Digital holographic measurement has the capability to measure features of intricate surfaces and has in turn emerged as a favorite technique in optical engineering for interferometry-based measurement. With time, scholars have unearthed that phase unwrapping is critical step for accurate reconstruction when utilizing this technique. The phases of the complex amplitudes computed from interferometric holograms are usually disturbed by speckle noise, therefore making it very cumbersome to obtain reliable unwrapping results. Over the years, researchers have developed unwrapping algorithms, unfortunately, these methods have been proven to be sensitive to spikes and prone to generating unreliable results in practice. Generally, current unwrapping algorithms are either inefficient or weak for dealing with noisy phase maps, thereby raising the need for further research purposed to resolve this shortcoming and provide a lasting solution.
Recently, a team of researchers led by professor Xiangchao Zhang from the Shanghai Engineering Research Center for Ultra-Precision Optical Manufacturing at Fudan University in China developed a robust phase unwrapping algorithm. Specifically, they employed the non-subsampled contourlet transform for accurate segmentation due to its shift-invariance and sparse representation of directional edge features. Moreover, they aspired to demonstrate a superior and universal algorithm to the digital holographic interferometry that could be proven both numerically and experimentally. Their work is currently published in the research journal, Optics Communications.
The technique employed commenced by exploiting the multiscale and directional decomposition of the non-subsampled contourlet transform to enhance the boundary between adjacent phase levels. Next, the researchers segmented the wrapped phase map into several regions corresponding to different phase levels. Lastly, an unwrapped phase map was attained by elevating the phases of a whole segment instead of individual pixels to avoid unwrapping errors caused by local spikes.
The authors observed that region segmentation with the non-subsampled contour-let transform was capable of accurately extracting region boundaries in the transform domain. In addition, the researchers noted that the phase-level elevation was conducted on the whole phase levels instead of on single pixels, hence the local errors could not affect the entire phase map even under severe noise pollution.
Xiangchao Zhang and research team truly advanced a novel unwrapping algorithm consisting of region segmentation, phase denoising, and phase-level elevation. X. Zhang and colleagues mainly observed that their technique outperformed other unwrapping algorithms operated in the spatial domain. More so, they recorded that their technique has the capability to deal with seriously polluted interferograms and complex wave-fronts with high fringe densities, especially in the digital holographic interferometry. Altogether, the proposed technique has potential to be extended to unwrapping other interferometric phase maps such as the biomedical samples such as; living cells and synthetic joints.
Xiaolei Zhang, Xiangchao Zhang, Min Xu, Hao Zhang, Xiangqian Jiang. Phase unwrapping in digital holography based on non-subsampled contourlet transform. Optics Communications, volume 407 (2018) page 367–374..Go To Optics Communications