Forest Fire Detection using Convolutional Neural Networks

Forest-fire detection model using TensorFlow and Keras, achieving 91.38% training accuracy and 90.10% validation accuracy. This convolutional neural network (CNN) significantly enhances early detection capabilities for forest fires, integrating advanced deep learning techniques for accurate prediction and prevention.

Table of Contents

• Project Overview
• Dataset
• Dependencies
• Project Structure
• Code Explanation
  • Data Loading
  • Data Preprocessing
  • Model Architecture
  • Model Training
  • Evaluation and Visualization
• Results
• Model Saving
• Usage
• License

Project Overview

This project involves building a binary classifier using a CNN to detect forest fires. The model is trained on a dataset of images, categorized as either 'Fire' or 'Non-Fire'.

Dataset

The dataset used for training consists of images stored in two directories:

• Fire: Contains images with forest fires.
• Non_Fire: Contains images without forest fires.

The dataset is loaded from a directory structure and is processed into a pandas DataFrame for further processing.

Dependencies

The project relies on the following Python libraries:

• pandas
• numpy
• matplotlib
• tqdm
• keras
• tensorflow
• seaborn
• sklearn

Accuracy and Loss Graphs

You can install the required packages using:

pip install -r requirements.txt