Smart Irrigation System

Overview

This project implements a deep learning-based smart irrigation system that predicts irrigation requirements based on environmental parameters. It leverages sequential (time-series) data using a hybrid architecture combining 1D Convolutional layers and LSTM-based Recurrent Neural Networks (RNN). Hyperparameter optimization has been performed using the GridSearchCV method to enhance model performance.

📁 Project Contents

• parameter_tuning.ipynb: Hyperparameter tuning using Keras Tuner.
• rnn.ipynb: RNN model for irrigation prediction.
• inference.ipynb: Inference pipeline for predicting irrigation needs.
• Smart_irrigation_project_dataset.csv: Dataset used for training and evaluation.
• README.md: Guide

📊 Dataset

The dataset used in this project is sourced from: Kulkarni, Rohan (2023), “Smart_irrigation_project_dataset”, Mendeley Data, V1 DOI: 10.17632/krsjvfvbsk.1

License: CC BY 4.0
You are free to share and adapt the dataset with proper attribution.

🧠 Model Details

• Architecture: Hybrid Deep Learning model combining 1D Convolutional layers (Conv1D) with Long Short-Term Memory (LSTM) layers.
• Model Type: Sequential Recurrent Neural Network (RNN) with temporal feature extraction.
• Frameworks: TensorFlow
• Techniques:
  • Conv1D for spatial pattern recognition in time-series data
  • Stacked LSTMs for capturing long-term dependencies
  • Hyperparameter optimization using GridSearchCV

🚀 Get Started

Follow the steps below to set up and run the Smart Irrigation System on your local machine:

1. Clone the Repository

git clone https://github.com/atharva-khachane/smart-irrigation.git
cd smart-irrigation

2. Install Dependencies

Ensure you have Python 3.7+ installed, then install required packages:

pip install -r requirements.txt

3. Dataset

Place the dataset file (Smart_irrigation_project_dataset.csv) in the root directory.

Before training, manually split the dataset into training and testing sets while preserving the sequential order of the data.
Note: Avoid using train_test_split from scikit-learn, as it shuffles the data and may disrupt the time-series sequence required for accurate model predictions.

4. Run Notebooks

You can open and run the following notebooks in Jupyter or VS Code:

• parameter_tuning.ipynb – Hyperparameter tuning using GridSearchCV.
• rnn.ipynb – Builds and trains the Conv1D + LSTM model.
• inference.ipynb – Runs inference on test data using the trained model.

📜 License

This project is licensed under the MIT License. See the LICENSE file for details.

🙌 Acknowledgements

Thanks to Rohan Kulkarni for the open dataset.