【正文】 I / 68浙 江 省 碩 士 學(xué) 位 論 文論文題目：嵌入式高光譜數(shù)據(jù)庫的開發(fā)及目標(biāo)分類應(yīng)用研究授予學(xué)位學(xué)科專業(yè)：控制工程學(xué)科專業(yè)代碼：085210研究方向：智能信息融合與處理省編號：II / 68摘 要高光譜遙感技術(shù)在過去三十年中取得了飛速的發(fā)展。它已經(jīng)成為人們獲得地物等感興趣目標(biāo)信息尤為重要的手段，在民用與軍用領(lǐng)域均發(fā)揮著巨大的作用。高光譜遙感具有數(shù)據(jù)波段眾多，波段間相關(guān)性強等特點，這為數(shù)據(jù)處理技術(shù)提出了更高的要求。由于高光譜遙感數(shù)據(jù)特有的“圖譜一體化”特性，因此收集并積累各類不同地物的光譜響應(yīng)數(shù)據(jù)信息，即光譜指紋，一直以來都是高光譜遙感影像領(lǐng)域的基礎(chǔ)研究工作和不可或缺的重要環(huán)節(jié)。而目前較成熟的光譜數(shù)據(jù)庫系統(tǒng)都是大型數(shù)據(jù)庫系統(tǒng)，可移植性差，使用環(huán)境受限。因此，對創(chuàng)建輕量型可移植嵌入式光譜指紋數(shù)據(jù)庫系統(tǒng)作了研究，并通過光譜數(shù)據(jù)庫系統(tǒng)中存儲的光譜指紋樣本，對未知高光譜遙感數(shù)據(jù)進(jìn)行目標(biāo)識別。同時，將 RX異常檢測算法、支持向量機算法移植到 ARM 平臺上，與創(chuàng)建的嵌入式數(shù)據(jù)庫協(xié)同使用，實現(xiàn)嵌入式平臺下無人工干預(yù)的高光譜遙感目標(biāo)分類，數(shù)據(jù)管理一體化。主要工作如下：(1) 介紹了高光譜遙感的研究背景，進(jìn)而闡述了研究內(nèi)容和研究目的。在此基礎(chǔ)之上，重點綜述了光譜數(shù)據(jù)庫系統(tǒng)的概念及發(fā)展現(xiàn)狀。結(jié)合研究內(nèi)容，著重介紹目前流行的嵌入式數(shù)據(jù)庫系統(tǒng)。(2) 針對本文數(shù)據(jù)庫系統(tǒng)開發(fā)的需要，介紹了嵌入式環(huán)境下的數(shù)據(jù)庫系統(tǒng)知識，對比常用嵌入式數(shù)據(jù)庫系統(tǒng)，選擇 SQLite 嵌入式數(shù)據(jù)庫。并對 SQLite數(shù)據(jù)庫系統(tǒng)的體系結(jié)構(gòu)、接口函數(shù)、事務(wù)和鎖以及 SQLite 的 SQL 語句進(jìn)行研究。最后在 ARM 開發(fā)板上實現(xiàn) SQLite 數(shù)據(jù)庫的移植，在 ARM 環(huán)境下對SQLite 數(shù)據(jù)庫進(jìn)行測試。(3) 借鑒學(xué)習(xí)國內(nèi)外成熟光譜數(shù)據(jù)庫系統(tǒng)，介紹了嵌入式光譜數(shù)據(jù)庫系統(tǒng)開發(fā)的設(shè)計思路。搭建開發(fā)環(huán)境，交叉編譯 Linux 下 GUI 界面開發(fā) Qt/E，并移植到 ARM 開發(fā)板。從界面設(shè)計、數(shù)據(jù)表設(shè)計和功能設(shè)計三個方面詳細(xì)介紹了數(shù)據(jù)庫系統(tǒng)的開發(fā)過程。(4) 從已建立的數(shù)據(jù)庫系統(tǒng)中提取樣本，運用 RX 異常檢測算法，通過光譜角映射和支持向量機對高光譜遙感數(shù)據(jù)進(jìn)行目標(biāo)識別，分析總結(jié)仿真結(jié)果，并移植 RX 異常檢測算法和支持向量機算法到 ARM 嵌入式平臺。關(guān)鍵字：高光譜遙感，嵌入式數(shù)據(jù)庫，光譜數(shù)據(jù)庫，RX 異常檢測，支持向量機III / 68ABSTRACTHyperspectral remote sensing technology has achieved rapid development in the past 30 has bee a particularly important means that people obtain feature information of which they’re interested in objects, and plays a huge role in both the civil and military remote sensing technology with features such as numerous data bands,interband correlation, puts forward higher requirements for data hyperspectral remote sensing data has the characteristics of integration of diagram and spectrum, it has always been a basic research work indispensable important segment in the field of hyperspectral remote sensing image to collect and accumulate information from all kinds of different ground objects spectrum response data. But at present most of the mature spectral database system is too large, poor portability, limited by the environment. Therefore, this paper makes a research on creating a lightweight portable embedded database system for spectrum curve, classifies target of unknown hyperspectral remote sensing data through samples stored in the database system. At the same time, the RX anomaly detection algorithm, support vector machine algorithm is transplanted to the ARM platform, working with embedded database cooperatively, realizing hyperspectral remote sensing target classification and data management on the embedded platform without manual intervention. The main work is as follows:(1) This paper introduces the background of hyperspectral remote sensing, and expounds the research content and research purposes. On this basis, this paper focuses on introducing the concept of spectral database system and summarizing its current situation of the development. Combined with the research content, this paper introduces some popular embedded database system at present.(2) Aimed at the needs of developing the database system, this paper introduces the knowledge of embedded database system. Comparing to monly used embedded database system, this paper choose the SQLite database system. This paper researches the architecture, the interface functions, the transactions, the lock and SQLite SQL statements of SQLite. Finally, this paper transplants SQLite on the ARM development board, and tests SQLite under the ARM environment.(3) Using domestic and overseas mature spectral database system for reference, IV / 68the design idea of embedded spectral database system development was introduced in this paper. The development environment was built and the paper crosspiles Linux GUI interface Qt/E. Then, they were transplanted to the ARM development board. The development process of database was introduces in detail from the aspects of interface design, data table design and function design.(4) Samples were extracted from the established database system and anomaly detected using RX. And spectrum angle matching and support vector machine were applied for target identification for hyperspectral remote sensing data. Then, simulation results were analyzed and summarized. Finally, RX anomaly detection algorithm and support vector machine algorithm were transplanted to the ARM embedded platform.Keywords:Hyperspectral remote sensing, embedded database system, spectral database system, anomaly detection, support vector machineV / 68目 錄摘 要 ............................................................................................................................IABSTRACT.................................................................................................................III目 錄 ...........................................................................................................................V第一章 緒論 ..................................................................................................................1 研究背景 ............................................................................................................1 研究內(nèi)容及目的 ................................................................................................2 光譜數(shù)據(jù)庫系統(tǒng)概念及發(fā)展現(xiàn)狀 ....................................................................2 嵌入式數(shù)據(jù)庫發(fā)展現(xiàn)狀 ....................................................................................5第二章 嵌入式數(shù)據(jù)庫系統(tǒng) ..........................................................................................7 引言 ....................................................................................................................7 嵌入式數(shù)據(jù)庫系統(tǒng)及特點 ................................................................................7 SQLITE 嵌入式數(shù)據(jù)庫 .....................................................................................9 SQLITE 數(shù)據(jù)庫系統(tǒng)及特點 ......................................................................9 SQLITE 層次調(diào)用與體系架構(gòu) ................................................................10 SQLIT