基于局部支持向量機(jī)的高分辨率遙感圖像分類
發(fā)布時(shí)間：2018-08-31 11:33
【摘要】：隨著圖像采集傳感技術(shù)的發(fā)展以及社會(huì)需求的不斷提高,遙感圖像呈現(xiàn)出高分辨率的發(fā)展趨勢(shì)。分辨率越高的遙感圖像所含信息內(nèi)容越豐富,對(duì)各種應(yīng)用領(lǐng)域的發(fā)展促進(jìn)更大。因此,高分辨率的遙感圖像成為地物目標(biāo)識(shí)別的重要來源。遙感圖像分辨率的日益增加也給遙感圖像處理領(lǐng)域帶來越來越大的挑戰(zhàn)。隨著遙感圖像分辨率的不斷提升,應(yīng)用領(lǐng)域?qū)ζ渚鹊男枨蟛粩嘣黾?導(dǎo)致需要處理的數(shù)據(jù)量呈指數(shù)級(jí)的增長(zhǎng)。因此,傳統(tǒng)的人工目視解譯的圖像處理模式已完全不能滿足現(xiàn)實(shí)需求,利用數(shù)據(jù)挖掘、機(jī)器學(xué)習(xí)等依靠計(jì)算機(jī)自動(dòng)進(jìn)行圖像分類成為遙感圖像處理領(lǐng)域的主流。然而,即使是利用計(jì)算機(jī)自動(dòng)分類也因?yàn)閿?shù)據(jù)量的爆炸性增長(zhǎng)而導(dǎo)致處理時(shí)間長(zhǎng),分類精度達(dá)不到預(yù)期效果等現(xiàn)象；另一方面,高分辨率遙感圖像具有更多的圖像特征,除了包括傳統(tǒng)圖像具有的光譜信息以外,還有紋理特征、幾何特征以及上下文信息等等。而傳統(tǒng)的遙感圖像處理主要依據(jù)單個(gè)特征進(jìn)行分類,這種方式因?qū)Φ匚镏写嬖诘摹巴锂愖V”以及“同譜異物”問題無法很好解決而導(dǎo)致分類結(jié)果差。支持向量機(jī)(SVM)模型是依靠訓(xùn)練樣本集中的支持向量學(xué)習(xí)而來,具有稀疏性好,分類精度高以及運(yùn)行速度快等優(yōu)勢(shì),特別適合處理訓(xùn)練數(shù)據(jù)集小、非線性以及高維特征的分類問題。針對(duì)具有海量數(shù)據(jù)的高分辨率遙感圖像分類,支持向量機(jī)具有較大的優(yōu)勢(shì)。然而,這種傳統(tǒng)支持向量機(jī)沒有全局一致性,為了進(jìn)一步提高其分類精度,融合近鄰分類器KNN和SVM的局部支持向量機(jī)算法被應(yīng)用于遙感圖像的分類問題中。局部支持向量機(jī)結(jié)合了兩個(gè)分類器的優(yōu)勢(shì),因此分類精度更高。根據(jù)高分辨率遙感圖像分類問題的特點(diǎn),為了進(jìn)一步提高對(duì)高分辨率遙感圖像的分類效果,降低其計(jì)算的復(fù)雜度,文章在對(duì)現(xiàn)有研究成果的梳理下,對(duì)局部支持向量機(jī)在高分辨率遙感圖像分類中的應(yīng)用進(jìn)行了深入研究,主要的研究工作如下：1、提出了基于不確定性的改進(jìn)KNNSVM的局部支持向量機(jī)—-BKNNSVM算法。詳細(xì)分析了具有代表性的局部支持向量機(jī)KNNSVM算法存在的算法時(shí)間復(fù)雜度高的不足。對(duì)KNNSVM算法進(jìn)行深入研究后,發(fā)現(xiàn)KNNSVM算法是通過嚴(yán)格的K近鄰分類器(SKNN)和局部支持向量機(jī)(LSVM)兩個(gè)分類器的融合來確定樣本類標(biāo)。分析KNNSVM算法的時(shí)間開銷曲線,發(fā)現(xiàn)需要建立SVM分類器確定類標(biāo)的樣本越多,時(shí)間開銷越大。考慮到對(duì)任何一個(gè)未標(biāo)記樣本其近鄰分布都是滿足二項(xiàng)式分布的,借助Beta分布對(duì)未標(biāo)樣本不確定性的計(jì)算,論文提出了BKNNSVM算法,該算法通過設(shè)定不確定性閾值及近鄰K值兩個(gè)參數(shù)值的大小,通過增加(減少)參與KNN分類器的分類樣本數(shù),從而減少(增加)需要建立SVM分類器的個(gè)數(shù)來調(diào)節(jié)BKNNSVM算法的時(shí)間復(fù)雜度。實(shí)驗(yàn)結(jié)果顯示,設(shè)置合理的閾值及K值的BKNNSVM能夠在保持KNNSVM算法精度的同時(shí),明顯地降低算法的時(shí)間復(fù)雜度。2、提出了基于距離的局部支持向量機(jī)算法DLSVM。通過對(duì)SVM錯(cuò)分樣本在超平面附近的分布特點(diǎn),發(fā)現(xiàn)超平面附近的樣本分類錯(cuò)誤率最高,分類精度最低,但這些樣本利用局部支持向量機(jī)進(jìn)行分類時(shí),分類精度較高。為了尋找這些被錯(cuò)分的點(diǎn),我們借助了主動(dòng)學(xué)習(xí)的相關(guān)理論,認(rèn)為離超平面越近,傳統(tǒng)支持向量機(jī)對(duì)樣本的錯(cuò)分率越高。因此,DLSVM首先計(jì)算未標(biāo)記樣本離超平面的距離,針對(duì)離超平面較近的樣本建立局部支持向量機(jī),提高其分類精度,而遠(yuǎn)距離的樣本直接采用傳統(tǒng)的SVM分類以減少分類時(shí)間。由于只對(duì)少數(shù)近距離樣本采用局部支持向量機(jī)算法,DLSVM在傳統(tǒng)支持向量機(jī)的基礎(chǔ)上時(shí)間復(fù)雜度增加不多,但分類精度有顯著提高,特別地,該算法的時(shí)間復(fù)雜度遠(yuǎn)小于KNNSVM算法。3、提出了多類分類問題的有向無環(huán)圖局部支持向量機(jī)(DAGLSVM)分類方法。支持向量機(jī)分類器SVM是用來解決二分類問題的。因?yàn)槠渚哂蟹诸惥雀�、泛化能力�?qiáng),因此也被推廣運(yùn)用于多分類問題。SVM解決多分類問題的主要方法有一對(duì)一,一對(duì)多、有向無環(huán)圖DAGSVM及二叉樹SVM等�？紤]到如果將KNNSVM算法直接應(yīng)用到一對(duì)一的多類分類問題中,建立局部支持向量機(jī)的數(shù)量將在二分類問題的基礎(chǔ)上呈二次方倍增,當(dāng)類數(shù)目較多時(shí),時(shí)間開銷將無法忍受。因此,為了降低時(shí)間復(fù)雜度,提出了基于有向無環(huán)圖(DAG)的多類分類局部支持向量機(jī)算法——DAGLSVM。該算法為每一個(gè)未標(biāo)記樣本建立近鄰訓(xùn)練樣本集,在DAG拓?fù)浣Y(jié)構(gòu)中的每個(gè)節(jié)點(diǎn),選擇未標(biāo)記樣本類不確定性最小的兩個(gè)類優(yōu)先進(jìn)行決策,以降低導(dǎo)致累積誤差的風(fēng)險(xiǎn)。最后的實(shí)驗(yàn)結(jié)果顯示,在合適的K值參數(shù)下,DAGLSVM算法在運(yùn)行時(shí)間消耗增加不大的情況下,能夠有效提高高分辨率遙感圖像的分類精度。4、設(shè)計(jì)了遙感圖像處理軟件平臺(tái),并通過該平臺(tái)完成文章中涉及的所有實(shí)驗(yàn)項(xiàng)目。該平臺(tái)在臺(tái)灣大學(xué)林智仁教授開發(fā)的開源代碼一-LIBSVM的基礎(chǔ)上,利用JAVA平臺(tái)設(shè)計(jì)開發(fā)了遙感圖像支持向量機(jī)分類軟件。該軟件以LIBSVM算法為核心,綜合數(shù)據(jù)挖掘軟件平臺(tái)——WEKA平臺(tái),實(shí)現(xiàn)了KNN、KNNSVM、BKNNSVM、DLSVM、DAGSVM以及DAGLSVM等多種分類算法。完成了對(duì)遙感圖像的顯示,圖像特征提取、分類、分類結(jié)果存儲(chǔ)以及結(jié)果顯示等功能。總之,由于局部支持向量機(jī)不僅能夠提高傳統(tǒng)支持向量機(jī)的精度,還能保持其泛化性好、支持小訓(xùn)練樣本集的特點(diǎn),因此,本文針對(duì)局部支持向量機(jī)算法提出的各種優(yōu)化算法能夠提高高分辨率遙感圖像的分類性能,降低其運(yùn)算的時(shí)間復(fù)雜度,有利于進(jìn)一步提高遙感圖像的處理速度,發(fā)揮其更大的社會(huì)經(jīng)濟(jì)效益。
