FireFly
收藏资源简介:
FireFly数据集是由南加州大学创建的合成数据集,用于检测森林火灾中的余烬。该数据集包含19,273张图像,涵盖了多种环境条件,如天气、光照和大气能见度。数据集的创建过程涉及使用虚幻引擎4(UE4)自动生成带有可调参数的标签数据,以增加数据的多样性和定制性。FireFly数据集主要应用于提高机器学习模型在真实世界森林火灾场景中的检测精度,特别是在减少误报和漏报方面。
The FireFly dataset is a synthetic dataset developed by the University of Southern California for detecting embers in forest fires. It consists of 19,273 images covering a wide range of environmental conditions including weather, lighting, and atmospheric visibility. The dataset was created using Unreal Engine 4 (UE4) to automatically generate labeled data with adjustable parameters, thereby enhancing data diversity and customizability. The FireFly dataset is primarily used to improve the detection accuracy of machine learning models in real-world forest fire scenarios, especially in reducing false positives and false negatives.

- 1FireFly A Synthetic Dataset for Ember Detection in Wildfire南加州大学 · 2023年



