基于自适应短时傅里叶变换的结构光场单帧三维成像算法  

Structured Light Field Single-Shot 3D Imaging Algorithm Based on Adaptive Short-Term Fourier Transform

作　　者：贾尚恩 安俊睿 周光泉[1] 何思渊[1] 周平[1] Jia Shangen;An Junrui;Zhou Guangquan;He Siyuan;Zhou Ping(School of Biological Science&Medical Engineering,Southeast University,Nanjing 210096,Jiangsu,China)

机构地区：[1]东南大学生物科学与医学工程学院,江苏南京210096

出　　处：《激光与光电子学进展》2023年第8期324-329,共6页Laser & Optoelectronics Progress

基　　金：国家自然科学基金(52071075)。

摘　　要：单帧结构光图像的相位恢复、解相是三维成像领域的重要问题。理论上,在选择合适参数的条件下,短时傅里叶变换(STFT)可以有效地完成相位恢复,但“合适参数”这一前提条件往往因缺少先验信息而难以实现。在传统STFT的基础上,采用结构光场系统充分利用光场成像的特性,提出一种基于自适应STFT的结构光场单帧三维成像算法。所提算法将结构光场图像从纹理、瞬时频率和深度估计信息等3个角度进行融合,将光场图像分割为满足STFT约束的若干区域,对每个区域依据瞬时频率估计对应的STFT的参数,进而恢复相对相位。随后,采用结构光场的解相方法实现三维成像。实验结果表明,所提自适应STFT算法能够有效完成被测物体的三维成像。The phase recovery and dephasing of single-frame structured light images are crucial in the field of threedimensional(3D) imaging.Theoretically,short-time Fourier transform(STFT) can efficiently complete phase recovery under the condition of selecting appropriate parameters;however,the prerequisite of “appropriate parameters” is often challenging to achieve due to the lack of prior information.Herein,a single-frame 3D imaging algorithm of structured light field based on adaptive STFT is proposed using a structured light field system to take full advantage of the characteristics of light field imaging.The suggested technique fuses the structured light field image from three angles of texture,instantaneous frequency,and depth estimation information and further divides the light field image into several regions that meet the STFT constraint.The associated STFT parameters are estimated for each zone using the instantaneous frequency,and then the relative phase is recovered.Next,the phase separation method of the structured light field is used to achieve 3D imaging.The experimental results show that the proposed adaptive STFT algorithm can successfully accomplish the 3D imaging of the measured object.

关 键 词：成像系统 光场 结构光 相位调制 三维成像 表面测量 轮廓 

分 类 号：O439[机械工程—光学工程]