[Abstract]:With the development of image acquisition and sensing technology and the continuous improvement of social demand, remote sensing images show a trend of high resolution. The higher the resolution of remote sensing images, the richer the information content, the greater the development of various applications. Therefore, high resolution remote sensing images become an important source of object recognition. The increasing resolution of remote sensing image also brings more and more challenges to the field of remote sensing image processing. With the continuous improvement of the resolution of remote sensing image, the demand for its accuracy in the application field is increasing, resulting in an exponential increase in the amount of data to be processed. Unable to meet the actual needs, the use of data mining, machine learning and other computer-based automatic image classification has become the mainstream in the field of remote sensing image processing. High-resolution remote sensing images have more image features, including texture features, geometric features and context information besides the spectral information of traditional images. However, traditional remote sensing image processing is mainly based on a single feature classification, which is due to the existence of "similarities and differences" and "similarities and differences" in terrain. Support Vector Machine (SVM) model is based on the support vector learning of training sample set. It has the advantages of good sparsity, high classification accuracy and fast running speed. It is especially suitable for dealing with the classification problem of small training data set, nonlinear and high-dimensional features. However, this traditional support vector machine does not have global consistency. In order to further improve the classification accuracy, the local support vector machine (LSVM) algorithm which combines the nearest neighbor classifier KNN and SVM is applied to the classification of remote sensing images. According to the characteristics of high-resolution remote sensing image classification problem, in order to further improve the classification effect of high-resolution remote sensing image and reduce the computational complexity, this paper combs the existing research results, on the basis of local support vector machine in high-resolution remote sensing image classification. The application of image classification is studied deeply. The main research work is as follows: 1. An improved KNNSVM-BKNNSVM algorithm based on uncertainty is proposed. The shortcomings of KNNSVM algorithm with high time complexity are analyzed in detail. After that, it is found that the KNNSVM algorithm determines the sample class scale by the fusion of the strict K-nearest neighbor classifier (SKNN) and local support vector machine (LSVM). By analyzing the time cost curve of the KNNSVM algorithm, it is found that the more samples need to be established to determine the class scale by SVM classifier, the greater the time cost. The nearest neighbor distribution satisfies the binomial distribution. The BKNNSVM algorithm is proposed by calculating the uncertainty of unmarked samples with Beta distribution. By setting the uncertainty threshold and the size of the nearest neighbor K value, the algorithm reduces (increases) the number of samples participating in the KNN classifier by increasing (decreases) the number of samples participating in the classification. The experimental results show that BKNNSVM with reasonable thresholds and K values can significantly reduce the time complexity while maintaining the accuracy of the KNNSVM algorithm. 