Semi-supervised Remote Sensing Image Segmentation Using Dynamic Region Merging

This paper introduces a remote sensing image segmentation approach by using semi-supervised and dynamic region merging. In remote sensing images, the spatial relationship among pixels has been shown to be sparsely represented by a linear combination of a few training samples from a structured dictionary. The sparse vector is recovered by solving a sparsity-constrained optimization problem, and it can directly determine the class label of the test sample. Through a graph-based technique, unlabeled samples are actively selected based on the entropy of the corresponding class label. With an initially segmented image based semi-supervised, in which the many regions to be merged for a meaningful segmentation. By taking the region merging as a labeling problem, image segmentation is performed by iteratively merging the regions according to a statistical test. Experiments on two datasets are used to evaluate the performance of the proposed method. Comparisons with the state-of-the-art methods demonstrate that the proposed method can effectively investigate the spatial relationship among pixels and achieve better remote sensing image segmentation results.