Fourier-based layout for grating function structure in spatial filtering velocimetry
Fourier-based layout for grating function structure in spatial filtering velocimetry
复制标题
DOI：
10.1088/1361-6501/aa6059
复制
发表时间：
2017-04
影响因子：
2.4
通讯作者：
M. Schaeper;N. Damaschke
中科院分区：
文献类型：
--
作者：
M. Schaeper;N. Damaschke
光学空间滤波测速技术（SFV）已经被用于速度测量几十年了。自20世纪90年代以来，电荷耦合器件（CCD）线传感器已被用于实现空间滤波系统的光栅功能，使用一个专门的时钟制度的固有实现。另一种方法是通过利用具有软件实现的光栅功能的阵列检测器（CCD或CMOS）来实现光学SFV系统，特别是用于二维速度测量。为观察到的场景选择合适的光栅函数可能是使用SFV时的障碍，并且依赖于用户的经验。考虑到这一点，本文概述了如何组装光学空间滤波系统。在对空间滤波系统中信号产生的一般描述之后，提出了一种利用傅里叶分析识别匹配谐波光栅函数的简单方法。这种方法对于观察到的具有周期性结构图案的场景具有特别的优点，当结合固定光栅功能使用SFV时，这是有问题的。谐波光栅函数的匹配周期可以被发现为成像场景的谱密度分布中的峰值。一旦找到匹配的光栅函数，就可以用SFV进行信号处理，这比计算全帧的互相关更简单，并且适合于实时应用。然后讨论了基于阵列探测器的空间滤波测速仪的布局标准。
Optical spatial filtering velocimetry (SFV) has been used for several decades for velocity measurements. Since the 1990s, charge-coupled device (CCD) line sensors have been used for the realization of spatial filtering systems by the inherent implementation of grating functions using a specialized clock regime. Another approach is the realization of optical SFV systems by utilizing array detectors (CCD or CMOS) with software-implemented grating functions, especially for two-dimensional velocity measurements. Choosing a suitable grating function for the observed scene can be an obstacle when using SFV, and relies on the experience of the user. With this in mind, this contribution presents an overview of how to assemble an optical spatial filtering system. After a general description of signal generation in spatial filtering systems, a straightforward approach to identifying matching harmonic grating functions by using Fourier analysis is presented. This approach has particular advantages for observed scenes with a periodically structured pattern, which were problematic when using SFV in connection with a fixed grating function. Matching periods of harmonic grating functions can be found as peaks in the spectral density distribution of the imaged scene. Once a matching grating function has been found, the signal processing can be made with SFV, which is simpler than calculating the cross-correlation of full frames and is suitable for real-time application. Criteria for the layout of an array-detector-based spatial filtering velocimeter are then discussed.