摘要
在森林火灾监测系统中,为了能够实时高效地定位森林视频监测图像中可疑的火灾着火点的具体位置,研究了将CMYK彩色模型下的火焰的彩色特征与遗传算法求最优适应度函数相结合,通过合理配置检测窗口邻域中颜色的M、Y、K各低阶矩特征在欧氏距离分类适应度函数中的因子,使森林监测图像的可疑的火灾点初定位控制在0.05s以内,然后通过火灾样本图像与初定位的可疑的火灾着火点的颜色互信息的计算和效验而给出准确的报警信号。实验结果表明,该算法具有良好的火焰区域定位效果及运算效率并易于移植于前端探测的嵌入式火灾探测系统。
In the monitor system of forest flame, in order to orientate efficiently the region of the flame area in a forest image, a combination algorithm of color property based on CMYK color space and genetic algorithm is proposed. M, Y, K elements are implements and configured as the adaptation function factors in Euclidean distance classification and neighborhood region~ s color lower order moments are used as the features and a accurate primary orientation of forest flame of a frame image in 0. 05 s is gained. It will do further verification to the primary orientation of forest flame by color mutual information calculation between key template image and searched window region image. The initial chrome distribution for a quicker orientation is optimized. The experiments show that the algorithm has a better orientation result and calculation efficiency and is easier to transfer the algo- rithm to embedded fire detecting system.
出处
《计算机工程与设计》
CSCD
北大核心
2012年第10期3903-3907,共5页
Computer Engineering and Design
关键词
图像匹配
森林火灾
彩色空间
遗传算法
互信息
image matching
forest flame
color space
genetic algorithm
mutual information
