Yolov8 Fire Detection

Fire detection with YOLOv8 is an amazing project aimed at utilizing the powerful YOLOv8 object detection algorithm to detect fires in images or videos. Our repository provides a implementation of fire detection using YOLOv8, including training scripts, pre-trained models, and inference tools.

Fire Detection With Yolov8

Installation

Download the face detection repository:

# Clone repo
git clone https://github.com/Yusuf-ozen/Yolov8_Fire_Detection.git

Navigate to the project directory:

cd Yolov8_Fire_Detection

Install all necessary library:

pip install -r requirements.txt

Testing On Real-Time Webcam

Run this code at git bash or cmd:

python yolov8s_live_test.py

Testing on an Image

Run this code at git bash or cmd and change /path/image according your files. Using --resize_width 400 and --resize_height 400 the size of output image can change:

python yolov8s_image_test.py /path/image.jpg --resize_width 400 --resize_height 300

Testing on a Video

Run this code at git bash or cmd and change /path/image according your files. Using --resize_width 400 and --resize_height 400 the size of output of the video can change:

python yolov8s_video_test.py /path/video.mp4 --resize_width 1280 --resize_height 720

Results

-This results produced after 50 epochs with yolov8s model and Fire-Dataset.

F1 Curve	P Curve	PR Curve

Confusion Matrix	R Curve	results

Predictions

label	Prediction

Test Model with Streamlit

-Using Streamlit framework we can show the results of yolov8 model in web application.

Usage

streamlit run main.py

Image Testing

Youtube Video Testing

References

https://roboflow.com/

https://github.com/ultralytics/ultralytics

https://github.com/CodingMantras/yolov8-streamlit-detection-tracking/tree/master