🌲🔥 Forest Fire Detection using CNN

This project detects forest fires from images using Convolutional Neural Networks (CNN). It is a deep learning-based image classification system that distinguishes between fire and non-fire images, supporting early detection of wildfires to reduce damage and protect the environment.

🔍 Project Overview

• Goal: Binary classification of images into Fire and No Fire.
• Dataset: Collected from Kaggle – Wildfire Image Dataset.
• Framework: TensorFlow/Keras for building and training the CNN model.
• Application: Can be used in real-time monitoring systems for fire-prone regions.

---

🧠 Deep Learning Concepts Used

Task	Method
Processing visual data	Convolutional Neural Networks (CNN)
Image classification	Supervised learning (Binary classification)
Vision-based AI detection	Roboflow + YOLO (alternative/extension)

🔧 Note: While YOLO is often used for object detection, this project focuses on image classification using CNN. YOLO integration can be done in future versions for real-time video surveillance.

---

🧪 Tools & Technologies

• Python
• TensorFlow / Keras
• OpenCV
• NumPy / Pandas
• Matplotlib / Seaborn
• Roboflow (for dataset visualization or deployment, optional)

---

📈 Project Pipeline

1. Data Collection & Loading

   • Dataset from Kaggle: fire and non-fire images.
   • Binary classification task.
   • Ensure all images are resized to the same dimensions (e.g., 224x224).

2. Image Preprocessing & Augmentation

   • Resize, normalize, and convert images to arrays.
   • Augment data (flip, rotate, zoom) using ImageDataGenerator.

3. Model Building

   • Construct a CNN using TensorFlow/Keras.
   • Input layer → Convolutional layers → MaxPooling → Dense layers → Output layer with sigmoid.

4. Model Training

   • Train the model on the preprocessed dataset.
   • Use validation split for performance monitoring.
   • Evaluate accuracy and loss using graphs.

5. Testing & Deployment

   • Test model on unseen images.
   • Optional: Deploy using Flask or Streamlit.