Fire detection project

[Cho-Hong-Seok]

Search before asking

• I have searched the YOLOv5 issues and discussions and found no similar questions.

Question

@glenn-jocher
I am working on a project to detect fire using YOLOv5&YOLOv8 models. The project model I would like to implement is a program that detects the initial fire with a lightweight model like YOLOv8s, and then uses a more accurate model like YOLOv8x to track -> segment and finally alarm as the fire and smoke grows. I was wondering if anyone could advise me on any libraries, frameworks, equipment, etc. to implement this?

The first picture below is a crude representation of the sort of process I'd like to implement.
The second picture is an image captured from roboflow youtube and it shows an implementation using the gradio library to switch to bbox->segmentation when a boundary is crossed using the YOLO-World + EfficientSAM model. My goal is to implement something similar to this for the fire object.

Thanks as always!

Additional

[glenn-jocher]

@Cho-Hong-Seok hello! 🚀 Your fire detection project sounds quite innovative, and using YOLOv5 and YOLOv8 for initial detection and then tracking to segmentation and alarm as the fire grows is a smart approach. For implementing this multi-stage process, you would want to consider a few aspects:

1. Libraries/Frameworks: For integration and switching between models (like YOLOv8s for initial detection and YOLOv8x for detailed tracking), consider using a deep learning framework like PyTorch which YOLO is based on. It offers flexibility in model management and deployment. For segmentation, you might look into segmentation models available in PyTorch or libraries like Detectron2 which are compatible with YOLO outputs.

2. Equipment: Depending on the deployment environment (real-time detection, offline processing, etc.), ensure you have the necessary computational resources. GPUs are highly recommended for model training and inference, especially for the more accurate (thus computationally intensive) models like YOLOv8x.

3. Application logic: For the application logic that switches between detection and tracking/segmentation based on certain criteria, you may need to implement custom code that monitors detection confidence, object size, or other metrics that signal when to transition from one model to another.

Remember that any usage of Ultralytics models, architectures, or code in your project at any stage requires an Ultralytics Enterprise License or fully open-sourcing your entire project under AGPL-3.0.

I hope these suggestions are helpful! 😊 For further details and documentation on Ultralytics YOLOv5, please visit the Ultralytics Docs. Best of luck with your fire detection project!

[github-actions]

👋 Hello there! We wanted to give you a friendly reminder that this issue has not had any recent activity and may be closed soon, but don't worry - you can always reopen it if needed. If you still have any questions or concerns, please feel free to let us know how we can help.

For additional resources and information, please see the links below:

• Docs: https://docs.ultralytics.com
• HUB: https://hub.ultralytics.com
• Community: https://community.ultralytics.com

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLO 🚀 and Vision AI ⭐