Smart Proactive Warning System for Early Prediction of Potato Late Blight Via Machine Learning

Abstract

Precision agriculture is one of the applications which may use the Machine learning technology to maximize the crops productivity, optimize the quality of the crops, and minimize the negative environmental impact like crops diseases. Crop diseases are generally one of the most critical problems that threaten the worlds agriculture, causing large losses in agricultural production of about 25% per year. Some metrological data like relative humidity, temperature, wind speed, wind direction, and pressure play an essential role in the creation and spreading of late blight disease of potato. The proposed system aims to predict a potato late blight using a machine learning (ML) IoT based system trained on dataset collected from field sensors at crops environment through forecasting weather data from weather stations. The proposed system is designed based on Logistic regression and Neural Networks where model results show performance and accuracy improvements compared with the result of built in service created on Microsoft Azure cloud using the same data set.