摘要
从火焰的颜色特征出发,综合考虑已有颜色模型的优缺点,提出一种改进的基于颜色模型的火灾检测方法。首先结合HSI颜色模型和YCbCr颜色模型对火灾图像进行预处理提取出可能火焰区域,然后在HSI空间上采用颜色空间距离法去除噪声。实验结果表明,该方法提高了火灾图像识别的准确度,并能在一定程度上消除干扰,具有较好的适用性。
Starting from the features of flame color, considering the advantages and disadvantages of existing color model, an improved fire detection method is proposed based on the color model. Firstly, HSI color model and YCbCr color model are combined to preprocess fire image and extract fire area, then, the color space distance method is used for removing noise in the HSI space. The experimental results show that this method increases the accuracy of fire image recognition, and ＇also eliminates interference to a certain extent, and has better applicability.
出处
《电视技术》
北大核心
2012年第19期173-176,185,共5页
Video Engineering
基金
陕西省科技攻关计划项目(2011k06-32)
中央高校基本科研业务费专项资金项目(CHD2011TD012)
长安大学科技创新重点项目
陕西省道路交通智能检测与装备工程技术研究中心开放基金项目
关键词
火灾检测
颜色模型
图像识别
fire detection
color model
image recognition