2. A distance-based local support vector machine algorithm DLSVM is proposed. Near the hyperplane, the sample classification error rate is the highest and the classification accuracy is the lowest, but these samples are classified by local support vector machine with high classification accuracy. Therefore, DLSVM firstly calculates the distance between unlabeled samples and hyperplane, and establishes local support vector machine (LSVM) for the samples nearer to the hyperplane to improve the classification accuracy. However, the traditional SVM classification is directly used for the remote samples to reduce the classification time. In particular, the time complexity of this algorithm is much less than that of KNNSVM. 3. A directed acyclic graph local support vector machine (DAGLSVM) classification method for multi-class classification problems is proposed. Support vector machine classifier SVM is used to solve the two-class classification problem. Because of its high classification accuracy and strong generalization ability, SVM is also widely used in multi-class classification problems. The main methods of SVM to solve multi-class classification problems are one-to-one, one-to-many, directed acyclic graph DAGSVM and binary tree SVM. Considering that if KNNSVM algorithm is directly applied to one-to-one multi-class classification problems, the locality is established. The number of support vector machines will be quadratically multiplied on the basis of binary classification problem, and the time cost will be unbearable when the number of classes is large. Therefore, in order to reduce the time complexity, a multi-class classification local support vector machine algorithm based on directed acyclic graph (DAG) called DAGLSVM is proposed. Neighborhood training sample sets, each node in the DAG topology, choose two classes with the least uncertainty of unlabeled sample classes to make decisions in order to reduce the risk of cumulative errors. Finally, the experimental results show that the DAGLSVM algorithm can effectively improve the performance of the DAGLSVM algorithm under the appropriate K-value parameters without increasing the running time. The classification accuracy of high-resolution remote sensing image is 4.The software platform of remote sensing image processing is designed, and all the experimental projects involved in this paper are completed through this platform.Based on the open source-LIBSVM developed by Professor Lin Zhiren of Taiwan University, the classification software of remote sensing image is designed and developed by using JAVA platform. The software takes LIBSVM algorithm as the core and integrates data mining software platform-WEKA platform. It implements various classification algorithms such as KNN, KNNSVM, BKNNSVM, DLSVM, DAGSVM and DAGLSVM. It completes the functions of remote sensing image display, image feature extraction, classification, classification result storage and result display. It can not only improve the accuracy of traditional support vector machine, but also maintain its good generalization and support the characteristics of small training sample set. Therefore, various optimization algorithms proposed in this paper for local support vector machine algorithm can improve the classification performance of high-resolution remote sensing images, reduce the time complexity of its operation, which is conducive to further improving remote sensing. The speed of image processing will bring greater social and economic benefits.
【學(xué)位授予單位】：中國(guó)地質(zhì)大學(xué)
