Artificial Intelligence in Agriculture (Jan 2021)
Early stage detection of Downey and Powdery Mildew grape disease using atmospheric parameters through sensor nodes
Abstract
Grape diseases are major factors causing severe diminution in its fruit development. Unfavorable climatic conditions are one of the principal dangers for grape disease development. Downy Mildew, Powdery Mildew, Anthracnose, Stem borer, Black Rot, Leaf Blight are widespread grape leaf vermin and diseases, which cause stern monetary losses to the grape industry. Devices ready to quantify the climate conditions in real-time for disease onset are hence crucial to perform timely diagnosis and precise detection of grape leaf diseases. This will ensure the healthy growth of grape plants, further controlling the spread of diseases. This paper discusses the requirements for building a consistent grape disease detection framework that would encourage headways in agribusiness. The primary aim of this work is to adapt an Internet of Things (IoT) based approach to predict the occurrence of Downey and Powdery Mildew grape diseases at an early stage. The sensor values received are transmitted to the Central Server with the help of the IoT device NodeMCU. At the server side, an analysis is made based on weather conditions. Further notification to the farmer is sent if weather properties are conducive for disease onset. The exclusivity of the system lies in using a rain gauge sensor along with the temperature sensor to predict the occurrence of grape diseases. This system realizes an overall accuracy of 94.4% for Downey Mildew and 96% for Powdery Mildew. Experimental results suggest the projected model can proficiently recognize Downey and Powdery Mildew grape diseases.
