主管：中华人民共和国应急管理部
主办：应急管理部天津消防研究所
ISSN 1009-0029  CN 12-1311/TU
消防科学与技术

消防科学与技术 ›› 2025, Vol. 44 ›› Issue (9): 1281-1286.

• • 上一篇    下一篇

基于改进CycleGAN的大跨空间场所火灾合成样本增强方法

李晓旭1,2,3, 张曦1,2,3, 于春雨1,2,3, 李泊宁1,2,3   

  1. （1.应急管理部沈阳消防研究所，辽宁 沈阳 110034； 2.辽宁省火灾防治技术重点实验室，辽宁 沈阳 110034； 3.消防与应急救援国家工程研究中心，辽宁 沈阳 110034）
  • 收稿日期:2025-05-22 修回日期:2025-08-04 出版日期:2025-09-15 发布日期:2025-09-15
  • 作者简介:李晓旭，应急管理部沈阳消防研究所研究第二研究室研究实习员，主要从事火灾探测研究，辽宁省沈阳市皇姑区文大路218-20号，110034，15998138995@163.com。
  • 基金资助:
    应急管理部重点科技计划项目（2024EMST111105）；应急管理部消防救援局科技计划项目（2020XFZD12）

Enhancement method for synthetic samples of large-span space fire based on improved CycleGAN

Li Xiaoxu1,2,3, Zhang Xi1,2,3, Yu Chunyu1,2,3, Li Boning1,2,3   

  1. (1. Shenyang Fire Science and Technology Research Institute of MEM, Shenyang Liaoning 110034, China; 2. Liaoning Key Laboratory of Fire Prevention Technology, Shenyang Liaoning 110034, China; 3. National Engineering Research Center of Fire and Emergency Rescue, Shenyang Liaoning 110034, China)
  • Received:2025-05-22 Revised:2025-08-04 Online:2025-09-15 Published:2025-09-15

摘要： 为了解决大跨空间场所火灾图像样本库不足导致火灾检测准确率低的问题，提出了大跨空间场所火灾合成样本增强方法。该方法基于改进的CycleGAN网络，构建火焰块生成网络和火灾图像生成网络，火焰块生成网络生成火焰块图像，火灾图像生成网络将火焰块融入工程场景指定区域，生成可用于目标检测模型的火灾样本。结果表明：模型生成的合成样本显著提升目标检测模型性能，mAP@0.5提升25.11%，验证了数据增强方法的有效性。通过生成高质量、多样化的大跨空间火灾图像，有效缓解了火灾样本库不足的问题，为开发应用于特定场景的基于人工智能图像火灾探测算法提供支撑。

关键词: 生成对抗网络；数据集增强；火灾检测；样本合成

Abstract: In order to solve the problem of insufficient sample library of fire images in large-span spatial places, which leads to low accuracy of fire detection, a method of enhancing synthesized samples of large-span spatial fire is proposed. This method is based on an improved CycleGAN network, which constructs a flame block generation network and a fire image generation network. The flame block generation network generates flame block images, and the fire image generation network integrates flame blocks into designated areas of the engineering scene to generate fire samples that can be used for object detection models. The results show that the synthesized samples generated by the model improved the performance of the object detection model significantly, with a increase of 25.11% in mAP@0.5, verifying the effectiveness of the data augmentation method. Therefore, by generating high-quality and diverse large-span fire images, the problem of insufficient fire sample libraries has been effectively alleviated, providing strong support for the development of artificial intelligence based image fire detection algorithms for specific scenarios.

Key words: generative adversarial network; dataset augmentation; fire detection; sample synthesis