Smart Water and Fertilizer Control System for Greenhouse Tomatoes

Abstract

To address the issues of strong empirical dependence, delayed regulation, and low resource utilization in traditional greenhouse tomato water and fertilizer management, this study designed and implemented a multi-source data-driven smart water and fertilizer control system. The system integrates multiple sensing modalities, including environmental sensing, crop stem flow monitoring, leaf thermal imaging and reflectance spectral analysis, and substrate/nutrient solution detection. Relying on a LoRa wireless network and a cloud-edge collaborative computing architecture, it achieves dynamic data acquisition and preprocessing in greenhouses (incorporating Savitzky-Golay filtering, Huffman coding compression, and $3 \sigma$ outlier removal). Furthermore, the system incorporates a coupled water-fertilizer prediction model based on the growth stage and residual soil nutrients. The execution layer, centered on the ESP32S2 microcontroller, employs Pulse Width Modulation (PWM) combined with flow feedback to achieve closed-loop precision control. System testing demonstrates that the proposed architecture meets the typical application requirements of facility agriculture in terms of communication stability, control accuracy, and response real-time performance. This study provides a comprehensive system design and technical reference for the intelligent water and fertilizer management of greenhouse tomatoes