基于PCNN分割的火焰图像检测技术  

Realization of fire detection technology based on PCNN

作　　者：厉谨[1,2] 王慧琴[1] 袁洪琳[2] 李力[2] 

机构地区：[1]西安建筑科技大学信息与控制学院,陕西西安710055 [2]西安工程大学电子信息学院,陕西西安710048

出　　处：《西安工程大学学报》2010年第2期217-221,共5页Journal of Xi’an Polytechnic University

摘　　要：基于PCNN分割的火灾探测方法,提出了基于计算机视觉和数字图像处理技术的火焰特征计算方法,根据色度学原理对火焰图像进行阈值分割、去除噪声、寻找火焰边界、计算周长和面积、找出火焰特征坐标等处理.从理论分析及实验结果中得出了该方法具有精度高、适应性强、易于计算等特点,可以抵抗常见的干扰对系统的影响,可以提高火灾探测的准确率减少漏判和误判.Based on PCNN image segmentation,the method of fire detection technology was discussed.The calculation method of flame characters were presented based on computer vision and digital image processing.According to chromaticity principle,fire image was treated including segmentation of threshold value,removing noise,seaching flame boundary,calculating perimeter and area,finding out flame feature coordinate.The experimental results show that the method has the features of high precision,good adaptability and easy to realize for computer.At the same time,the effect of common interference on the system is resisted and the accuracy rate of fire detection is improved and the missing and misjudgement is reduced.

关 键 词：火焰图像 区域周长 视频 区域分割 PCNN 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]