UAVs-FFDB: A high-resolution dataset for advancing forest fire detection and monitoring using unmanned aerial vehicles (UAVs)

Md Najmul Mowla ¹, Davood Asadi ¹, Kadriye Nur Tekeoglu ²

⁰Alparslan Türkeş Science and Technology University, Adana 1250, Turkey.

Data in Brief +

|

July 30, 2024

View abstract on PubMed

Summary

A new Unmanned Aerial Vehicles (UAVs) based forest fire database (UAVs-FFDB) offers crucial real-time data for early fire detection. This comprehensive dataset aids in developing advanced AI methods for forest fire monitoring and management.

Area Of Science

Forestry and Environmental Science
Computer Science and Artificial Intelligence
Remote Sensing and Geospatial Technology

Background

Forest ecosystems are increasingly vulnerable to wildfires, necessitating rapid and accurate detection systems.
Current real-time forest fire data accessibility and timeliness are insufficient for effective fire management.
Developing robust datasets is critical for advancing AI-driven solutions in wildfire surveillance.

Purpose Of The Study

To introduce the Unmanned Aerial Vehicles (UAVs) based forest fire database (UAVs-FFDB) to address the need for accessible, real-time forest fire data.
To provide a comprehensive dataset for training and validating AI models for early fire detection and monitoring.
To support the development of advanced methodologies for automated fire classification, recognition, detection, and segmentation.

Main Methods

Collected 1653 high-resolution RGB raw images using a UAV equipped with a RaspiCamV2 camera.
Augmented the raw image dataset to a total of 15560 images, increasing diversity and coverage.
Annotated all images using Makesense.ai for precise fire boundary demarcation and utilized UAVs for data acquisition at varying altitudes (5-15m) and speeds (2 m/s).

Main Results

The UAVs-FFDB comprises 1653 raw images (692 MB) and 15560 augmented images (6.76 GB), captured in a forested area near Adana, Turkey.
The dataset features images with varying dimensions, ensuring a comprehensive representation of potential fire scenarios.
Accurate annotations facilitate the development of precise AI algorithms for fire detection and segmentation.

Conclusions

The UAVs-FFDB provides a vital data infrastructure for researchers developing innovative early fire detection and continuous monitoring methodologies.
This dataset is foundational for advancing AI-based systems for automated forest fire analysis, improving ecosystem protection and human safety.
The UAVs-FFDB supports sustainable forest management practices through enhanced wildfire surveillance capabilities.

Keywords:

Convolutional neural networks Deep Learning Detection and classification Forest fire Machine learning Unmanned aerial vehicles