THzViSAR极坐标格式算法的几何失真校正  

Geometric Distortion Correction for THzViSAR Polar Format Algorithm

作　　者：尚双丽 李鑫宇 陈燕燕 王亚哲 郑其斌 

机构地区：[1]上海理工大学健康科学与工程学院,上海 [2]上海无线电设备研究所,上海 [3]杭州乾晶半导体有限公司,浙江 杭州

出　　处：《建模与仿真》2023年第2期1612-1622,共11页Modeling and Simulation

摘　　要：极坐标格式算法(PFA)基于平面波前假设对差分距离进行二阶近似,导致线性和二次相位误差的存在,并分别导致几何失真和图像散焦。太赫兹视频SAR (THzViSAR)的特点在于有效场景范围大、成像区域小,因此可以忽略二次相位误差(QPE),而线性相位误差(LPE)只与成像几何有关,不可忽略。考虑到不同帧坐标下的图像旋转,本文对偏移后的目标位置进行了公式推导,并且采用图像域重采样的方法进行几何失真校正,同时实现图像旋转。最后通过仿真数据处理验证了本文所述的校正方法能够有效地改善PFA图像的失真及旋转情况。The polar format algorithm (PFA) makes a second-order approximation of the differential range based on the planar wavefront assumption. This leads to the existence of linear and quadratic phase errors, respectively resulting in geometric distortion and defocusing. Terahertz video synthetic ap-erture radar (THzViSAR) is characterized by the small region and the large permissible region of Quadratic Phase Error (QPE), so QPE can be ignored. However, Linear Phase Error (LPE) is only re-lated to imaging geometry and cannot be ignored. Considering the image rotation in different frame coordinates, this paper deduces the formula for the offset position, and adopts the image domain resampling for geometric distortion correction and rotation. Finally, simulation experiments verify that the correction method can effectively improve the distortion and rotation of PFA images.

关 键 词：图像旋转 几何失真 重采样 二阶近似 公式推导 太赫兹 成像区域 SAR 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]