Scribd
Upload
10 views8 pages

Week_3_Project_PPT_Template1

The document outlines a project focused on using Convolutional Neural Networks (CNNs) for detecting forest fires through image classification. It details the methodology including data collection, preprocessing, model architecture, and training, highlighting the need for an automated detection system to improve response times to wildfires. The project successfully built a CNN that achieved good accuracy and can serve as a foundation for real-time detection systems.

Uploaded by

hicomo9645
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
10 views8 pages

Week_3_Project_PPT_Template1

The document outlines a project focused on using Convolutional Neural Networks (CNNs) for detecting forest fires through image classification. It details the methodology including data collection, preprocessing, model architecture, and training, highlighting the need for an automated detection system to improve response times to wildfires. The project successfully built a CNN that achieved good accuracy and can serve as a foundation for real-time detection systems.

Uploaded by

hicomo9645
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 8

Forest Fire Detection Using Deep

Learning
Learning Objectives

•Understand how to use Convolutional Neural Networks (CNNs) for


image classification tasks.
•Learn how to preprocess image datasets and train models on Google
Colab.
•Gain experience using ImageDataGenerator for real-time data
augmentation.
•Evaluate and visualize model performance using matplotlib.
GOAL

Source : www.freepik.com/
Tools and Technology used

•Programming Language: Python 🐍


•Deep Learning Framework: TensorFlow + Keras
•Dataset Source: KaggleHub
•Platform: Google Colab
•Version Control: GitHub
•Visualization: Matplotlib
•Libraries: NumPy, PIL, TensorFlow, Matplotlib
Methodology
1. Data Collection:
Kaggle Wildfire Dataset with images labeled as fire and nofire.
2. Preprocessing:
•Resized to (150x150)
•Rescaled pixel values
•Used ImageDataGenerator
3. Model Architecture:
•3 Conv2D + MaxPooling layers
•Flatten → Dense(512) + Dropout
•Output layer: Sigmoid for binary classification
4. Training & Validation:
•Epochs: 12
•Batch size: 32
•Evaluated using Accuracy & Loss curves
5. Testing:
•Final accuracy evaluated on the test set
•Single image predictions tested
Problem Statement:

Problem:
Wildfires are destructive and can spread rapidly, causing damage to
ecosystems, property, and human life. Manual detection is slow and may lead
to delayed response times.

Need:
There is a need for an automated system to detect fire from images, which
can help authorities respond faster.
Solution:

The project builds a CNN-based image classifier that:


• Automatically detects fire in images
• Can be extended to a real-time video-based detection system
• Reduces human delay in identifying forest fires
Screenshot of Output:
Conclusion:

✅ Successfully built and trained a CNN to classify wildfire


images
✅ Achieved good accuracy and performance on unseen data
✅ This model can serve as a base for real-time wildfire
detection systems
✅ Learned end-to-end machine learning workflow from data
collection to deployment

You might also like