基于代数迭代法的激光成像目标重构方法  

Laser imaging target reconstruction method based on algebraic iteration method

作　　者：柳洁冰[1] 张文娟[1] LIU Jiebing;ZHANG Wenjuan(Zhangjiakou University,Zhangjiakou Hebei 075000,China)

机构地区：[1]张家口学院,河北张家口075000

出　　处：《激光杂志》2022年第12期98-102,共5页Laser Journal

基　　金：河北省自然科学基金项目(No.A2019209533)。

摘　　要：受到现实情况中各种客观条件的限制,获得的激光成像回波数据存在缺失现象,导致目标图像重构的精确度受到了极大干扰。针对上述问题,提出一种基于代数迭代法的激光成像目标重构方法。激光束从多个角度向目标发射激光束,并利用接收器采集激光回波数据。对采集到的激光回波数据实施预处理,包括去噪、数据修复两个部分。以预处理好的数据为基础,通过代数迭代法建立目标重构方程,并利用Kaczmarz松弛法实现迭代求解,完成激光成像目标重构。结果表明:所研究方法应用下,目标结构相似度要高于基于“边缘对折,单边滚动平滑”的重构方法、基于加权一阶类高斯匹配滤波算法的重构方法、基于卷积神经网络的重构方法的应用效果,说明利用所研究方法重构的目标图像更接近实际结果。Limited by various objective conditions in reality,the obtained laser imaging echo data is missing,which greatly interferes with the accuracy of target image reconstruction.To solve the above the above problems,a laser imaging target reconstruction method based on algebraic iteration method is proposed.The laser beam emits the laser beam to the target from multiple angles,and the receiver is used to collect the laser echo data.Pre-processing is carried out on the collected laser echo data,including two parts:denoising and data repair.Based on the preprocessed data,the target reconstruction equation is established by algebraic iteration method,and the iterative solution is realized by kaczmarz relaxation method to complete the laser imaging target reconstruction.The results show that under the application of the research method,the target structure similarity is higher than the reconstruction method based on"edge folding and unilateral rolling smoothing",the reconstruction method based on weighted first-order Gaussian like matched filter algorithm and the reconstruction method based on convolution neural network.It shows that the target image reconstructed by the research method is closer to the actual results.

关 键 词：代数迭代法 激光成像 预处理 目标重构 

分 类 号：TN209[电子电信—物理电子学]