摘要
针对远距离探测成像系统受湍流杂波影响导致目标图像出现低信噪比、模糊和几何畸变等退化降质情况,提出一种基于空频域先验的湍流降质目标增强方法。所提方法首先利用潜在低秩分解(Latent Low-Rank Decomposition, LatLRD)将单帧湍流降质目标图像分解为代表全局结构的低秩分量、局部结构的纹理分量和高频噪声分量。LatLRD分解滤出目标噪声后,接着对低秩分量和纹理分量的小波频率域空间分别采用基于区域分割先验和基于像素导向的方式抽取小波系数,最后通过小波逆变换实现湍流模糊降质目标增强。针对典型的湍流退化降质目标增强进行对比实验,结果表明所提算法能够平均提高湍流降质目标信噪比约8dB,同时所提方法只需单帧图像信息便能有效地将目标空频域先验嵌入湍流目标增强处理。
In view of the degradation and quality reduction of target images in long-distance detection imaging systems,such as low signal-to-noise ratio,blurring,and geometric distortion,which are caused by the influence of turbulent clutter,a turbulent-degraded target enhancement method based on spatio-frequency domain prior is pro-posed.The proposed method first utilizes the Latent Low Rank Decomposition(LatLRD)method to decompose a sin-gle frame turbulence-degraded target image into low rank component with global structural features,texture component with local structural features,and high-frequency noise component.After filtering out target noise through LatLRD de-composition,wavelet coefficients are extracted for low rank and texture components using region segmentation prior and pixel-oriented methods,respectively.Finally,turbulence distortion and blur are removed through inverse wavelet transform.A comparative experiment was carried out for the enhancement of typical turbulence-degraded targets.The results show that the proposed algorithm can increase the signal-to-noise ratio of turbulence-degraded targets by a-bout 8 dB on average.At the same time,the proposed method only needs the information of a single-frame image to effectively embed the target spatial-frequency domain prior into the turbulence target enhancement processing.
作者
徐兴贵
李红
郭沫海
任维贺
XU Xing-gui;LI Hong;GUO Mo-hai;REN Wei-he(School of Information Science and Engineering,Yunnan University of Finance and Economics,Kunming Yunnan 650051,China;Eighth Brigade of the Second Mobile Corps of the People's Armed Police Force,Mengzi Yunnan 661100,China;The Institute of Beijing Space Electromechanical Research,Beijing 100039,China)
出处
《计算机仿真》
2025年第9期362-369,396,共9页
Computer Simulation
基金
国家自然科学基金(62161051,62202415)
兴滇英才支持计划(20220632)。
关键词
目标增强
湍流杂波
对偶小波
低秩分解
Target enhancement
Turbulent clutter
Dual wavelet
Low-rank decomposition
