Scribd
Upload
357 views10 pages

CNN Model For Smoke and Fire Detection

The document presents a CNN model designed for real-time smoke and fire detection, detailing its architecture, training process, and various applications. It emphasizes the importance of early detection for safety, property protection, and emergency response, and outlines the steps for dataset preparation, model training, and performance evaluation. The model achieves high accuracy, precision, and recall, with potential applications in smart homes, public safety, industrial safety, and drone surveillance.

Uploaded by

Prateek Prusty
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
357 views10 pages

CNN Model For Smoke and Fire Detection

The document presents a CNN model designed for real-time smoke and fire detection, detailing its architecture, training process, and various applications. It emphasizes the importance of early detection for safety, property protection, and emergency response, and outlines the steps for dataset preparation, model training, and performance evaluation. The model achieves high accuracy, precision, and recall, with potential applications in smart homes, public safety, industrial safety, and drone surveillance.

Uploaded by

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

CNN Model for Smoke and

Fire Detection
This presentation explores a Convolutional Neural Network (CNN) model
for real-time smoke and fire detection, analyzing its architecture, training,
and applications.
Importance of Smoke and Fire Detection
Safety and Security Property Protection Emergency Response
Early detection is crucial for Smoke and fire detectors safeguard Prompt detection enables efficient
preventing catastrophic damage and buildings and valuable assets from emergency response, minimizing
loss of life. destructive fires. casualties and damage.
Overview of Convolutional
Neural Networks (CNNs)
Image Recognition Feature Extraction
CNNs excel at image CNNs use filters and
recognition tasks, extracting convolutional layers to
features from images and identify patterns and features
classifying them. within images.

Pattern Recognition
By analyzing features, CNNs can identify specific patterns, such as
smoke or flames.
CNN Architecture for Smoke and Fire Detection
Input Layer
1 Receives raw image data as input.

Convolutional Layers
2
Extract features from the image.

Pooling Layers
3
Downsample feature maps to reduce dimensionality.

Fully Connected Layers


4
Combine features and classify the image.

Output Layer
5
Provides the final prediction of smoke or fire.
Dataset Preparation and Preprocessing
Data Collection
1
Gather images of smoke, fire, and non-fire scenarios.

Data Labeling
2
Label each image as smoke, fire, or non-fire.

Data Augmentation
3
Increase dataset size by manipulating existing images.

Data Normalization
4
Standardize pixel values for consistent model training.
CNN Model Training and
Optimization

Training Optimization
Feed the model labeled images to Adjust model parameters to
learn patterns. minimize error and improve
accuracy.

Validation
Evaluate model performance on
unseen data to prevent overfitting.
Performance Evaluation and
Benchmarking

95% 98%
Accuracy Precision
Percentage of correct predictions. Proportion of true positive
predictions.

97% 96%
Recall F1 Score
Proportion of correctly identified Harmonic mean of precision and
fire or smoke events. recall.
Real-world Applications and
Future Developments
Smart Homes
1 Automatic fire detection and notification in residential
settings.

Public Safety
2 Early fire detection in public buildings for rapid response.

Industrial Safety
3 Preventing fires in hazardous industrial environments.

Drone Surveillance
4 Using drones to detect smoke and fire in remote areas.
Acknowledgment
This project acknowledges the invaluable contributions of the
researchers, engineers, and developers involved in the development of
the CNN model for smoke and fire detection.

You might also like