Year: 2026
Type: article
Abstract: Smallholder agriculture often relies on manual observation and uncalibrated soil assessment, which leads to inefficient resource management, improper fertilization, and unsuitable crop selection. Internet of Things (IoT) technologies have lowered the cost of field monitoring, but many existing systems only visualize sensor data and do not turn measurements into actionable agricultural recommendations. To address this limitation, this paper presents an end-to-end, solar-powered smart agriculture monitoring and decision-support system built around an ESP32 microcontroller edge node. The system continuously monitors five parameters: soil moisture, air temperature, relative humidity, ammonia (NH3), and methane (CH4). The firmware averages 10 samples of each analog sensor reading and applies a two-point linear calibration to the soil moisture sensor. The processed data are transmitted over two concurrent Wi-Fi paths: to the Blynk IoT platform every 3 seconds for real-time monitoring, and to a Flask REST server every 15 seconds as JSON payloads. At the server, a Decision Tree machine learning model, trained on an agricultural dataset of 50 crop varieties, analyzes the sensor data to estimate the expected crop yield percentage and to generate a ranked list of suitable crops. Sensor data and prediction results are logged in an SQLite database and displayed on an auto-refreshing web dashboard, and the predictions are also sent to Blynk virtual pins for remote mobile access. A photovoltaic solar panel, a charge controller, and a rechargeable lithium-ion battery power the node, allowing autonomous operation. The proposed architecture combines field-level sensing, wireless communication, cloud monitoring, database management, and machine learning-based prediction into an affordable, energy-independent decision-support framework for precision agriculture. Keywords—Precision agriculture, Internet of Things, ESP32, Decision Tree, crop recommendation, yield prediction, solar energy harvesting.
Language: English
Topic: Smart Agriculture and AI
Subfield: Plant Science
Domain: Life Sciences
Open Access status: bronze
OpenAlex ID: W7225331792