Xu Wenjun, Tang Chen, Su Yonggang, Li Biyuan, Lei Zhenkun
Appl Opt. 2018 Feb 1;57(4):861-871. doi: 10.1364/AO.57.000861.
In this paper, we propose an image decomposition model Shearlet-Hilbert-L with better performance for denoising in electronic speckle pattern interferometry (ESPI) fringe patterns. In our model, the low-density fringes, high-density fringes, and noise are, respectively, described by shearlet smoothness spaces, adaptive Hilbert space, and L space and processed individually. Because the shearlet transform has superior directional sensitivity, our proposed Shearlet-Hilbert-L model achieves commendable filtering results for various types of ESPI fringe patterns, including uniform density fringe patterns, moderately variable density fringe patterns, and greatly variable density fringe patterns. We evaluate the performance of our proposed Shearlet-Hilbert-L model via application to two computer-simulated and nine experimentally obtained ESPI fringe patterns with various densities and poor quality. Furthermore, we compare our proposed model with windowed Fourier filtering and coherence-enhancing diffusion, both of which are the state-of-the-art methods for ESPI fringe patterns denoising in transform domain and spatial domain, respectively. We also compare our proposed model with the previous image decomposition model BL-Hilbert-L.
在本文中，我们提出了一种图像分解模型Shearlet-Hilbert-L，它在电子散斑图案干涉术（ESPI）条纹图案去噪方面具有更好的性能。在我们的模型中，低密度条纹、高密度条纹和噪声分别由剪切波平滑空间、自适应希尔伯特空间和L空间描述，并分别进行处理。由于剪切波变换具有卓越的方向敏感性，我们提出的Shearlet-Hilbert-L模型对各种类型的ESPI条纹图案，包括均匀密度条纹图案、中等密度变化条纹图案和密度变化极大的条纹图案，都取得了值得称赞的滤波效果。我们通过将其应用于两个计算机模拟的以及九个实验获得的具有各种密度和低质量的ESPI条纹图案，来评估我们提出的Shearlet-Hilbert-L模型的性能。此外，我们将我们提出的模型与窗口傅里叶滤波和相干增强扩散进行比较，这两种方法分别是变换域和空间域中ESPI条纹图案去噪的最先进方法。我们还将我们提出的模型与之前的图像分解模型BL-Hilbert-L进行比较。
