and 0.005%.