Please wait a minute...
img

官方微信

遥感技术与应用  2011, Vol. 26 Issue (3): 340-347    DOI: 10.11873/j.issn.1004-0323.2011.3.340
图像与数据处理     
并联结构组合分类器的误差分析
郑忠,曾永年,刘慧敏,徐艳艳,于菲菲
(中南大学地球科学与信息物理学院,中南大学空间信息技术与可持续发展研究中心,湖南 长沙410083)
The Errors Analysis of Combined Classifier based on Parallel Structure
ZHENG Zhong,ZENG Yong-nian,LIU Hui-min,XU Yan-yan,YU Fei-fei
(School of Geosciences and Geomatics,Center for Geomatics and Sustainabal Development Research,
Central South University,Changsha 410083,China)
 全文: PDF(4985 KB)  
摘要:

虽然遥感图像分类器发展迅速,但是单分类器分类精度仍然不能满足实际应用的需要,因此组合分类器便成为遥感分类技术研究的一个重要方面。与串联结构相比,并联结构是实际应用过程中研究较早、发展较充分和应用最广泛的,因而主要就并联结构组合分类器的误差进行分析。通过理论分析得出:并联结构组合分类器的精度变化与单分类器的误差集合分布模式有关。相离时,其精度提升幅度最大;相交时,组合分类器精度得到提升,其精度提升幅度大小与组合后的误差像元集合的大小成反比;相包含时,其精度位于两单分类器之间,且更靠近精度较高的单分类器的精度。同时以长沙市局部区域为实验区进行了验证性的实验,实验结果有效地验证了并联结构组合分类器的误差分析所得到的推论。最后从理论上讨论了组合分类器提高遥感图像分类精度的可行性,并指出单个类别的组合分类结果也与单分类器对这个类别的误差集合分布模式有关,为组合分类器发展提供了一个较好的突破方向。

关键词: 组合分类器分类精度土地利用Landsat TM    
Abstract:

The classification algorithm of remote sensing image rapidly develops,as the classification accuracy of single classifier still cant meet the needs of practical application,the combined classifier becames an important aspect of the remote sensing classification.There are various types of combined classifiers.Compared with serial structure,Parallel structure is early studied ,fully developed and widely used.This paper analyzed classification accuracy of the combined classifiers based on parallel structure.The results indicated that the precision of combined classifier is related to the position of incorrectly classified pixels in each single.While the incorrectly classified pixels by single classifiers are separated ,the accuracy of combined classifier is the highest;while the incorrectly classified pixels by single classifiers are intersected in classified results,the precision of combined classifier is higher than that of each single classifier and the improvement is inversely proportional to the size of error set in combined classified results;while the incorrectly classified pixels by one classifier are included in another classifier,the accuracy of combined classifier is located in between the high and low accuracy,which is near to the higher one.This paper also experiments in Changsha local area,and the result effectively tested the deduction which we get in the analysis.Finally,this paper discussed theoretically the feasibility of improvement which combined classifier can get in the remote sensing image classification and provides an effective way to improve the performance of combined classifier.

Key words: Combined classifier    Classification accuracy    Land use    Landsat TM
收稿日期: 2011-02-17 出版日期: 2013-01-23
:  TP 75  
基金资助:

国家自然科学基金项目(40771198),湖南省自然科学基金项目(08JJ6023),中南大学研究生学位论文创新基金。

通讯作者: 曾永年(1959- ),男,青海西宁人,博士,教授,主要从事遥感与地理信息系统的应用研究及教学工作。Email:ynzeng@mail.csu.edu.cn。     E-mail: Middle880104@139.com
作者简介: 郑忠(1988- ),男,四川广元人,硕士研究生,主要从事遥感应用研究。Email:Middle880104@139.com。
服务  
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章  
郑忠
曾永年
刘慧敏
徐艳艳
于菲菲

引用本文:

郑忠,曾永年,刘慧敏,徐艳艳,于菲菲. 并联结构组合分类器的误差分析[J]. 遥感技术与应用, 2011, 26(3): 340-347.

ZHENG Zhong,ZENG Yong-nian,LIU Hui-min,XU Yan-yan,YU Fei-fei. The Errors Analysis of Combined Classifier based on Parallel Structure. Remote Sensing Technology and Application, 2011, 26(3): 340-347.

链接本文:

http://www.rsta.ac.cn/CN/10.11873/j.issn.1004-0323.2011.3.340        http://www.rsta.ac.cn/CN/Y2011/V26/I3/340

[1]Liu Xiaoping,Li Xia,He Jinqiang,et al.Classification of Remote Sensing Images based on Ant Colony Optimization[J].Journal of Remote Sensing,2008,12(2):253-262.[ 刘小平,黎夏,何晋强,等.基于蚁群智能的遥感影像分类新方法[J].遥感学报,2008,12(2):253-262.]

