🧯 Web-Based Fire Safety Inspection Platform Using ML Services 🤖📸

This project is a web-based tool designed to simplify the process of fire safety inspection and documentation utilizing FSE ML services. It provides an intuitive interface for users to annotate images, manage datasets, and iteratively improve object detection models.

This repository is based on the research presented in the following publications:

• Aziz, A., & König, M. (Forthcoming). Creation of a web-based tool for the visual inspection of building equipment.
• Soultana, A., & Aziz, A. (2023). Active learning approach for object detection in technical building equipment images. Ruhr-Universität Bochum.

Overview

This repository provides:

• A Backend API built with FastAPI to handle model inference, training, and data management.
• A Frontend Interface for annotating images, managing datasets, and visualizing results.
• Pre-trained and executable ML models for detecting FSE objects and FSE-related information in images.
• An Active Learning Framework for object detection tasks, enabling iterative model improvement.
• Tools for data preparation, including downloading and managing validation and rare images.´
• Dockerized setup for easy deployment and scalability.

Setup

A NVIDIA GPU is required to run the backend!

Step 1 (Clone the Git Repository):

URL: https://github.com/RUB-Informatik-im-Bauwesen/fse-web-tool.git (github.com)

Step 2

Download docker desktop: https://www.docker.com/products/docker-desktop/

Step 3:

Build and start the services defined in the docker-compose.yml file:

docker compose build
docker compose up -d

Step 4:

Acess Localhost at: http://127.0.0.1:5173/

Step 5:

Enjoy!

Downloading Models and Images:

To use this application, you need to obtain YOLOv8 weights and organize them as described below. The weights for the YOLOv8 services can be requested by contacting Angelina Aziz via email at [email protected]. Additionally, the current implementation of the Active Learning framework is based on YOLOv5.

Directory Structure for Weights

Organize the weights for each service in the following structure:

storage/
└── Visual_Annotation_Tool/
    ├── Detection_Condition_amodal_Yolov8/
    │   └── best.pt
    ├── Detection_Condition_modal_Yolov8/
    │   └── best.pt
    ├── Detection_fire_class_symbols_Yolov8/
    │   └── best.pt
    ├── Detection_FSE_Yolov8/
    │   └── best.pt
    └── Detection_marking_Yolov8/
        └── best.pt

• Download validation images from this link: (https://drive.google.com/file/d/1nueXWhQNkAwiC7LDdTp8bbO2GRrOHo88/view?usp=sharing) and replace the folder in Storage.
• Download rare images from this link: (https://drive.google.com/file/d/1jgORHgpLZ_uT1sp0gbRu2Gpfae_wyDKU/view?usp=sharing) and replace the folder in Storage.