Quantum Machine Learning

2Profiling Crop Water Stress Using Modular Neural Network Classifier for an IoT-Enabled Crop Irrigation System

Jeyapandian Munisamy and Durga Karthik^*

Department of Computer Science and Engineering, Srinivasa Ramanujan Centre, SASTRA Deemed University, Kumbakonam, India

Abstract

In agriculture, water is the most essential factor for crop production and its related activities. Farmers face difficulties in the irrigation process due to scarce groundwater, flooding, under-irrigation, over-irrigation, unpredictable rainfall, etc. It is the need of the hour to conserve water due to poor irrigation management system so as to save plants and crops from water stress and to protect the water table. The proposed work aims to reduce the crop water stress for a cotton crop land by combining infield sensor data and metrological data. Infield climate data such as soil moisture, air temperature, and relative humidity of the cotton crop land is collected with the help of smart sensors. Metrological data such as air temperature, wind speed, evaporation, and rainfall was collected from Indian Rice Research Institute (IRRI), Aduthurai, Tamil Nadu. A modular neural network classifier (MNNC) is implemented to identity crop water stress from the collected infield sensor data and metrological data. The proposed MNNC model is formed by combining two sequential neural networks: a first module that processes the input features and a second module that combines the output from the first module, integrates ...