火灾图像分割与目标识别方法研究.pdfVIP

机械工程师

MECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEERMECHANICALENGINEER

火灾图像分割与目标识别方法研究

1a,1b21a,1b

赵汝海,王宏亮,朱广

（1.安徽建筑大学a.机械与电气工程学院；b.建筑机械故障诊断与预警技术重点实验室，合肥230601；2.安徽省淮南第一中学，

安徽淮南232001）

火灾现场环境复杂，获取的图像对比度较低，边缘模糊，使得火灾图像分割和火焰区域识别困难。为此，提出了一

摘要：

种基于色彩空间的火灾图像分割与多特征量融合的火灾图像识别方法。首先通过颜色特征找出可疑的火焰发生区域，然后

通过色彩空间的H、S、I特征的线性组合对可疑区域进行分割，接着分析和提取可疑区域的火焰特征，输入提取的火焰颜色、

纹理和形状特征,利用BP神经网络建立火灾识别模型，并以光、气焰和烛焰为干扰，通过神经网络的训练，最终分类和输出。

仿真结果表明，火灾识别精度高，鲁棒性好。

图像分割；特征提取；火灾识别

关