Abstract Details
  1. Home
  2. Abstract List
  3. Abstract Details
160 / 2023-07-05 20:00:36
Research on Fire identification Method based on multi-feature matching
Fire identification；Adaptive median filtering；Flame outline；Color model
2. Mine ventilation, fire, water damage prevention
Draft Accepted
Chao Wang / Heilongjiang University of Science and Technology
斌 沈 / 黑龙江科技大学
针对传统火灾识别算法降噪保护细节困难、实时性能和抗干扰能力差等问题，提出一种基于多特征匹配的火灾识别方法。首先，采用优化改进的自适应中值滤波算法检测动态火焰区域;然后，通过分离颜色通道建立颜色模型，以进一步区分火焰区域。最后，采用帧差法对火焰轮廓进行分割提取，得到火焰实施面积。实验结果表明，改进的自适应中值滤波能够更好地处理图形细节，与颜色模型和帧差法相结合时，能够更准确、实时地判断和提取火焰，具有较强的抗干扰能力。该算法系统的实验检测率达到92.7%，高于目前通用水平的90%。与传统的中值滤波算法和传统的混合高斯算法相比，它可以更好地保护图像细节，避免误报和漏检。

 
Important Date

  • Conference Date

    Aug 18

    2023

    to

    Aug 20

    2023

  • Jul 07 2023

    Draft paper submission deadline

  • Aug 20 2023

    Registration deadline

Sponsored By
International Committee of Mine Safety Science and Engineering
Organized By
Heilongjiang University of Science and Technology
Contact Information
Login to view full contact details
Previous Conferences