利用雷達遙測影像於災害監測,其好處在於全天候及不受雲雨影響.然而目前衛載雷達系統如ERS-2及RADARSAT-1為單偏極雷達系統,雖然ERS-2為單偏極雷達系統,但配合INSAR(干涉雷達)技術仍然可以進行地物分類之進行.INSAR為利用兩組之雷達影像使其干涉而得到相位差影像,相位差會反映出地表高程資訊.而利用兩組以上之雷達影像之相位差(差分干涉雷達)更可得出地表高程之微量變化.另外,INSAR所提供之另一項重要資訊為同調性影像,同調性會反映出地表於兩次雷達觀測之時間差內變化之幅度,若地表改變大則同調性低,若地表改變小則有高同調性.利用同調性影像及兩次觀測之雷達影像,便有更豐富之特徵可供判斷地物或地表高程是否有變化.目前差分干涉雷達能偵測公分級之微小變量,但對於公尺級之變化則難以偵測.本年度將嘗試利用偵測同調性變化來偵測公尺級以上之地表位移。 The main applicability of radar remote sensing on disaster relies on its all weather and day-and-night observation. The proposed project employs ERS-2 radar image on terrain deformation. SAR Interferometry (INSAR) has broadly applied in digital elevation model generation; differential SAR interferometry additionally has the excellent ability to detection the centi-metric deformation of terrain. However, on metric deformation of terrain, it is too rapid on phase change for differential radar interferometry to detect. Therefore, the coherence image will be applied to detection metric deformation in the project this year. A coherence image reveals the degree of change on terrain and land-cover between two observations. Coherence becomes lower as the terrain and land-cover changes. Metric terrain deformation will heavily decrease coherence level. The goals of this project are to detect centi-metric deformation by differential SAR interferometry and metric deformation by cooperating with coherence image. 研究期間：9201 ~ 9212