【學(xué)位級(jí)別】：博士
【學(xué)位授予年份】：2015
【分類號(hào)】：P237
[Abstract]:With the development of image acquisition and sensing technology and the continuous improvement of social demand, remote sensing images show a trend of high resolution. The higher the resolution of remote sensing images, the richer the information content, the greater the development of various applications. Therefore, high resolution remote sensing images become an important source of object recognition. The increasing resolution of remote sensing image also brings more and more challenges to the field of remote sensing image processing. With the continuous improvement of the resolution of remote sensing image, the demand for its accuracy in the application field is increasing, resulting in an exponential increase in the amount of data to be processed. Unable to meet the actual needs, the use of data mining, machine learning and other computer-based automatic image classification has become the mainstream in the field of remote sensing image processing. High-resolution remote sensing images have more image features, including texture features, geometric features and context information besides the spectral information of traditional images. However, traditional remote sensing image processing is mainly based on a single feature classification, which is due to the existence of "similarities and differences" and "similarities and differences" in terrain. Support Vector Machine (SVM) model is based on the support vector learning of training sample set. It has the advantages of good sparsity, high classification accuracy and fast running speed. It is especially suitable for dealing with the classification problem of small training data set, nonlinear and high-dimensional features. However, this traditional support vector machine does not have global consistency. In order to further improve the classification accuracy, the local support vector machine (LSVM) algorithm which combines the nearest neighbor classifier KNN and SVM is applied to the classification of remote sensing images. According to the characteristics of high-resolution remote sensing image classification problem, in order to further improve the classification effect of high-resolution remote sensing image and reduce the computational complexity, this paper combs the existing research results, on the basis of local support vector machine in high-resolution remote sensing image classification. The application of image classification is studied deeply. The main research work is as follows: 1. An improved KNNSVM-BKNNSVM algorithm based on uncertainty is proposed. The shortcomings of KNNSVM algorithm with high time complexity are analyzed in detail. After that, it is found that the KNNSVM algorithm determines the sample class scale by the fusion of the strict K-nearest neighbor classifier (SKNN) and local support vector machine (LSVM). By analyzing the time cost curve of the KNNSVM algorithm, it is found that the more samples need to be established to determine the class scale by SVM classifier, the greater the time cost. The nearest neighbor distribution satisfies the binomial distribution. The BKNNSVM algorithm is proposed by calculating the uncertainty of unmarked samples with Beta distribution. By setting the uncertainty threshold and the size of the nearest neighbor K value, the algorithm reduces (increases) the number of samples participating in the KNN classifier by increasing (decreases) the number of samples participating in the classification. The experimental results show that BKNNSVM with reasonable thresholds and K values can significantly reduce the time complexity while maintaining the accuracy of the KNNSVM algorithm. 