摘要
利用森林早期火灾图像中火焰的特点,对火灾进行识别,主要以火焰颜色分析、火焰面积增长特性、火焰尖角特性和火焰闪烁频率作为火灾发生的四个判据。同时,对烟雾图像进行分割、提取,并结合各个方向采集到的烟雾图像进行整合,得出火灾的蔓延方向估计。火灾识别与蔓延方向估计两者相结合,可有效判断森林早期火灾的发展形势,且有较高的稳定性和可靠性,具有一定的实际意义。
The characteristic of flame in early forest fire image is used to identify forest fire.And the criterions of fire recognition have been studied in detail.Especially four criterions are suggested,such as flame color analysis,flame area increasing,flame angle and flicker frequen-cy.Moreover,smoke images of different orientation have been also extracted and segmented to estimate the expanding orientation of fire.Considering both recognition and expanding orientation estimation is very important.Because it can indicate the situation of forest fire more effectively.In other words,the method is more stable and reliable.
作者
刘文礼
何鹏
于泓博
李会
LIU Wen-li,HE Peng,YU Hong-bo,LI Hui(Dept.of Communication,Qiqihar University,Qiqihar 161006,China)
出处
《电脑知识与技术》
2010年第5期3479-3480,共2页
Computer Knowledge and Technology
基金
黑龙江省教育厅科学技术研究项目(11541399)
关键词
火灾识别
闪烁频率
图像提取
方向估计
图像分割
fire recognition
flicker frequency
image extraction
orientation estimation
image segmentation