A New Technique of Soluble NPK Nutrient Detection System Using Deep Learning-Based Recurrent Neural Networks (RNN)

Modern agriculture relies heavily on effective process monitoring, control, and remote data collection, making accurate and cost-effective sensors crucial for innovation. This research focuses on the development of a spectroscopy-based device for real-time and remote monitoring of nutrient levels in hydroponic systems. The device utilizes IoT architecture, incorporating principles of light absorbance to quantify concentrations of essential nutrients during crop growth. Three deep learning techniques, Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), and Recurrent Neural Networks (RNN) were adopted for data analysis. Results demonstrate that tailored implementation of deep learning can significantly enhance the performance of sensor network system (SNS). Specifically, the RNN format achieved an impressive 98% model accuracy during both training and testing phases. The technological advancements in addressing contract variation and photometric levels contribute to increased accuracy as more sensors become available. The findings serve as a comprehensive guide for collecting extensive agricultural data, encompassing environmental factors throughout the entire cultivation period, using state-of-the-art remote control sensor equipment. This work provides a practical, efficient, and cost-effective solution for precise nutrient control and continuous data monitoring in agricultural production.