2. A distance-based local support vector machine algorithm DLSVM is proposed. Near the hyperplane, the sample classification error rate is the highest and the classification accuracy is the lowest, but these samples are classified by local support vector machine with high classification accuracy. Therefore, DLSVM firstly calculates the distance between unlabeled samples and hyperplane, and establishes local support vector machine (LSVM) for the samples nearer to the hyperplane to improve the classification accuracy. However, the traditional SVM classification is directly used for the remote samples to reduce the classification time. In particular, the time complexity of this algorithm is much less than that of KNNSVM. 3. A directed acyclic graph local support vector machine (DAGLSVM) classification method for multi-class classification problems is proposed. Support vector machine classifier SVM is used to solve the two-class classification problem. Because of its high classification accuracy and strong generalization ability, SVM is also widely used in multi-class classification problems. The main methods of SVM to solve multi-class classification problems are one-to-one, one-to-many, directed acyclic graph DAGSVM and binary tree SVM. Considering that if KNNSVM algorithm is directly applied to one-to-one multi-class classification problems, the locality is established. The number of support vector machines will be quadratically multiplied on the basis of binary classification problem, and the time cost will be unbearable when the number of classes is large. Therefore, in order to reduce the time complexity, a multi-class classification local support vector machine algorithm based on directed acyclic graph (DAG) called DAGLSVM is proposed. Neighborhood training sample sets, each node in the DAG topology, choose two classes with the least uncertainty of unlabeled sample classes to make decisions in order to reduce the risk of cumulative errors. Finally, the experimental results show that the DAGLSVM algorithm can effectively improve the performance of the DAGLSVM algorithm under the appropriate K-value parameters without increasing the running time. The classification accuracy of high-resolution remote sensing image is 4.The software platform of remote sensing image processing is designed, and all the experimental projects involved in this paper are completed through this platform.Based on the open source-LIBSVM developed by Professor Lin Zhiren of Taiwan University, the classification software of remote sensing image is designed and developed by using JAVA platform. The software takes LIBSVM algorithm as the core and integrates data mining software platform-WEKA platform. It implements various classification algorithms such as KNN, KNNSVM, BKNNSVM, DLSVM, DAGSVM and DAGLSVM. It completes the functions of remote sensing image display, image feature extraction, classification, classification result storage and result display. It can not only improve the accuracy of traditional support vector machine, but also maintain its good generalization and support the characteristics of small training sample set. Therefore, various optimization algorithms proposed in this paper for local support vector machine algorithm can improve the classification performance of high-resolution remote sensing images, reduce the time complexity of its operation, which is conducive to further improving remote sensing. The speed of image processing will bring greater social and economic benefits.
