A Fast Algorithm for SAR Image Segmentation Based on Key Pixels

摘要: Recent high-performance clustering methods process all pixels when segmenting an image, which results in a very large time complexity of these algorithms. Additionally, the performance such algorithms can be severely affected by noise dealing with highly polluted images. To address problems, we propose new unsupervised algorithm for synthetic aperture radar images based on fuzzy approach, called fast C-means key pixels. Our first selects subset special “key” rule local extrema, and then performs image segmentation only using combined nonlocal information. Next, remaining non-key rapidly segmented combining similarity metric is robust to speckle noise. This approach greatly accelerates overall because time-consuming operation performed small We show effectiveness our proposed series experiments including twelve simulated four real Moreover, validate results, compare obtained those seven other state-of-the-art from literature. The experimental suggest that outperforms both computational speed suppression.