基于改进Canny算子的气液两相泡状流图像分割算法  

Bubble Image Segmentation of Gas/Liquid Two-Phase Flow Based on Improved Canny Operator

作　　者：张问银[1,2] 金宁德[1] 

机构地区：[1]天津大学电气与自动化工程学院,天津300072 [2]临沂师范学院信息学院,山东临沂276005

出　　处：《计算机工程与科学》2009年第8期137-139,共3页Computer Engineering & Science

基　　金：博士后基金资助项目(20070410757);国家自然基金资助项目(50674070)

摘　　要：图像分割是多相流高速视频分析的前提和基础。本文基于改进的Canny算子给出一种泡状气液两相流的图像分割方法。该方法充分考虑原始图像的灰度分布信息,利用高斯Sigma平滑滤波器有效地保持了边缘及消除了噪声,通过梯度的非极大值抑制和滞后阈值化处理后,获得了清晰的气液两相流泡状流图像边缘,为泡状流流动特性分析奠定了分析基础。Image segmentation is the foundation of multi-flow high-speed video analysis. Based on an improved Canny operator, this paper provides a method for bubble gas/liquid two-phase flow image segmentation, which takes more gray distribution information into account, uses the Gauss Sigma smooth filter to maintain edges and eliminate noise. After the non-Max control of grads and post-threshold processing, a clear bubble edge image is obtained. The method is much effective in the analysis and recognition of bubble multi-flow.

关 键 词：气液两相流 图像分割 CANNY算子 高斯平滑滤波 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]