Investigation on Different Interferometric Coherence Optimization Methods

The interferometric coherence optimization methods are based on the definition of projection vectors, which is the way to combine the information occuring from the polarimetric channels of both interferometric images. An important point is the physical interpretation of the projection vectors in terms of scattering characteristics. An optimization algorithm has been first proposed by Cloude and Papathanassiou(referred as C & P algorithm [2]) , based on the selection of the best pair of projection vectors. The C & P algorithm leads to a 3 × 3 complex eigenvector problem, each projection vector representing a selected mechanism for each image. We prove first that the C & P algorithm corresponds mathematically to the most general optimization of the interferometric coherence. This means that the interferograms formed by using these mechanisms have the highest possible coherence. Nevertheless a major drawback of the C & P algorithm is that in neighboring resolution cells the optimal scattering mechanisms can vary significantly, introducing discontinuities into the extracted height differences. Moreover, the C & P algorithm does not fix uniquely the phase difference between the two projection vectors. Although an additional condition can be freely fixed, an attempt to apply the C & P algorithm to a selected set of our data led to an unexploitable height mapping. In this paper an alternative coherence optimization algorithm using the same projection vector for both images is proposed. This method is tested and compared with the first one in terms of physical interpretation and results ; the properties of the projection vectors obtained by both methods are discussed. Finally, some tests have been carried out in order to assess the height estimations obtained from both methods.