Agri Doctor: Intelligent Crop Disease Detection for Sustainable Farming

Prior years’ indices reveal that deep learning is one of the most efficient methodologies to study diseases in plants. An identification of the diseases of the leaves of several crops is made in this research using MobileNet which is a simpler CNN. The compute and speed of MobileNet has been boosted hence it can be implemented online or in most scenarios that demands real time implementation ignoring the powering concerns. In the image identification tasks, there is high accuracy due to the reduction of the number of parameters and the required computations using depthwise separable convolutions whereby the model is relieved of numerous parameters. Thus, the automated framework, MobileNet, is successfully able to detect specific features associated with different diseases, which infect plants during the training phase using a large set of leaves’ images and, consequently, classify plants as either healthy or diseased plant. The identification of diseases is very relevant to the farmers as it helps in prevention of diseases that may cause severe infection of the produce. In addition to increasing food production, it also promotes organic farming that leads to reduction in use of pesticide. Consequently, the system proposed in this paper using the MobileNet satisfies the early detection of diseases in agricultural systems and in addition, foods security through improvement of sustainable farming globally.