An Algorithm for Supporting Dual-frequency and Multi-polarimetry SAR Image Interpretation

概要

論文の詳細を見る
This paper presents a pair-wise comparative strategy for supporting interpretation of dual-frequency and multi-polarimetry SAR images (i.e., Pi-SAR images). The basic requirement of Pi-SAR image-interpretation is to clarify the differences with respect to multi-polarized and multi-band information. Color composite images of Pi-SAR data are generally used for image interpretation. There is, however, no theoretical ground on combination cases for assigning multi-polarized Pi-SAR images to "Red (R)" "Green (G), " and "Blue (B)" planes, respectively. As a measure, we have introduced the pair-wise comparative strategy for Pi-SAR image-interpretation, which consists of following steps : Stepl) The dissimilarity measures between pair-wise comparative cases of Pi-SAR images (i.e., LHH, LHV, LVV, XHH, XHV, XVV) are calculated ; Step2) Through the quantification method type-IV, the scatter-diagrams are delineated with respect to items corresponding to each Pi-SAR image and ; Step3) As a final product, the composite images are produced by assigning the dissimilarity images to the color-planes of red, green and blue, respectively. Through the experiments, we conclude that the proposed composite images jointly with the pair-wise comparative table enable effective cross-interpretation for multi-polarized Pi-SAR images in stead of typical color composite images.
社団法人日本リモートセンシング学会の論文