摘要
针对传统火灾探测技术存在的不稳定、误判率高等缺点,通过分析室内火灾图像与常见干扰光源图像的特点,提出一种基于人工神经网络的火焰图像检测技术。对火焰图像的基本特性进行分析,利用火焰图像序列的面积重叠率和中心相对移动率以及颜色等信息,结合实现学习向量量化(LVQ)神经网络融合技术,对视频序列图像中火焰的自动检测。仿真试验结果表明,基于LVQ神经网络的信息融合算法的网络收敛速度较快,有较高的火灾火焰识别准确率。
Considering the instability and high rate of erroneous recognition with traditional methods for fire detection,through analyzing the characteristics of fire image and the familiar interference light source,a flame recognition algorithm is proposed based on LVQ neural network.The basic characteristics of the flame image are analyzed.Some information of the flame image sequences,such as the center area of overlap,center relative mobility and color is applied,and the LVQ network convergence technology is combined to achieve automatic detection of the flame in the video image sequences.The simulation experiments show that LVQ neural network fire detection algorithm has a relatively high convergence rate and correct recognition rate.
出处
《中国安全科学学报》
CAS
CSCD
北大核心
2011年第6期60-64,共5页
China Safety Science Journal