基于自注意力机制和CycleGAN的高分三号ScanSAR图像的扇贝效应抑制

高分三号卫星是我国首颗分辨率达到1 m的C波段 多极化合成孔径雷达(synthetic aperture radar,SAR) 卫星,其中扫描 式合成孔径雷达(scan synthetic aperture radar,ScanSAR)模式是高分三号卫星重要的工 作模式之一,由于该模式的工作机制导致生成的图像可能发生扇贝效应,一般呈现为明暗相 间的条纹。本文针对高分三号卫星ScanSAR模式下存在的扇贝效应,提出自注意力机制与循 环一致对抗网络(cycle-consistent adversarial networks,CycleGAN)结合的模型对Scan S AR图像进行处理,从而抑制扇贝效应产生的条纹现象。本文所示方法与传统扇贝效应抑制方 法和深度学习相关算法进行比较,并通过亮度均值、平均梯度等指标进行分析。实验结果表 明,本文方法可以对高分三号ScanSAR图像存在的扇贝效应进行较好的处理,有效抑制图像 的条纹现象,使得图像质量得到提升,具有较大的实用意义。

Descalloping of GF-3 ScanSAR image based on self-attention mechanism and CycleGAN

GF-3 satellite is the first C-band multi-polarimetric synthetic apertu re radar satellite with a space resolution up to 1 m in China,in which scan synthetic ape rture radar (ScanSAR) is one of the important mode of GF-3.The working mechani s m of this mode results in the phenomena of serious nonuniformity,generally show ing bright and dark stripes,also known as scalloping.In view of scalloping in ScanSAR mode of GF-3,this paper proposes a model combining self-attention mec ha nism and cycle-consistent adversarial networks (CycleGAN),so as to perform des c alloping.The proposed descalloping method is compared with traditional descalloping methods a nd deep learning related algorithms,and analyzed by indicators such as brightne ss average and average gradient.The experimental results demonstrate that the p roposed method in this paper can better complete descalloping in the GF-3 ScanSAR image, effectively suppress the stripes phenomenon of the image,and improve the image quality,which is of great practical significance.