Recognition of roads and bridges in SAR images

A region-based method is presented for the recognition of roads and bridges in fully polarimetric synthetic aperture radar (SAR) images. For the recognition of bridges, a CFAR detector is first used to extract strong backscatterers from bridge fences; the backscatterers are then grouped using a Hough transform to find potential bridge fences; finally, bridges are delineated according to the polarimetric features of the regions between the potential fences. After the delineation of bridges, another Hough transform is used to recognize roads. Morphological filtering is used to suppress speckle in SAR images. Results on high resolution SAR images obtained from Lincoln Laboratory are given.