SAR Image Segmentation Based on Fast Superpixel Clustering
Wei Zhang, Wenbo Jing, Deliang Xiang, Yi Su · 2021 CIE International Conference on Radar (Radar) · 2021
Accurate segmentation on synthetic aperture radar (SAR) images is a challenging task. In this paper, we propose a SAR image segmentation method based on the fast superpixel clustering algorithm. First, our proposed multi-scale superpixel generation method is used to generate a superpixel map with high quality. Second, the edge-aware superpixel smoothing is performed on the superpixels to alleviate the effect of speckle noise, which facilitates improving segmentation accuracy. Finally, the fuzzy C-means (FCM) clustering is implemented on the superpixels to obtain the final segmentation result. Experimental results on two Sentinel-1 C band SAR images demonstrate that the proposed method exceeds the comparing methods both quantitatively and qualitatively, confirming its advantages in SAR image segmentation.