DEEP LEARNING-ENABLED INTERNET OF THINGS SYSTEMS FOR INTELLIGENT SOIL HEALTH MONITORING IN PRECISION AGRICULTURE

Abstract

The relevance of sustainable agriculture, coupled with challenges posed by climate change, soil degradation and resource mismanagement.  At present, the need for intelligent soil health monitoring system has gained increasing significance. Soil health is very critical for productivity of the crops due to influence on nutrient availability, water retention capability, microbial activities and sustainability of the ecosystem. Conventional approaches in soil assessment are based on manual soil sampling and lab analysis, which are extremely laborious, costly and time consuming, and hence unsuitable for precision agriculture. The emergence of Internet of Things (IoT) technology and Deep Learning (DL) offers new horizons for development of intelligent soil monitoring systems that would continuously collect, analyze and interpret the soil data to support decision making in agriculture. The integration of sensors and deep learning would enable AI-powered soil monitoring systems to monitor the soil conditions such as moisture, temperature, pH, electrical conductivity, and nutrients content and to process huge amounts of data to search for complex patterns and make forecasts about soil conditions and make well-informed decisions in farming. In this research, the whole system of deep learning-based IoT system for intelligent soil health monitoring for precision agriculture has been proposed. The technologies of IoT sensor networks, cloud computing, and data preprocessing techniques, CNN (convolutional neural network) and LSTM (long short-term memory) networks with decision support mechanisms are proposed for the combination.