請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/70391
標題: | 極化雷達與極化熵應用於物體辨識之研究與雷達天線設計準則 Polarimetric Radar and Entropy Applied to Target Identification and Criterion for Radar Antenna Design |
作者: | Je-Ruei Bai 白哲睿 |
指導教授: | 陳士元 |
關鍵字: | 極化不變量,極化熵,極化雷達,雷達目標物辨識,散射矩陣, invariant properties of polarization,polarimetric entropy,polarimetric radar,radar target identification,scattering matrix, |
出版年 : | 2018 |
學位: | 碩士 |
摘要: | 雷達目標物辨識是近年來很熱門的研究議題,通常應用於遙測系統與軍事科技發展。截至目前有許多目標物辨識的方法被提出,其中最有力且廣為人知的方法為極化熵辨識方法。本論文一開始介紹了何謂極化熵、何謂目標物的散射矩陣、整理一套完整的運算過程將散射矩陣轉換為極化熵並說明其物理意義。接著我們討論這套演算法在實際應用上會面臨的困難,以及在矩陣運算過程中可能存在的問題,然後針對問題提出可能的解決方法。我們提出了極化雷達天線於設計時應遵守的三個準則,可有效提升利用極化熵方法來辨識目標物的成功機會。為了驗證所提出的設計準則,模擬和實驗是必須的。首先我們使用全波模擬軟體建立了模擬雷達目標物辨識的環境,搭配以所提出的準則設計的天線,得到多組散射矩陣,經後處理計算得到極化熵,其結果驗證了所提出理論的正確性與可行性。與傳統雙極化雷達天線相較,採用符合吾人所提出設計準則的雷達天線在極化熵的數值和目標辨識的效果都有較優異的表現。接著我們實際製作該天線與目標物,並在無反射實驗室建立了簡易的量測環境,實驗數據也證明了利用極化熵進行目標物辨識的可行性與所提出天線設計準則的功效。此外,我們初步研究了極化不變量,該技術可用來輔助並提高極化熵之目標物辨識能力。我們整理了文獻中所提出的數組極化不變量的例子,並引入機器學習中廣泛被使用的支撐向量機方法,來幫助我們進行目標物辨識,而其初步的模擬結果也一併附上。本論文的最後整理了結論與數個未來可能的研究方向,期望給有興趣的研究者一些啟發與思考方向。 Target identification has drawn lots of attentions in remote sensing and radar applications. Among various algorithms, the polarimetric entropy is a powerful means to discriminate between different targets. This thesis begins with a brief review of the statistical entropy, the polarimetric radar and scattering matrix, and the formulation of polarimetric entropy and its physical interpretation. Then, the formulas are discussed qualitatively, and based on the discussion, a criterion for radar antenna design is proposed to enhance the target-identifying efficacy. Next, an exemplary scenario is considered, in which the radar antenna is designed following the proposed criterion. After performing full-wave scattering simulations on the scenario, the polarimetric entropies can readily be calculated, and the results outperform those in the conventional paradigm. Additionally, an experiment in anechoic chamber is designed and performed to demonstrate the feasibility and performance of this algorithm in realistic environment. Besides, some invariance properties of polarimetric radar are also introduced to be used in conjunction with polarimetric entropy to enlighten prospective works in the field of radar target identification. To this end, the support vector machine (SVM), well known in machine learning, is adopted. Preliminary simulations are conducted to demonstrate the effectiveness of the concept. Finally, a brief conclusion of this work and some potential future works are given at the end of this thesis. We hope that they can inspire those who are interested in radar target identification. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/70391 |
DOI: | 10.6342/NTU201803339 |
全文授權: | 有償授權 |
顯示於系所單位: | 電信工程學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-107-1.pdf 目前未授權公開取用 | 3.19 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。