Integration of Convolutional Neural Networks and Recurrent Neural Networks for Foliar Disease Classification in Apple Trees

Automated methods intended for image classification have become increasingly popular in recent years, with applications the agriculture field including weed identification, fruit classification, and disease detection plants trees. In convolutional neural networks (CNN) already shown exceptional results but problem these models is that cannot extract some relevant features of input image. On other hand, recurrent network (RNN) can fully exploit relationship among features. this paper, performance combined CNN RNN evaluated by extracting on images diseased apple leaves. This article suggested a combination pre-trained LSTM, particular type RNN. With use transfer learning, deep were extracted from several connected layers i.e. Xception, VGG16, InceptionV3. The layer concatenated fed into to allow proposed model be more focused finding information data. Finally, class labels foliar are determined integrated experimental findings demonstrate approach outperforms individual models.