摘要
基于燃气燃烧试验台,利用工业电荷耦合器件(CCD)相机获取不同工况下的扩散火焰与预混火焰图像,通过选择合适的图像处理算法获得表征火焰形状、位置和亮度等特征的6个特征量。以不同类型火焰的6个特征量数据为训练样本,通过支持向量机的分类方法对程序进行分类训练,并在燃烧试验台上进行实时监测和稳定性判断。结果表明:各类型火焰检测正确率在99%以上。
Based on the gas combustion test platform, a large number of diffusion and premixed flame images were taken by industrial CCD camera under different conditions, following which, six characteristic variables that representing the shape, position and brightness of the gas flames were obtained by choosing and adopting appropriate image processing algorithm. Taking the six characteristic data of different kinds of flames as the training samples, the classification program was trained by support vector machine, and subsequently the real-time monitoring and stability evaluation were conducted on the combustion test platform. Results show that via the method proposed, the detection accuracy may achieve 99% for different types of flames.
作者
田正林
余岳峰
朱小磊
王宇
张忠孝
TIAN Zhenglin;YU Yuefeng;ZHU Xiaolei;WANG Yu;ZHANG Zhongxiao(School of Mechanical Engineering,Shanghai Jiaotong University,Shanghai 200240,China)
出处
《动力工程学报》
CAS
CSCD
北大核心
2019年第10期811-817,共7页
Journal of Chinese Society of Power Engineering
基金
国家重点研发计划资助项目(2017YFF0209801)
关键词
燃气火焰
图像处理
实时监测
支持向量机
gas flame
image processing
real-time detection
support vector machine
