Plant Disease Detection using Deep Learning

Abstract

Abstract: Farming is critical to the nation's economy and progress. Precision Farming (PA), on the other hand, is still in its development when it comes to technology-driven growth. Various plant diseases have caused pain to untold millions of people around the world over the years, with an estimated annual yield loss of 14% globally. Computerized disease segmentation and diagnosis from based on leaf photos has the potential to be more effective than the current method. Image capture, preprocessing, and segmentation are followed by augmentation, feature extraction, and classification using models for automatic plant disease diagnosis. This project employs VGG-16, ResNet-50, AlexNet, DenseNet-169, and InceptionV3 Deep Learning models to identify plant illnesses from photos in the Plant Village Dataset and reliably classify them into two classes. The results of the experiment revealed that the ResNet-50 has achieved highest accuracy of 97.80 % as compare to other applied deep learning models for disease classification. Keywords: VGG16, ResNet50, Inception V3, CNN, GoogleNet, AlexNet