Power line detection from synthetic aperture radar imagery using coherence of co-polarisation and cross-polarisation estimated in the hough domain

Power lines are an important target type in topographic mapping, as well as in post-disaster and post-strike damage detection. A method of detecting power lines from polarimetric synthetic aperture radar (POLSAR) data using the coherence between the co- and cross-polarisation is proposed based on the azimuth symmetry of the clutter. First, the scattering of power lines is analysed from the aspects of radar cross-section and polarimetric response. Next, the covariance and coherence between the co- and cross-polarisation of power lines and clutter are characterised, and a constant false-alarm rate detector based on the coherence statistics is proposed. In order to obtain a high accuracy, the coherence is estimated in the Hough domain, and the proposed method is improved by considering the sign of the polarimetric orientation angle and using only the real part of the coherence. In addition, the coherence in the special case of azimuth symmetrical power lines is estimated using rotated scattering matrix. Finally, the proposed detector is demonstrated using both simulated and real P-band POLSAR data.