计及偏移量的非线性光学显微成像误差校正方法  

Nonlinear optical microscopic imaging error correction method considering offset

作　　者：翟伟芳 冯娟 刘永立 ZHAI Weifang;FENG Juan;LIU Yongli(Baoding University of Technology,Baoding Hebei 071000,China)

机构地区：[1]保定理工学院,河北保定071000

出　　处：《激光杂志》2023年第12期235-239,共5页Laser Journal

基　　金：河北省省级科技计划软科学研究专项资助(No.21555401D)。

摘　　要：针对因非线性光学显微成像以客观像素特征为驱动源,一旦出现特征缺失,推断过程很容易形成误差,导致的光学显微图像失真现象严重、清晰度低问题,提出计及偏移量的非线性光学显微成像误差校正方法。利用数据驱动法全面、快速以及敏感地捕捉光学数据,划分光学显微图像的边缘和非边缘区域,设定标准滑动窗口,计算不同区域内像素点在该窗口内的像素灰度值,根据灰度值判定边缘点和非边缘点的误差形成概率。在此基础上,将误差校正看做一种角度或相位偏离补偿问题,设定偏离中心,计算误差概率较大的图像区域内各节点与中心间角度和相位偏移量,根据偏移量给出相应正向和负向补偿,完成有效误差校正。实验数据证明:所提方法误差校正精准度高,校正后成像失真、细节丢失以及分辨率低问题被解决,误差点的分布离散程度降低,算法整体实用性较强。Aiming at the problem of serious distortion and low definition of optical microscopic image caused by the fact that nonlinear optical microscopic imaging is driven by objective pixel characteristics,once the characteristics are missing,the inference process is easy to form errors,and the offset is taken into account,a nonlinear optical microscopic imaging error correction method is proposed.The data driven method is used to capture optical data comprehensively,quickly and sensitively,divide the edge and non edge areas of optical microscopic image,set a standard sliding window,calculate the pixel gray value of pixels in different areas within the window,and determine the error formation probability of edge points and non edge points according to the gray value.On this basis,error correction is considered as an angle or phase deviation compensation problem.By setting the deviation center,the angle and phase offset between each node and the center in the image area with a high probability of error are calculated,and corresponding positive and negative compensations are given based on the offset amount to complete effective error correction.Experimental data show that the proposed method has high accuracy of error correction,and the problems of imaging distortion,detail loss and low resolution after correction are solved.The distribution of error points is less Statistical dispersion,and the overall practicability of the algorithm is strong.

关 键 词：数据驱动 非线性光学显微成像 误差校正 相位偏移量 边缘点 

分 类 号：TN249[电子电信—物理电子学]