PlantsCare IoT
Agricultural IoT from hardware to web: custom ESP32 nodes (water, air, camera), MQTT/CoAP, InfluxDB, EfficientNet-B0 image classification and AWS deployment.
What it is
A complete, hardware-to-web agricultural monitoring system, built during my time at the Aragón Institute of Technology (ITA). Not just a dashboard: it includes the sensor-node firmware, the IoT communication layer, the time-series database, AI image classification and the cloud deployment.
What I built
- Three ESP32 node types with custom firmware: water (irrigation/humidity), air (environment) and camera — each with its own dedicated firmware.
- Real IoT communication: MQTT and CoAP between nodes and server.
- Time series in InfluxDB, built for continuous sensor readings.
- Plant-image classification with EfficientNet-B0 (TensorFlow Lite) for early problem detection.
- Monitoring web app with Astro + React and a Node backend.
- Deployed on AWS EC2, with the electronics documented (sensor and microcontroller datasheets in the repo).
Why it matters
This is the project that joins my two worlds: real electronics/IoT (firmware, protocols, physical sensors) and monitoring software (time series, applied AI, web). The same combination I apply today to critical infrastructure at OFITECO.
Stack
ESP32 (C/C++ firmware) · MQTT · CoAP · InfluxDB · TensorFlow Lite (EfficientNet-B0) · Astro · React · Node.js · AWS EC2.