本文關(guān)鍵詞： SAR圖像處理 道路檢測 海陸分割 馬爾科夫隨機(jī)場　出處：《西安電子科技大學(xué)》2014年碩士論文　論文類型：學(xué)位論文

【摘要】：SAR圖像處理在各個(gè)領(lǐng)域已經(jīng)得到廣泛應(yīng)用,是人們獲取信息的重要途徑之一。隨著SAR成像技術(shù)的日益成熟,SAR數(shù)據(jù)的獲取速度與質(zhì)量也已得到迅速發(fā)展,這些都將增加對SAR圖像解譯的需求,在大量高分辨率SAR實(shí)測數(shù)據(jù)的支持下,實(shí)現(xiàn)快速、自動(dòng)化的SAR圖像處理是一個(gè)重要的研究方向。本文從概率角度出發(fā),以馬爾科夫隨機(jī)場模型為基礎(chǔ),結(jié)合SAR圖像數(shù)據(jù)的統(tǒng)計(jì)特性,實(shí)現(xiàn)圖像中的目標(biāo)檢測與分割。論文主要工作總結(jié)如下:1.介紹了SAR圖像處理中道路檢測和海陸分割的研究背景、發(fā)展現(xiàn)狀及意義。概括了論文的主要工作。2.介紹了馬爾科夫隨機(jī)場模型及基礎(chǔ)理論。對于場的概念,鄰域定義與子團(tuán)的定義以及MRF(Markov Random Field)模型進(jìn)行了詳細(xì)的介紹。在應(yīng)用馬爾科夫隨機(jī)場時(shí)最重要的理論支撐就是MRF與吉布斯隨機(jī)場的等價(jià)性,本章對吉布斯隨機(jī)場進(jìn)行了介紹并對兩者的等價(jià)性進(jìn)行了介紹。最后介紹了常用的能量最小化算法。3.研究了SAR圖像中海陸分割的算法的實(shí)現(xiàn)。采用合適的概率模型對SAR數(shù)據(jù)中海面和陸地兩類樣本的統(tǒng)計(jì)特性進(jìn)行描述,得到兩類樣本的似然信息。然后結(jié)合馬爾科夫隨機(jī)場中的鄰域與子團(tuán)定義,將樣本的先驗(yàn)信息考慮在內(nèi),通過能量函數(shù)的定義以及能量最小化算法完成海陸分割。4.研究了SAR圖像中道路檢測的算法實(shí)現(xiàn)。運(yùn)用邊緣檢測完成圖像中疑似道路的點(diǎn)檢測,通過Hough變換將邊緣檢測的結(jié)果轉(zhuǎn)換為線,實(shí)現(xiàn)由點(diǎn)到線的轉(zhuǎn)換。SAR圖像中的道路有著特定的屬性,將這些屬性作為道路的先驗(yàn)信息,借助馬爾科夫隨機(jī)場完成道路檢測。實(shí)驗(yàn)結(jié)果表明了檢測效果與場景的復(fù)雜度有著很大的關(guān)系,合理有效的預(yù)處理非常重要。
[Abstract]:SAR image processing has been widely used in various fields, and it is one of the important ways for people to obtain information. With the development of SAR imaging technology, the acquisition speed and quality of SAR data have been developed rapidly. All these will increase the demand for SAR image interpretation. With the support of a large number of high-resolution SAR measured data, it is an important research direction to realize fast and automatic SAR image processing. Based on Markov random field model and combined with the statistical characteristics of SAR image data, target detection and segmentation are realized. The main work of this paper is summarized as follows: 1. The research background of road detection and land and sea segmentation in SAR image processing is introduced. The main work of this paper is summarized. 2. Markov random field model and basic theory are introduced. The definition of neighborhood, the definition of sub-cluster and the MRF(Markov Random field model are introduced in detail. The most important theoretical support in applying Markov random field is the equivalence of MRF and Gibbs random field. In this chapter, Gibbs random field is introduced and its equivalence is introduced. Finally, the commonly used energy minimization algorithm .3.The realization of land and sea segmentation algorithm in SAR image is studied. To describe the statistical characteristics of sea and land samples in SAR data. The likelihood information of two kinds of samples is obtained, and the priori information of samples is taken into account by combining the definitions of neighborhood and sub-cluster in Markov random fields. Through the definition of energy function and the energy minimization algorithm to complete the land and sea segmentation. 4. The algorithm of road detection in SAR image is studied. The edge detection is used to detect the suspected road points in the image. The result of edge detection is transformed into a line by Hough transform. The path in the image is converted from point to line. The road has special attributes, which are regarded as the priori information of the road. The experimental results show that the detection effect is closely related to the complexity of the scene, and the reasonable and effective pretreatment is very important.
【學(xué)位授予單位】：西安電子科技大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：TN957.52

【共引文獻(xiàn)】

相關(guān)博士學(xué)位論文 前2條

1 林偉;極化SAR圖像分類的投影尋蹤方法研究[D];西北工業(yè)大學(xué);2007年

2 張?bào)K祥;小波變換和馬爾可夫隨機(jī)場在圖像處理中的應(yīng)用研究[D];天津大學(xué);2007年

，

本文編號：1550105