pHorest: AI-Driven Precision Soil Sustainability

Table of Contents

• Overview
• Features
• ML Algorithm & Training
• Project Structure
• Setup
• App and Script flags
• Browser Dashboard

Overview

Overview.mp4

pHorest is a cyber-physical system designed to combat soil acidification and nutrient runoff—major environmental concerns in modern chemistry. By integrating IoT sensors with a Random Forest Machine Learning model, the system identifies the soil's chemical state and provides precise neutralization strategies through recommendations of crops and fertilizers.

Features

• Hardware Sensing: Real-time pH, TDS, and Temperature monitoring via Arduino.
• Cloud Rainfall Input: 30-day rainfall is fetched online by location (Open-Meteo) for model input.
• AI Inference: Random Forest Classifier trained on agricultural datasets.
• Smart Recommendations: Crop prediction + fertilizer suggestion with pH-aware chemistry advice.

ML Algorithm & Training

Model used

• Algorithm: RandomForestClassifier
• Input features (strict order): Nitrogen, phosphorus, potassium, temperature, humidity, ph, rainfall
• Target label: label (crop name)
• Train/test split: 80/20 using train_test_split(..., test_size=0.2, random_state=42)
• Model params used: n_estimators=100, random_state=42
• Saved model artifact: random forest/model/soil_model.pkl (via joblib.dump)

Training environment

• Trained in Jupyter Notebook.
• Training code used: same as random forest/training/train.ipynb in this repo
• Dataset used: Crop_recommendation.csv (same schema as random forest/dataset/Crop_recommendation.csv in this repo).

Using the trained model in this project

1. Place random forest/model/soil_model.pkl in the random forest/model/ directory.
2. Ensure runtime feature order exactly matches training feature order.
3. Run scripts/ser_script.py, scripts/sim_script.py, or app/app.py to load the model and predict crops.

Project Structure

• app/app.py: Browser dashboard (live readings + crop + fertilizer recommendation).
• scripts/: Contains utility scripts.
  • ser_script.py: Local Python bridge between Arduino and ML model.
  • sim_script.py: Simulation script for software-only demonstration.
• random forest/: Contains all machine learning related files.
  • dataset/: Contains the dataset.
    • Crop_recommendation.csv: The dataset used for training.
  • model/: Contains the trained model.
    • soil_model.pkl: Serialized Random Forest model.
  • training/: Contains the training notebook.
    • train.ipynb: Jupyter notebook for training the model.
• docs/setup.md: Required hardware components and Arduino wiring map.
• arduino/: Directory containing Arduino firmware files for data acquisition.

Setup

Before proceeding, Set up hardware per docs/setup.md and flash arduino/arduino_full/arduino_full.ino via Arduino IDE.

1. Connect Arduino (for live/serial mode), commonly /dev/ttyACM0 or /dev/ttyUSB0 for Linux/macOS, or COM3 for Windows.

2. Install dependencies:

   Linux / macOS:

   python3 -m pip install pandas joblib pyserial scikit-learn flask

   Windows:

   py -m pip install pandas joblib pyserial scikit-learn flask

Run

To run the scripts, use the appropriate command for your operating system:

• Terminal:

  Linux / macOS:

  # Simulation
  python3 scripts/sim_script.py
  
  # Serial
  python3 scripts/ser_script.py --port=/dev/ttyACM0

  Windows:

  # Simulation
  py scripts/sim_script.py
  
  # Serial
  py scripts/ser_script.py --port=COM3

• Browser dashboard:

  Linux / macOS:

  # Simulation (default)
  python3 app/app.py
  
  # Serial (explicit port)
  python3 app/app.py --serial /dev/ttyACM0
  
  # Serial (auto-detect first available Arduino-like port)
  python3 app/app.py --serial

  Windows:

  # Simulation (default)
  py app/app.py
  
  # Serial
  py app/app.py --serial COM3

Flags

App flags (app/app.py)

• app/app.py → starts in SIM mode
• app/app.py --sim → explicitly starts in SIM mode
• app/app.py --serial <PORT> → starts in SERIAL mode
• app/app.py --serial → starts in SERIAL mode and auto-detects the port if /dev/ttyACM0 is unavailable
• --lock-mode → hides the mode selector and prevents switching modes during runtime
• --no-check → skips strict pH input range validation (chemical status still shown)

Usage examples:

# Start in simulation mode
python3 app/app.py

# Start in serial mode with a specific port
python3 app/app.py --serial /dev/ttyACM0

# Start in serial mode with auto-port detection and locked UI
python3 app/app.py --serial --lock-mode

the mode selector can be hidden by pressing 'm' in normal mode.

Script Flags (scripts/sim_script.py, scripts/ser_script.py)

These scripts support optional flags to provide rainfall data:

• --location <LAT> <LON>: Fetches 30-day rainfall data from an online weather API (Open-Meteo) based on the provided latitude and longitude.
  • If the API call fails, the script will fall back to using the --rainfall_data value if provided, otherwise it will use a default value (100.0mm) and print a warning.
• --rainfall_data <VALUE>: Directly specifies the 30-day rainfall value in millimeters.
  • If both --location and --rainfall_data are provided, --location is attempted first. If it fails, --rainfall_data is used as a fallback.

Usage examples:

# Simulation with location-based rainfall
python3 scripts/sim_script.py --location 34.0522 -118.2437

# Simulation with direct rainfall data
python3 scripts/sim_script.py --rainfall_data 150.0

# Serial with location-based rainfall
python3 scripts/ser_script.py --port=/dev/ttyACM0 --location 34.0522 -118.2437

# Serial with direct rainfall data
python3 scripts/ser_script.py --port=/dev/ttyACM0 --rainfall_data 150.0

🌐 Browser Dashboard

Open http://127.0.0.1:5000 after starting app/app.py.

• Set your coordinates using Use Browser Location or manual latitude/longitude.
• The app fetches rolling 30-day rainfall (mm) online and feeds it to the model.
• If location/weather fetch is unavailable, the app falls back to cached or default rainfall and shows the source status in UI.