Forest Fire Dataset

This dataset comprises information related to forest fires and is intended for training algorithms designed for forest fire detection, alongside data for object detection. The section dedicated to fire classification consists of 2974 images, divided into two categories: the first category includes images depicting forest fires, while the second category contains images of intact forests without fires. As for the object detection data, it encompasses 1690 images, suitable for object detection purposes. These data have been distributed across training, validation, and test sets with proportions of 80%, 15%, and 5%, respectively. [1] A. Khan and B. Hassan, “Dataset for forest fire detection,” Mendeley Data, vol. 1, p. 2020, 2020, Accessed: Nov. 18, 2023. [Online]. Available: https://data.mendeley.com/datasets/gjmr63rz2r/1 [2] https://www.kaggle.com/datasets/phylake1337/fire-dataset These data were gathered from various sources on the internet and were manually filtered to ensure data integrity. Additionally, a portion of this data was generated manually by simulating forest fires after obtaining the necessary approvals from relevant authorities. This project is part of a master's thesis titled "Development of a Deep Learning-Based Surveillance System for Forest Fire Detection and Monitoring using UAV (İHA KULLANILARAK ORMAN YANGINLARININ TESPİTİ VE GÖRÜNTÜLENMESİ İÇİN DERİN ÖĞRENME TABANLI GÖZETLEME SİSTEMİNİN GELİŞTİRİLMESİ)" at Karabuk University in Turkey, conducted by the student Ibrahim Shmata and supervised by Dr. Batıkan Erdem Demir.本数据集包含与森林火灾相关的各类信息，旨在为森林火灾检测算法训练以及目标检测任务提供数据支撑。其中火灾分类模块包含2974张图像，分为两类：第一类为森林火灾场景图像，第二类为无火灾的完好森林图像。目标检测数据集包含1690张图像，可用于目标检测相关任务。上述数据已按照80%、15%、5%的比例分别划分为训练集、验证集与测试集。[1] A. Khan与B. Hassan，《用于森林火灾检测的数据集》，Mendeley Data，第1卷，第2020页，2020年，访问时间：2023年11月18日。[在线资源] 可获取链接：https://data.mendeley.com/datasets/gjmr63rz2r/1 [2] https://www.kaggle.com/datasets/phylake1337/fire-dataset 本数据集的原始数据均采集自互联网各公开渠道，并经过人工筛选以保障数据完整性。此外，部分数据是在获得相关主管部门许可后，通过模拟森林火灾的方式人工生成的。本数据集所属项目为土耳其卡拉布克大学（Karabuk University）学生易卜拉欣·什马塔（Ibrahim Shmata）在巴蒂坎·埃尔登·德米尔（Dr. Batıkan Erdem Demir）博士指导下完成的硕士学位论文《基于深度学习的无人机（Unmanned Aerial Vehicle, UAV）森林火灾检测与监测监控系统开发》（原土耳其语标题：İHA KULLANILARAK ORMAN YANGINLARININ TESPİTİ VE GÖRÜNTÜLENMESİ İÇİN DERİN ÖĞRENME TABANLI GÖZETLEME SİSTEMİNİN GELİŞTİRİLMESİ）的一部分。

搜集汇总

背景与挑战

背景概述

该数据集是一个专门用于森林火灾检测和物体检测的综合性图像资源，包含2,974张分类图像（分为活跃火灾和无火森林）和1,690张物体检测图像，数据按80%训练、15%验证和5%测试划分。数据集结合了真实在线收集和受控模拟数据，经过手动过滤以确保高质量，适用于机器学习与计算机视觉研究，旨在支持火灾检测系统开发，促进森林保护和灾害管理。