Deep Learning Models Used to Recognize Plant Disease Detection

The quality and production of crops depend on the timely and accurate detection of crop diseases so that the right treatments may be implemented. Nevertheless, expertise and extensive experience in plant pathology are necessary for disease diagnosis. As a result, the development of an early detection system for diseases is a crucial area for agricultural automation. Building a convolutional neural network (CNN) algorithm using five pre-trained models and photos of sick and healthy plants in pairings was necessary to produce this system. Classification of crops, disease detection, and illness classification are the three stages that make up the disease detection model. To make the model applicable to a broad range of situations, the "unknown" is grouped into categories. With a precision of 98.08%, the disease identification model successfully identified different kinds of crops and diseases in the validation test. Incorporating non-model crops into the training dataset increased the poor accuracy of the model, suggesting that the model is expendable. As we expand our training dataset to include other crop types, our model will see extensive application and may find a home in smart farming for Solanaceae crops.