Comparative analysis of polarimetric SAR image exploitation algorithms for environmental stress-change monitoring

摘要

Recent theoretical advances in scene classification using single and multi-frequency POL-SAR techniques have provided a wealth of tools for application to important environmental issues involving ground cover classification and the discriminationof change due to natural versus manmade stress. Among these are techniques which have evolved from the Navy International Cooperative Opportunities Program (NICOP), and although initial emphasis within the discussion to follow has been placed on NICOPresults, a broader base of algorithms is currently being evaluated, with partial results included in this paper. A portion of the multi-frequency POL-SAR data being considered is from the Southern Maryland area, and in particular the entropy based schemeof Cloude [17] [15, chap. 5] and the fuzzy c-means technique [25] are used as an initial baseline for this work. The three-component scattering mechanism described by Krogager [1] and other techniques will be incorporated as the work continues.