Survey on Smart Multi Crop Disease Detection and Advisory Systems Using IoT and Deep Learning

Abstract

Food production/production serves as the backbone of the economies of some countries. Unfortunately, disease in crops affects growth/production and reduces yield and quality of crop. Most currently, disease detection in crops is performed manually by visual examination, which is both time-consuming and often inaccurate. The use of Deep Learning (DL) and IoT technologies have significantly enhanced the ability to monitor diseases and manage crops in modern agriculture over recent years. This survey discusses various techniques used for multi crop disease detection and farmer advisory system mainly for Tomato, Capsicum annuum and Banana crops. The study talks about various DL models such as CNN, VGG16, MobileNet, RESNET and Transformer based approaches which are being used for classification of plant diseases. Most of the previous studies focused on disease detection using lab datasets and the recent studies aimed at improving the model performance under practical field conditions with different light, weather changes, shadows and complex backgrounds. The survey also highlights a big disadvantage in a lot of existing systems. Most existing models are capable of disease classification but they are not giving appropriate instructions to farmers regarding irrigation, pesticide application and crop maintenance. To solve this problem, the proposed research workr introduces a smart multi-crop observing and advisory framework by combining IoT-based environment for sensing and image-level disease detection. The parameters such as soil moisture, temperature and humidity can be useful in early detection of potential disease conditions. Additionally, the proposed framework also offers farmer advisory features such as irrigation recommendations, weed monitoring and mobile-based notifications. Local processing and edge AI solutions can also help to reduce the dependence on the internet in rural farming locations. Overall, the survey suggests that a combination of disease detection, environmental monitoring and farmer support systems can improve crop protection and promote sustainable precision agriculture.