For an efficient crop yield, early identification of diseases is necessary in agriculture. Plant disease is an ongoing problem for smallholder farmers, which threatens income and food security. It is necessary to have rapid identification and recognition of crop diseases. In Image recognition, Convolutional Neural Networks have been in recent trends and provide the potential to have a simple and definitive diagnosis. A new architecture to successfully classify different plant diseases is proposed in this article. Three types of plants make up the dataset with healthy and disease affected leaves collected from various resources. The new CNN model is trained and tested and has provided an accuracy of 95%. The CNN model's training speed is improved by 83 % when it is run on an ARM processor. These experimental findings indicate that three distinct forms of plant disease can be correctly identified by the proposed method.
CNN Based Early Identification of Plant Disease Using Leaf Images
2021-12-02
2389475 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch