基于随机波数序列合成的激光三维干涉测量理论、系统及验证

Laser three-dimensional interferometry (LTI) can perspectively measure the volumetric displacement and strain field inside the resin matrix composite with high precision, scale-span and three-dimension, which will become one of the development future for experimental characterization. However, the depth resolution of LTI is as low as millimeter because of limited wavenumber tuning range, resulting in the restricted application. This project plans to design a laser wavenumber discontinuous tuning interferometry, whose random wavenumber series are synthesized so that the wavenumber tuning range is enlarged. Moreover, the demodulation model based on randomly sampling compressed sensing for interference signal is presented. It uses sparse optimization to reconstruct interference spectrum, resulting in the maximal removal of window function blurry convolution. By combining the hardware and theory, the depth resolution of LTI is up to the level of micrometer. The project offers a comprehensive solution of depth resolution in LTI, which suggests that the promising application of LTI is improved a lot, such as nondestructive examination of micron-size defects inside the resin matrix composite.

激光三维干涉测量可高精度、跨尺度、三维透视测量树脂基复合材料内部在不同载荷下的体位移场和应变场，是材料力学性能实验表征测试的未来发展方向之一。然而因半导体激光器的波数调谐范围有限，激光三维干涉的深度测量分辨率低，仅为亚毫米级别，无法表征多层树脂基复合材料内部层间间距为几十个微米的界面层力学性能，应用十分有限。本项目提出设计激光波数间断调谐干涉测量系统：合成半导体激光器的多个随机波数序列，扩展波数调谐范围；建立基于随机采样压缩感知的干涉信号解调模型：使用稀疏优化方法重构出干涉频谱，最大限度去除窗函数卷积模糊效应。通过结合硬件和理论方法，本项目将激光三维干涉的深度测量分辨率提高至微米级别，全面解决激光三维干涉的深度测量分辨率过低问题，大幅度提升其应用前景，如材料内部微米级缺陷无损检测等。

激光三维干涉可高精度透视测量树脂复合材料内部的体变形场，是未来材料力学性能表征的重要发展方向之一。然而受无跳模激光扫频带宽的限制，传统激光三维干涉的深度测量分辨率仅为亚毫米级，应用有限。针对上述问题，本项目从信息处理角度出发，提出基于激光三维干涉的超分辨深度测量方法,研制相应测量系统。研究内容包括：1) 提出激光跳模波数合成算法，实现波数跳模前后的高效合成，扩展单个无跳模激光器的扫频带宽，在硬件层面上提高深度测量分辨率；2) 利用干涉频谱稀疏性特点，构造激光三维干涉信号的盲分离模型，采用波数域总体最小二乘重构高分辨层析相位场，在算法层面上提高深度测量分辨率；3) 提出基于相位噪声数目优化的包裹相位合成理论，实现无需破坏x, y成像分辨率下的散斑噪声降噪，高分辨重构出材料内部的位移场；4) 搭建了激光波数间断调谐干涉测量系统，采用软硬协调一体化技术手段，将激光三维干涉的深度测量分辨率从原来亚毫米级提升至~10μm。研究成果将有助于解决激光三维干涉因硬件瓶颈限制导致的深度测量分辨率受限难题，为其在复合材料力学性能表征、微米级缺陷检测等应用提供技术支撑。