Summary: | 碩士 === 國立中央大學 === 太空科學研究所 === 85 === 1991年7月歐洲發射偵測全球之遙感探測衛星一號(ERS-1),利用
這顆 衛星上搭載的主動式微波散射儀(Active Microwave
Scatterometer)來觀 測全球海面風場,但其空間解析度較差,對於近
岸之海面風場資訊,會因其 解析度而受限制;而另一個主動式微波儀器
-合成孔徑雷達(Synthetic Aperture Radar,SAR具有較高的空間解
析度,可改善此缺點並提高風場的 解析度,因此本研究的目的即利用此
兩種主動式微波儀器來推算海面的風場 ,並經由CMOD4模式得到背向散
射係數與風場的關係。 首先處理模擬風場資料,藉由模擬散射儀所
觀測到的雷達背向散射係數 資料來推算出海面的風速及風向,重建之方
法是採用動態學習類神經網路 (DLNN),並將重建風場與預設風場加以
比較,結果發現風速及風向之誤 差值分別為1.2m/s及10度;接著,分析
在台灣附近海面的ERS-1散射儀資料, 將重建風場與ECMWF(歐洲中尺度
氣象預報中心)風場比較,結果發現本 研究重建之風場誤差較小;最後
,利用SAR影像來推算海面的風場,針對幾 組台灣附近海面之ERS-1 SAR
影像作初步的研究,結果顯示重建之風場與散 射儀風場趨勢大致吻合。
In July 1991 the European Space Agency (ESA) launched the
European Remote Sensing Satellite (ERS-1), a forerunner of a new
generation of satellites for environmental monitoring, and used
the active microwave scatterometer on ERS-1 to obtain
information of the global sea surface wind. For the low
spatial resolution of scatterometer the information of the
coastal regions would be lost, however, AMI-SAR systems, as
opposed to scatterometers, have the potential to improve it due
to the higher spatial resolution. So the purposes of
thisresearch is to reconstruct wind field using these two
active microwave instruments. In the reconstruction we used
DLNN (Dynamic Learning Neural Networks). A CMOD4 model was used
to train the network to relate the backscattering coefficient
and wind vector.. Firstly, we verified our reconstruction
procedure by means of Monte Carlo simulation. Sets of wind
fields of verious speeds and directions were generated. The
corresponding backscattering coefficients were obtained through
CMOD4 and subsequendly fed into the DLNN for training and
reconstruction. It it found that error of wind speed and wind
direction are less then 1.2 m/s and 10 degree, respectively.
Secondly, we reconstructed weinds from ERS-1 Scatterometer data
around Taiwan water area, then compared the results with these
obtained by ECMWF (European Centre Medium-range Weather
Forecasting). We find that the present method obtains a more
consistent results. Finally, we retrieved surface wind vector
from ERS-1 SAR imagery near Taiwan, and primary results show
that the reconstructed wind vector from SAR imagery are
generally agree with those of Scatterometer.
|