【學(xué)位授予單位】：中國(guó)地質(zhì)大學(xué)
【學(xué)位級(jí)別】：博士
【學(xué)位授予年份】：2015
【分類號(hào)】：P237
【相似文獻(xiàn)】
相關(guān)期刊論文 前10條
1 馮劍豐;王洪禮;李勝朋;;基于支持向量機(jī)的浮游植物密度預(yù)測(cè)研究[J];海洋環(huán)境科學(xué);2007年05期
2 蔡佳;陳洪;;ν-支持向量機(jī)的收斂性[J];湖北大學(xué)學(xué)報(bào)(自然科學(xué)版);2005年04期
3 唐萬梅;;基于灰色支持向量機(jī)的新型預(yù)測(cè)模型[J];系統(tǒng)工程學(xué)報(bào);2006年04期
4 熊宇虹;溫志渝;梁玉前;陳勤;張波;劉妤;向賢毅;;基于支持向量機(jī)的模型傳遞方法研究[J];光譜學(xué)與光譜分析;2007年01期
5 左林;;一個(gè)新的稀疏支持向量機(jī)模型[J];山西師范大學(xué)學(xué)報(bào)(自然科學(xué)版);2008年03期
6 赫英明;王漢杰;;支持向量機(jī)在積雪檢測(cè)中的應(yīng)用[J];南京氣象學(xué)院學(xué)報(bào);2009年01期
7 賈亮;王禮力;;支持向量機(jī)在上市公司財(cái)務(wù)預(yù)警模型中的應(yīng)用[J];淮陰工學(xué)院學(xué)報(bào);2010年02期
8 郎宇寧;藺娟如;;基于支持向量機(jī)的多分類方法研究[J];中國(guó)西部科技;2010年17期
9 謝福民;;支持向量機(jī)理論研究[J];科技廣場(chǎng);2011年01期
10 王紅蔚;席紅旗;孔波;;一種新的半監(jiān)督支持向量機(jī)[J];鄭州大學(xué)學(xué)報(bào)(理學(xué)版);2012年03期
相關(guān)會(huì)議論文 前10條
1 余樂安;姚瀟;;基于中心化支持向量機(jī)的信用風(fēng)險(xiǎn)評(píng)估模型[A];第六屆（2011）中國(guó)管理學(xué)年會(huì)——商務(wù)智能分會(huì)場(chǎng)論文集[C];2011年
2 劉希玉;徐志敏;段會(huì)川;;基于支持向量機(jī)的創(chuàng)新分類器[A];山東省計(jì)算機(jī)學(xué)會(huì)2005年信息技術(shù)與信息化研討會(huì)論文集（一）[C];2005年
3 史曉濤;劉建麗;駱玉榮;;一種抗噪音的支持向量機(jī)學(xué)習(xí)方法[A];全國(guó)第19屆計(jì)算機(jī)技術(shù)與應(yīng)用（CACIS）學(xué)術(shù)會(huì)議論文集（下冊(cè)）[C];2008年
4 何琴淑;劉信恩;肖世富;;基于支持向量機(jī)的系統(tǒng)辨識(shí)方法研究及應(yīng)用[A];中國(guó)力學(xué)大會(huì)——2013論文摘要集[C];2013年
5 劉駿;;基于支持向量機(jī)方法的衢州降雪模型[A];第五屆長(zhǎng)三角氣象科技論壇論文集[C];2008年
6 王婷;胡秀珍;;基于組合向量的支持向量機(jī)方法預(yù)測(cè)膜蛋白類型[A];第十一次中國(guó)生物物理學(xué)術(shù)大會(huì)暨第九屆全國(guó)會(huì)員代表大會(huì)摘要集[C];2009年
7 趙晶;高雋;張旭東;謝昭;;支持向量機(jī)綜述[A];全國(guó)第十五屆計(jì)算機(jī)科學(xué)與技術(shù)應(yīng)用學(xué)術(shù)會(huì)議論文集[C];2003年
8 周星宇;王思元;;智能數(shù)學(xué)與支持向量機(jī)[A];2005年中國(guó)智能自動(dòng)化會(huì)議論文集[C];2005年
9 顏根廷;馬廣富;朱良寬;宋斌;;一種魯棒支持向量機(jī)算法[A];2006中國(guó)控制與決策學(xué)術(shù)年會(huì)論文集[C];2006年
10 侯澍e，
本文編號(hào)：2214858
本文鏈接：http://sikaile.net/kejilunwen/dizhicehuilunwen/2214858.html
最近更新
教材專著
