基于经验模态分解的激光散斑噪声抑制方法  

Laser speckle noise suppression method based on empirical mode decomposition

作　　者：詹晓江 甘楚立 丁毅 胡轶 许彬 习江涛 邓定南 ZHAN Xiao-jiang;GAN Chu-li;DING Yi;HU Yi;XU Bin;XI Jiang-tao;DENG Ding-nan(Faculty of Intelligent Manufacturing,Wuyi University,Jiangmen 529020,China;School of Electrical,Computer and Telecommunications Engineering,University of Wollongong,Wollongong 2522,Australia;School of Physics and Electronic Engineering,Jiaying University,Meizhou 514015,China)

机构地区：[1]五邑大学智能制造学部,广东江门529020 [2]伍伦贡大学电气计算机与通信工程学院,澳大利亚伍伦贡2522 [3]嘉应学院物理与电子工程学院,广东梅州514015

出　　处：《液晶与显示》2023年第4期495-506,共12页Chinese Journal of Liquid Crystals and Displays

基　　金：国家自然科学基金面上项目(No.62075168);广东省普通高校创新团队项目(No.2021KCXTD051);五邑大学港澳联合研发基金项目(No.2021WGALH17)。

摘　　要：为抑制激光散斑噪声对定量相位重建精度的影响,本文基于经验模态分解提出了一种激光散斑噪声抑制方法。该方法只需采集一张全息图像且无需添加额外硬件设备,避免了寻找最优处理参数的复杂过程。首先,采用经验模态分解突出图像细节信息。其次,使用边缘检测克服各向异性扩散方程中梯度算子抗噪声能力不强、不能识别伪边缘等问题,对细节突出的图像进行边缘检测。最后,由高精度的边缘检测结果引导扩散去噪过程达到更好的散斑抑制效果。实验结果表明,经过本文方法处理后,重建相位的结构相似性提高了12.9000%,边缘保持指数提高了14.3861%,散斑抑制指数降低了8.1299%,并且相位截面曲线与原始相位最接近。本文所提出的方法不仅具备更好的去噪效果,而且更有效地保留了细节信息,相位重建的精度更高。In order to suppress the influence of laser speckle noise on the accuracy of quantitative phase reconstruction,a laser speckle noise suppression method based on empirical mode decomposition is proposed in this paper.The method only needs to acquire a holographic image without adding additional hardware equipment and avoids the complicated process of finding optimal processing parameters.Firstly,empirical mode decomposition is used to highlight image details.Secondly,edge detection is used to overcome the problems of poor anti-noise capability and the inability to identify false edges of gradient operators in the anisotropic diffusion equation,and edge detection is performed on images with prominent details.Finally,the diffusion denoising process is guided by the high-precision edge detection results to achieve better speckle suppression.The experimental results show that after processing by the method in this paper,the structural similarity of the reconstructed phase is increased by 12.9000%,the edge preservation index is increased by 14.3861%,the speckle suppression index is reduced by 8.1299%,and the phase cross-section curve is the closest to the original phase.The method proposed in this paper not only has a better denoising effect but also preserves the detailed information more effectively,and the accuracy of phase reconstruction is higher.

关 键 词：散斑噪声抑制 经验模态分解 边缘检测 各向异性扩散方程 

分 类 号：O438.1[机械工程—光学工程] TP391.4[理学—光学]