基于幂次变换的SAR图像Otsu分割法

通过分析合成孔径雷达(synthetic aperture radar, SAR)目标切片图像的散射特性,提出一种适用于SAR目标识别的目标切片图像分割算法。算法首先对SAR图像做相干斑滤波,通过邻域平滑处理,提高背景区域和目标区域像素幅值一致性。然后自适应地选取变换幂次,对滤波后的SAR图像做幂次变换,以进一步增强目标区域像素幅值一致性。最后直接利用一维Otsu法对变换后的图像进行分割处理。实验表明,该算法对不同散射特性的目标切片图像都能够实现较为准确的分割,且计算复杂度小,利于工程应用。

Otsu segmentation algorithm for SAR images based on power transformation

By analyzing the scattering property of the synthetic aperture radar(SAR) images, a power transformation Otsu segmentation algorithm for SAR images is proposed, which can be applied in the preprocessing step of SAR automatic target recognition. Firstly, a speckle suppression method for SAR images is applied to improve the coherence of the amplitude in the background region and the target region using adjacent pixels. Secondly, the adaptive power transformation is employed to the filtered SAR images to enhance the coherence of the amplitude in the target region furthermore. At last, the original Otsu method is adopted to segment the transformed SAR images. The experimental results show that the proposed algorithm provides satisfactory segmentation results for SAR images with different amplitude distribution properties. In addition, the proposed method has low computation burden, which benefits the practical applications.