[2]Liu Xiaoping,Peng Xiaojuan,Ai Bin.An Automatic Classification Model of Remote\|Sensing Image based on Pixel Information Decomposition Combined with Decision Tree[J].Geography and Geo Information Science,2004,20(6):35-39.[刘小平,彭晓鹃,艾彬.像元信息分解和决策树相结合的影像分类方法[J].地理与地理信息科学,2004,20(6):35-39.]

[3]Lu D,Weng Q.A Survey of Image Classification Methods and Techniques for Improving Classification Performance[J].International Journal of Remote Sensing,2007,28(5):823-870.

[4]Schowengerdt R A.Remote Sensing:Models and Methods for Image Processing[M].Burlington:Academic Press,1997.

[5]Richards J A.Remote Sensing Digital Image Analysis[M].Berlin:Springer Verlag,1998.

[6]Bo Yanchen,Wang Jinfeng.Combining Multiple Classifiers for Thematic Classification of Remotely Sensed Data[J].Journal of Remote Sensing,2005,9(5):555-563.[ 柏延臣,王劲峰.结合多分类器的遥感数据专题分类方法研究[J].遥感学报,2005,9(5):555-563.]

[7]Liu X H,Skidmore A K,Van Oosten H.Integration of Classification Methods for Improvement of Land\|cover Map Accuracy[J].ISPRS Journal of Photogrammetry & Remote Sensing,2002,56(4):257-268.

[8]Matsuyama T.Knowledge-based Aerial Image Understanding Systems and Expert Systems for Image Processing[J].IEEE Transactions on Geoscience and Remote Sensing,1987,GE-25(3):305-316.

[9]Lv Yue,Shi Pengfei,Zhao Yuming.Voting Principle for Combination of Multiple Classifiers[J].Journal of Shanghai Jiaotong University,2000,34(5):680-683.[ 吕岳,施鹏飞,赵宇明.多分类器组合的投票表决规则[J].上海交通大学学报,2000,34(5):680-683.]

[10]Lu D,Weng Q.Spectral Mixture Analysis of the Urban Landscapes in Indianapolis with Landsat ETM+ Imagery[J].Photogrammetric Engineering and Remote Sensing,2004,70(9):1053-1062.

[11]Huang Z,Lees B G.Combining Non\|parametric Models for Multisource Predictive Forest Mapping[J].Photogrammetric Engineering and Remote Sensing,2004,70(4),415-425.

[12]Zhou Hongying,Lin Qizhong,Wu Yunzhao,et al.Application of Combining Classifier based on AdaBoost to Remote Sensing Classification[J].Application Research of Computers,2007,24(10):181-184.[周红英,蔺启忠,吴昀昭,等.基于AdaBoost的组合分类器在遥感影像分类中的应用[J].计算机应用研究,2007,24(10):181-184.]

[13]Zhou Yi,Jing Fengguang.The Classify Method of Remote Sensing Image by Multiple Classifier Combination[J].Urban Geotechnical Investigation & Surveying,2008,(2):56-59. [周夷,景奉广.多分类器组合的遥感图像分类的方法[J].城市勘测,2008,(2):56-59.]

[14]Chen Xuehong,Chen Jin,Yang Wei,et al.Study on Combined Classifier based on Error Analysis[J].Journal of Remote Sensing,2008,12(5):683-691.[ 陈学泓,陈晋,杨伟,等.基于误差分析的组合分类器研究[J].遥感学报,2008,12(5):683-691.]

[15]Zhang Dan,Yang Bin,Zhang Ruiyu.Multiple Classifiers Combination for Remote Sensing Image Classification[J].Remote Sensing Information,2009,5:41-43. [张丹,杨斌,张瑞禹.多种分类器融合的遥感影像分类[J].遥感信息,2009,5:41-43.]

[16]Gon alves L M S,Fonte C C,Caetano M.Using Uncertainty Information to Combine Soft Classifications[C]//13th International Conference on Information Processing and Management of Uncertainty.Berlin:Springer\|Verlag,2010,455-463.

[17]Maulik U,Chakraborty D.A Robust Multiple Classifier System for Pixel Classification of Remote Sensing Images[J].Fundamenta Informaticae,2010,101(4):286-304.

[18]Sun Hao,Du Peijun,Zhao Weichang.Target Identification from High Resolution Remote Sensing Image based on Multiple Classifier Combination[J].Geography and Geo\|Information Science,2009,25(1):32-35.[孙灏,杜培军,赵卫常.基于多分类器组合的高分辨率遥感影像目标识别[J].地理与地理信息科学,2009,25(1):32-35.]

