论文简介 |
Abstract. The adaptive spatial filtering method is commonly adopted to extract the þ1 term spectrum in digital
holography for real-time dynamic analysis. However, the typical filtering method is not satisfactory for automatic
analysis, because the reset of the filtering window is needed to extract the area of the þ1 term spectrum.
Therefore, an adaptive spatial filtering method based on region growing and the characteristic of the spectrum
separation is proposed. Its filtering window is automatically formed by region growing. The key parameters,
including threshold and seed point, are set by the intensity distribution of the hologram spectrum. Then the
adaptive filtering extracting the þ1 term spectrum is realized by multiplying the hologram spectrum by the filtering
window. Compared to the typical filtering method, the experimental results of a microhole array and a phase step
show that the proposed method has better adaptability and a higher precision. Moreover, the applicability of this
method for different uses is also demonstrated by experiments with a microhole array and a phase step. © 2015
Society of Photo-Optical Instrumentation Engineers (SPIE) [DOI: 10.1117/1.OE.54.3.031103]
Keywords: digital holography; spatial filtering; region growing.
Paper 140931SS received Jun. 10, 2014; revised manuscript received Aug. 3, 2014; accepted for publication Aug. 7, 2014; published
online Oct. 14, 2014.
1 Introduction
Digital holography is widely used in the fields of microelectromechanical
systems measurement,1–6 biological cell monitoring,
7–12 microscopic particles tracking,13,14 and ladar
observation systems15,16 due to its significant advantages
over other conventional three-dimensional (3-D) measurements
such as high resolution,17,18 real time,8,9,13,14,19 and
nonmechanical scanning.20–22 In order to get a good quality
reconstructed image, the zero order of diffraction and the real
image must be eliminated during digital reconstruction.23–25
Phase-shifting digital holography,26,27 off-axis digital holography,
25,28,29 and parallel phase-shifting digital holography30,31
are commonly used to eliminate the two undesired diffraction
images. The first one is not suitable for real-time
dynamic analysis due to the multiple holograms required.
The latter one is fit for real-time dynamic analysis but its
spatial resolution is low. However, off-axis digital holography
is not only fit for real-time dynamic analysis, but also
has high spatial resolution. So, off-axis digital holography is
widely employed in real-time dynamic analysis. However,
proper spatial filtering is necessary to determine the þ1 term
spectrum which will affect the quality of the reconstructed
image.
To realize real-time analysis, it is very important to automatically
and precisely find out the distribution of the þ1
term spectrum. There are two methods of spatial filtering
used to achieve the þ1 term spectrum, including manual spatial
filtering17,32 and automatic spatial filtering.33,34 Manual
spatial filtering is time consuming and depends on subjective
judgment. Therefore, it is not suitable for real time and
dynamic analysis. However, in typical automatic spatial filtering,
a key step is to set a regular filtering window (an elliptic
filtering window or a rectangular filtering window) for
eliminating the zero and −1 term spectra (representing the
zero order of diffraction and the |