作者: Hao Hu , Bin Liu , Weiwei Guo , Zenghui Zhang , Wenxian Yu
DOI: 10.1109/IGARSS.2017.8127074
关键词:
摘要: In this paper, we propose a superpixel generation method for synthetic aperture radar (SAR) images by using the density-based spatial clustering of applications with noise (DBSCAN) algorithm. The pixels is firstly grouped to generate initial superpixels probabilistic patch-based (PPB) dissimilarity. Then, small clusters are combined into their neighbor get final results through distance measurement defined statistical models. Experiments on simulated data sets exhibit high boundary adherence generated and demonstrate availability efficiency proposed method.