Use of Proper Orthogonal Decomposition Towards Time-resolved Image Analysis of Sprays
Use of Proper Orthogonal Decomposition Towards Time-resolved Image Analysis of Sprays
复制标题
使用适当的正交分解进行喷雾的时间分辨图像分析
DOI：
--
复制
发表时间：
2011
期刊：
影响因子：
--
通讯作者：
B. Eilers
中科院分区：
文献类型：
--
作者：
V. Narayanan;M. Lightfoot;S. Schumaker;S. Danczyk;B. Eilers
摘要：本文对气心旋流同轴（GCSC）喷油器喷射出的光学致密喷雾的高速电影进行了图像分析，以确定喷雾特性。作为其在分析高光密度喷雾中的功效的测试，在图像分析之前将本征正交分解（POD）应用于图像序列。喷雾形态，包括喷雾边界，宽度，角度和边界振荡频率的结果进行了比较，没有应用POD。最主要的模态重建（2和5之间）的喷雾被发现是足以定量描述的喷雾参数，而喷雾边界的运动频率是很好地预测的前两个模式。尽管是计算密集型的，但是使用POD对喷雾图像进行预处理具有以下优点：a）减少喷雾边界中的液滴噪声，从而导致用于边缘检测的更平滑的边界，以及B）处理表现不佳的喷雾的形态参数的能力。
Abstract : High-speed movies of optically dense sprays exiting a Gas-Centered Swirl Coaxial (GCSC) injector are subjected to image analysis to determine spray characteristics. As a test for its efficacy in analysis of high-optical-density sprays, Proper Orthogonal Decomposition (POD) was applied to the image sequence prior to image analysis. Results of spray morphology including spray boundary, widths, angles and boundary oscillation frequencies, are compared with and without the application of POD. The most dominant modal reconstructions (between 2 and 5) of the spray are found to be sufficient to quantitatively describe the spray parameters, while the frequency of movement of the spray boundary is well predicted by the first two modes alone. Despite being computationally intensive, preprocessing of spray images using POD has advantages in - a) reduction of droplet noise in the spray boundaries thereby resulting in a smoother boundary for edge detection, and b) ability to process morphological parameters of ill-behaved sprays.