摘要
针对目前视频火灾检测误检率高以及计算量大而需要高性能硬件支持的工程应用需求,提出了采用改进的Surendra背景更新算法实时提取视频背景。采用迭代法实时求取当前的最佳阀值,以及Surendra背景更新算法获取背景,依据背景求得图像中的运动区域,形态学处理后通过火焰颜色统计模型判别是否是火灾。实验表明,文中采取的检测算法快速高效,具有良好的市场应用前景。
The current flame detection algorithm cannot meet the requirements of high detection rate and requires high performance hardware because of large calculation in engineering application. An improved Surendra background updating algorithm is proposed to extract video background in real time. Firstly,the best threshold is calculated in real time by the iterative method. Then update background is performed by the Surendra background updating algorithm to get the moving areas. Finally,the flame is detected by the statistical color model after morphological processing. Experiments show that the proposed detection algorithm is fast and efficient with good application prospect.
出处
《电子科技》
2016年第5期169-171,共3页
Electronic Science and Technology
