Extraction of cross-sea bridges from GF-2 PMS satellite images using mathematical morphology

Cross-sea bridges are typical target of natural sceneries. Extracting cross-sea bridges from remote sensing images appears significant. In this study, for the complexity of ground objects in GF-2 PMS satellite images, we have selected direction-augmented linear structuring elements to be used for extraction of cross-sea bridges. Firstly, pre-processing of original image has been carried out; secondly, calculate NDVI and extract water bodies; thirdly, according to the direction of water bodies, select appropriate direction-augmented linear structuring elements, proceed mathematical morphology operations on the water bodies, and assisted by prior knowledge of the bridge, extract the bridge object; finally, select an experimental area to verify the effectiveness and applicability of the method. Our work shows that the direction-augmented linear structuring elements are effective and efficient for extraction of bridges with different directions in GF-2 PMS satellite images, and the approach in this study is important to expand the application areas of domestic high-resolution remote sensing images.