Validation of Jhulsacast Model using Human Participatory Sensing and Wireless Sensor Networks

Abstract

A wireless sensor network was deployed using four sensor nodes to collect temperature and humidity data from the farm in a region of interest. This study demonstrated implementation of the Central Potato Research Institute’s Jhulsacast model for forecasting of potato late blight disease using these sensor observations. A novel human participatory sensing mobile phone application was developed for diagnosis of the disease at the farm level. The validation of the Jhulsacast model, and its comparison with other models was done using the human participatory sensing approach. The study indicated that unlike the existing methods which rely on disease forecasting to control the plant diseases, the decisions on pesticide application can be made by jointly considering diagnosis and forecasting information. Experimental results depicted that the Jhulsacast Model has been found to be significantly accurate than the Ullrich, Fry, Winsteland Wallin models for the Hapur region of Uttar Pradesh, India.