An Effective Method of Bridge Detection Based on Polarimetric CSAR

This paper introduce an approach for detecting bridges when the difference in radar echo energy is not obvious. It consists of two steps: extracting targets with high anisotropy and detecting bridge. Firstly, by image segmentation based on the feature of multi-aperture polarimetric entropy, we get targets with high certainty. And then we utilize edge detection to extract straight line segmentations. Finally, the parallel straight line segmentations which meet the bridges' geometric properties are selected as bridges. The proposed approach has been test with polarimetric CSAR data, and the experimental results show that our method can detect bridges effectively.