多時域雷達干涉技術（MTI）於長期地表變形偵測之探討：以濁水溪沖積扇為例

ASSESSMENT OF MULTI-TEMPORAL INTERFEROMETRY FOR DETECTING LONG-TERM SURFACE DEFORMATION: A CASE STUDY IN CHOSHUI RIVER ALLUVIAL FAN

呂喬茵(CHIAO-YIN LU) ； 張志新(CHIH-HSIN CHANG) ； 蘇元風(YUAN-FONG SU) ； 胡植慶(JYR-CHING HU)

《工程環境會刊》 39期 (2019/12) Pp. 29-45

https://doi.org/10.6562/JEE.201912_(39).0002

摘要

透過多時域雷達干涉技術（Multi-Temporal Interferometry, MTI）可獲得長時間、大範圍、高密集點位與精度的地表三維變動資訊，於地層下陷、火山活動、山崩潛移等地表變形監測上有很高的應用潛力。本研究以近四十年來，嚴重地層下陷的濁水溪沖積扇為例，透過三種不同的合成孔徑雷達影像（ERS、ENVISAT、ALOS）進行MTI技術之分析，並利用與衛星影像同一時期之GPS站資料進行數據的融合，以獲取垂直方向的下陷趨勢與變化。分析結果中，除了ERS衛星影像成果因PS點之密度較低導致相位回復錯誤（phase unwrapping error）以及其所能偵測的最大速率限制使成果不良外，ENVISAT與ALOS衛星影像所得之下陷區域的空間分布與現地監測資料一致。且PS點的密度（115ps／km2）遠高於傳統測量點位（0.19水準點／km2），因此利用MTI技術可得更細緻的地表空間變化分布，彌補缺乏水準測量與GPS測站區域之地表變動情形，並透過與GPS資料進行融合處理及扣除研究區水平變形的影響，可有效獲得與現地監測資料更接近之地層下陷速率，提供政府相關單位後續可廣泛掌握精確下陷空間資訊之重要技術。

關鍵字

合成孔徑雷達 ； 多時域雷達干涉技術 ； 地層下陷 ； 地表變形監測

並列摘要

Multi-Temporal Interferometry (MTI) is a powerful remote sensing technique which could provide wide area, long-term, high density and precise information for surface deformation analysis, such as subsidence investigating, volcano monitoring and creep landslide detecting. In order to improve our understanding of the relevant land subsidence issues, we chose Choshui River alluvial fan as the study area where still suffering serious land subsidence hazard in recent years. This study applied three kinds of SAR images including ERS, ENVISAT and ALOS in order to acquire a long time series of subsidence information. Different from previous studies, we calibrated the results and removed the contribution of horizontal tectonic movement along the line of sight (LOS) direction by GPS data in the same time period. The spatial patterns of subsidence derived from ENVISAT and ALOS are not only agree with that from observations (leveling and GPS) data but reveal more details. However, the spatial and temporal phase unwrapping errors are evident in coastal area of the mean velocity map derived from ERS images. The reason for that may be the lower density of scatterers and limitation of maximum detecting velocity of ERS dataset. Due to the density of PS pixel (115 ps/ km2) is much higher than GPS and leveling (0.19 benchmark/km2), the spatial deformation pattern detected by PSInSAR analysis could provide more detailed spatial information than the traditional monitoring systems (leveling and GPS).

並列關鍵字

Synthetic Aperture Radar ； Multi-Temporal Interferometry ； Subsidence ； Surface deformation monitoring