基于多模光纤的多模干涉原理以及图像处理算法, 对多模光纤输出端面光斑图像相位谱的灰度共生矩阵特征值进行提取, 提出一种新型光斑图样处理算法, 并且将其应用于微位移传感。详细分析了以阶跃型多模光纤为基础的单模-多模光纤结构在两光纤径向产生微位移时, 位移量与端面光斑图像相位谱特征值之间的关系。对多种结构参数光斑图样的分析表明, 两光纤径向微位移与相关性及一致性两特征值成近似线性关系, 可以用光斑图像处理的方法进行微位移测量, 从而实现传感。实验结果表明与目前常用的计算光斑图像归一化强度内积处理算法相比, 多模光纤芯径较小时, 线性度、动态范围比较接近; 芯径较大时, 线性度提高的同时动态范围扩大约一倍。因此所提出的算法稳定性更好, 适用范围更广。

Based on the principle of multi-mode interference of multimode fiber and the algorithm of image processing, the characteristics of gray level co-occurrence matrix of phase spectrum of multi-mode fiber output speckle pattern were extracted, a new speckle pattern processing algorithm was proposed and applied to micro-displacement sensing. The relationship between the displacement and the characteristics of the phase spectrum of the output speckle pattern was analyzed in detail, when the single-mode-multimode fiber structure based on step-index multimode fiber generated micro-displacement in the radial direction of the two optical fibers. The analysis of speckle patterns of various structural parameters shows that, the approximation linear relationship between radial micro-displacement of the two optical fibers and the two characteristics including correlation and homogeneity can be obtained. Therefore, in order to achieve micro-displacement measurement, the method of image processing algorithm over speckle pattern can be used, making further efforts to realize optical fiber sensing. The experimental results show that compared with the commonly used algorithm of calculating normalized intensity inner product of the speckle pattern, when the core diameter of multi-mode fiber is small, the linearity and dynamic range are relatively close, when the core diameter is larger, the dynamic range is doubled while the linearity is improved. Therefore, the proposed algorithm has better stability and wider application range.