🌿 Smart Plant Monitoring and Control System

A complete AI-powered smart greenhouse built using Raspberry Pi, Arduino, and FastAPI, with support for Gemini and Ollama AI models. This system monitors temperature, humidity, and soil moisture, captures live plant images, and uses AI to recommend ideal conditions and control devices like pumps, fans, lights, and humidifiers automatically.

---

🧠 System Overview

This project combines:

• Arduino Nano/Uno → reads sensors and sends data via Serial.
• Raspberry Pi → runs Python backend, logs data, captures images, queries AI, and controls relays for automation.
• AI Models (Gemini/Ollama) → analyze plant images and sensor data to determine ideal conditions.

---

🧩 System Architecture

                ┌──────────────────────┐
                │      Sensors (DHT11, │
                │ Soil Moisture Sensor)│
                └──────────┬───────────┘
                           │
                     [Serial /dev/serial0]
                           │
                    ┌──────▼──────┐
                    │  Arduino    │
                    │  nano_multi │
                    └──────┬──────┘
                           │
                     USB / Serial
                           │
                ┌──────────▼──────────┐
                │ Raspberry Pi        │
                │ sensors.py          │
                │ server.py (FastAPI) │
                └──────────┬──────────┘
                           │
                     Camera + AI
                           │
                 ┌─────────▼─────────┐
                 │ Gemini / Ollama AI│
                 └───────────────────┘

---

⚙️ Hardware Components

Component	Description
Arduino Nano / Uno	Reads sensor values and sends via Serial
DHT11 Sensor	Measures temperature and humidity
Soil Moisture Sensor	Measures soil wetness
Raspberry Pi	Main controller (AI + automation)
Relay Module (5-Channel)	Controls external devices
Camera (Picamera2)	Captures plant images for AI analysis
Actuators	Water pump, fan, light, humidifier, exhaust

---

🔌 Pin Configuration

Raspberry Pi → Relay Mapping

Relay	Function	GPIO Pin
1	Water Pump	23
2	Grow Light	24
3	Fan	25
4	Humidifier	26
5	Exhaust Fan	27

Arduino Nano Sensor Pins

Sensor	Signal Pin	Description
DHT11	D2	Temperature & Humidity
Soil Moisture	A0	Analog moisture sensor
TX/RX	D0/D1	Serial communication with Raspberry Pi

---

🔧 Software Components

File	Description
nano_multi.ino	Arduino firmware — reads sensors, prints data to serial
DHT.h	DHT11 library for Arduino
logs.py	Reads Arduino serial data, parses and logs it to JSON
sensors.py	Controls GPIO relays, captures images, and executes AI-based actions
server.py	FastAPI backend providing API endpoints for monitoring and AI analysis
logs.json	Stores historical sensor readings
last_response.json	Stores AI-generated ideal plant conditions
req.txt	Python dependencies list

---

🧠 Arduino Firmware (nano_multi.ino)

The Arduino reads data from the DHT11 and Soil Moisture Sensor and sends it to the Raspberry Pi in this format:

Arduino: Moisture: 532 | Humidity: 64.0 % | Temperature: 30.0 *C

These values are continuously logged and used for AI analysis and decision-making.

Sample Output:

Arduino: Moisture: 450 | Humidity: 63.0 % | Temperature: 29.0 *C

---

💻 Raspberry Pi Software Setup

1. Clone the Repository

git clone https://github.com/yourusername/smart-plant-ai.git
cd smart-plant-ai

2. Install Dependencies

pip install -r req.txt

3. Connect Hardware

• Connect Arduino via /dev/serial0

• Wire relays and sensors as shown above

• Enable camera:

  sudo raspi-config

  → Interface Options → Enable Camera

4. Run the FastAPI Server

uvicorn server:app --host 0.0.0.0 --port 8000

---

🌐 API Endpoints

Endpoint	Method	Description
/sensors	GET	Reads current sensor values
/capture	GET	Captures a new image from the Pi Camera
/ai	GET	Sends data to Ollama AI model for analysis
/ai/gemini	GET	Sends data to Google Gemini model for analysis
/act	GET	Executes actions (pump, fan, etc.) based on AI output

---

🧠 AI Integration

Ollama (Local)

• Runs locally on another machine or network.
• Modify server.py → url = "http://<your_ollama_ip>:11434/api/generate"

Gemini (Cloud)

• Requires a valid Google API key.

• Add it in:

  client = genai.Client(api_key="YOUR_API_KEY")

Example AI Output

{
  "plant_name": "Zephyranthes sp. (Rain Lily)",
  "temperature": 24,
  "humidity": 60,
  "moisture": 500
}

This JSON is saved in last_response.json and used by sensors.py to control devices.

---

🤖 Automated Actions

The Raspberry Pi reads last_response.json and compares AI-recommended values with real-time sensor data:

Condition	Action
Moisture < Ideal	Activates water pump
Temperature < Ideal	Turns on heater/light
Humidity < Ideal	Turns on humidifier & fan
Else	Keeps system idle

All actuators are controlled via GPIO relays.

---

🪵 Data Logging

• logs.json → Continuous sensor data log
• last_response.json → AI recommendation record

You can easily visualize logs using any plotting tool or dashboard.

---

🧰 Troubleshooting

Issue	Possible Cause	Solution
Serial not reading	Wrong port	Check /dev/serial0 or ls /dev/tty*
Camera not working	Disabled in config	Enable via sudo raspi-config
Relays not toggling	Wrong GPIO mode	Ensure GPIO.setmode(GPIO.BCM)
AI not responding	Missing API key / bad URL	Check Gemini key or Ollama host IP

---

🌱 Future Enhancements

• Web dashboard for real-time visualization
• MQTT or WebSocket support for IoT integration
• Cloud data storage and analytics
• AI-based disease detection using captured images