Recognition of Bridge over Water in High-Resolution Remote Sensing Images

Bridge is an important artificial target in the field of remote sensing analysis. A method for automatic recognition of bridges over water in high-resolution remote sensing images is presented. Firstly, we establish bridge knowledge models. Based on top-down knowledge-driven, the flow is composed of two steps: hypothesis and testing. Hypothesis is rough positioning including such techniques: waters segmentation, ROI extraction with morphology operator and connectivity sign, candidate regions detection. In testing process, the possible bridge is authenticated using gray features, and the parameters such as coordinates and azimuth of bridge can be obtained. Experiments are executed on the high-resolution remote sensing images, and the results confirm the validity of the proposed method.