Authors: Deepika Soni, Neelesh Shrivtastav, Pradeep Tripathi
Abstract: Using imaging technology, we suggest that plant disease detection systems automatically identify the symptoms that occur on the leaves and stems of a plant, allowing for the cultivation of healthy plants on a farm to be improved. It is these systems that monitor the plant's characteristics, such as its leaves or stem, and any variations that are seen from those characteristics will be automatically recognised and sent to the user. The purpose of this paper is to conduct an evaluation of the available disease detection methods in plants. The most recent breakthrough in deep learning-based convolutional neural networks (CNNs) has resulted in a significant improvement in picture categorization accuracy. This Thesis, which is motivated by CNN's success in picture classification, uses a pre-trained deep learning-based technique for identifying plant illnesses to detect plant diseases. The contribution of this work may be divided into two categories: Predictions for a dataset may be made using the most powerful large-scale architectures available today, such as AlexNet GoogleNet, which are utilised for illness detection and the usage of baseline and transfer learning techniques for predictions. CNN's suggested model was trained and tested using data sets gathered from the website, according to the network. The results of training, testing, and experiments demonstrate that the suggested architecture is capable of realising and increasing GoogleNet model getting to 99.10 percent. when compared to other models, the accuracy.
