A Deep Learning Method for Classifying Crop Infections with an Herbicides Suggestion System

The environment, farmers, consumers, and the global economy are all seriously threatened by crop infections. Pests and infections destroy 35% of field crops in India alone, causing significant financial losses for farmers. Careless use of herbicides poses a major health risk, as many are toxic and subject to biomagnification. Using these chemical fertilizers without identifying the infections first may lead to severe health issues in humans as well as other living creatures. Identifying infections and selecting the appropriate herbicide is a time-consuming process. To address this, we have developed an automated system for efficient diagnosis and herbicide recommendation.However, access to specialists is often restricted for farmers. Using a smartphone app, farmers can quickly and accurately identify infections and obtain remedies by taking pictures of the affected plant parts. Cloud-based image processing, powered by the latest Artificial Intelligence (AI) algorithms, enables real-time diagnosis. The AI model consistently enhances its precision by learning from user-uploaded photos and expert suggestions. The automated CNN model was employed to diagnose the test images, with plant pathologists ensuring the accuracy of results. The accuracy of infection identification was over 95%. For farmers and professionals interested in environmentally sustainable crop production, our approach is scalable, accessible, and innovative tool for managing infections across various agricultural crops. It has the potential to be implemented as a cloud-based service.