Classification features in phase components of mechanism vectors in PolInSAR optimization

In our experiments, we tested the properties of both the amplitude and phase components of mechanism vectors. Then, we compared the results with coherence map of the optimized data. The result showed that as well as the amplitude components, the phase components of mechanism vectors can also be used for land classification. Moreover, the mechanism vectors have different classification features from coherence map. That is, the mechanism vectors are sensitive to low vegetation field while the coherence map is sensitive to the lake area. In addition, we supposed that the phase components of mechanism vectors can contain more land information than the amplitude components.