An Intelligent System for Plant Disease Diagnosis and Analysis Based on Deep Learning and Augmented Reality

Pathogens existing in crop fields cause harm to the plant and result in depletion of the production. These disease-causing microorganisms causes various diseases in the plant species and can be diminished with the use of microbial pesticides. The pesticides depend on the type of disease that is caused to the plant. This can be recognised and used based on the pattern of the disease caused, with the help of a deep learning model using argument reality. The proposed system uses various deep learning techniques including convolutional neural network (CNN), k-nearest neighbour (KNN), and InceptionV3. The accuracy of the model developed resulted in 96% for KNN, 96% for CNN, 98% for InceptionV3. Thus, the developed system relies on various algorithms, and compares the efficiency of each through the obtained results. InceptionV3 is observed as the efficient model among the three deep learning algorithms deployed, as the accuracy obtained through InceptionV3 is high compared to other models deployed for the classification process. To make the model interactive and more effective, the output of the model is displayed through augmented reality. This is achieved by importing the model in Augmented Reality unity engine and a 3D model employed for the same.