摘要
提出了一种计算机自动准确识别流型的新方法.该方法利用高速电荷耦合器件(CCD)获取管道内气液两相流的流动图像,经过图像处理后,提取二值化图像中气泡的面积、宽度、高度和中心坐标,结合模糊推理方法对流动图像进行流型识别.实验结果表明,该方法能够自动有效地识别水平管道内的多种基本流型,识别准确率分别为:气泡流93.3%,塞状流85.3%,分层流97.3%,弹状流98.6%,雾环状流92.7%,识别图像速度约为22帧/s.
A new method was developed for identifying flow pattern automatically and accurately. Flow images were captured by a high speed charge couple device （CCD）. The characteristics of binary images, such as area, width, height and center coordinates of bubbles were obtained using image processing techniques. Then, a fuzzy reasoning method was used to identify the flow pattern. Experimental results showed that the method is effective for identifying the basic flow patterns in a horizontal pipe. The identification accuracy was as follows： bubbly flow, 93.3%; plug flow, 85.3%; stratified or wavy flow, 97.3%; slug flow, 98.6%; annular flow, 92.7%. An estimation of the process time was 22 frame/s.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2005年第8期1128-1131,共4页
Journal of Zhejiang University：Engineering Science
关键词
两相流
流型识别
图像处理
模糊推理
two-phase flow
flow pattern identification
image processing
fuzzy reasoning
