Results of applying two cross-talk removal algorithms (Klein's and Quegan's methods) based on natural distributed targets to the Danish polarimetric airborne SAR (EMISAR) are presented. Also, a new approach of obtaining robust estimates by the algorithms is proposed. The approach is based on a masking procedure using the cross-correlation between co- and cross-polarized returns. The cross-talk estimated for the EMISAR data is low (below -30 dB), virtually independent of range, and stable between different acquisitions-within a mission and between missions. The cross-talk is reduced by 8 dB after correction to less than -38 dB.