[19]Li Chungan.Study on the Forest Sub-compartment Map Updating Technology with SPOT Imagery[D].Beijing:Beijing Forestry University.2009.[李春干.面向对象的SPOT5图像森林分类研究[D].北京:北京林业大学,2009.]

[20]Lu Y.Knowledge Integration in a Multiple Classifier System[J].Applied Intelligence,1996,6(2):75-86.

[21]De Stefano C,DElia C,Marcell A,et al.Using Bayesian Network for Combining Classifiers[C]//14th International Conference on Image Analysis and Processing\|ICIAP 2007.Washington:IEEE Computer Society,2007:73-80.

[22]Han Jianfeng,Yang Zhehai.Combined Classifiers and Its Application in Hyperspectral Classification[J].Journal of Zhengzhou Institute of Surveying and Mapping,2007,24(3):231-234.[韩建峰,杨哲海.组合分类器及其在高光谱影像分类中的应用[J].测绘科学技术学报,2007,24(3):231-234.]

[23]Zhang Shaojia.Remote Sensing Classification by Combing Mu\|ltiple Classifiers[D].Changsha:Central South University,2010.[ 张少佳.多分类器组合及其遥感分类研究[D].长沙:中南大学,2010.]

[24]Fu Bin.Multiple Classifier Portfolio based on Classifier Selection Method[D].Beijing:Beijing Jiaotong University,2009.[付彬.多分类器组合中的基分类器选取方法[D].北京:北京交通大学,2009.]

[25]Gong Dapeng,Li Zhixin,Han Jianfeng,et al.Hyperspectral Classification and the Utilization of Combined Classifiers[J].Hydrographic Surveying and Charting,2006,26(5):20-23.[宫大鹏,李之歆,韩建峰,等.高光谱影像分类及组合分类器的应用[J].海洋测绘,2006,26(5):20-23.]

[1] 胡云锋,商令杰,张千力,王召海. 基于GEE平台的1990年以来北京市土地变化格局及驱动机制分析[J]. 遥感技术与应用, 2018, 33(4): 573-583.
[2] 赵梦雨,薛亮. 咸阳市生境质量变化遥感监测研究[J]. 遥感技术与应用, 2017, 32(6): 1171-1180.
[3] 陈洋波,张涛,窦鹏,董礼明,陈华. 基于SVM的东莞市土地利用/覆被自动分类误差来源与后处理[J]. 遥感技术与应用, 2017, 32(5): 893-903.
[4] 曹广真,卢乃锰,侯 鹏,李贵才. 遥感湖泊面积MTMF提取方法及空间尺度效应研究[J]. 遥感技术与应用, 2016, 31(5): 864-871.
[5] 朱爽,张锦水. 样本特征对参数/非参数分类器分类精度的影响分析[J]. 遥感技术与应用, 2016, 31(4): 748-755.
[6] 谢 瀚,李爱农,汤家法,雷光斌,边金虎. 山地地表覆被野外采样系统研究及应用[J]. 遥感技术与应用, 2016, 31(3): 430-437.
[7] 管小彬,沈焕锋,甘文霞,张良培. 基于Landsat TM/ETM+影像的武汉市冬季NPP估算及其时空变化分析[J]. 遥感技术与应用, 2015, 30(5): 884-890.
[8] 张文博,覃志豪,刘含海. 基于分形理论的土地利用类型变化研究—以广州番禺为例[J]. 遥感技术与应用, 2015, 30(5): 952-958.
[9] 白晓燕,陈晓宏,王兆礼. 基于面向对象分类的土地利用信息提取及其时空变化研究[J]. 遥感技术与应用, 2015, 30(4): 798-809.
[10] 梁治华,刘勇. 近20年来腾格里沙漠南缘土地利用/土地覆盖变化及其成因探讨—以甘肃省古浪县冰草湾地区为例[J]. 遥感技术与应用, 2015, 30(2): 383-390.
[11] 宋晓阳,姜小三,江东,黄耀欢,万华伟,王昌佐. 基于面向对象的高分影像分类研究[J]. 遥感技术与应用, 2015, 30(1): 99-105.
[12] 韩婷婷,习晓环,王成,包玉海,骆社周. 基于决策树方法的云南省森林分类研究[J]. 遥感技术与应用, 2014, 29(5): 744-751.
[13] 郭钇宏,王博,刘勇,杨亦宁 . 综合优度法和不一致性法的最优分割参数选择方法[J]. 遥感技术与应用, 2014, 29(3): 489-497.
[14] 王少伟,张晓祥,杨晓英. 太湖湖滨敏感区的土地利用遥感分类研究[J]. 遥感技术与应用, 2014, 29(1): 114-121.
[15] 韦春竹,孟庆岩,郑文锋,李小江,魏曦,王靓. 广州市地表温度反演与土地利用覆盖变化关系研究[J]. 遥感技术与应用, 2013, 28(6): 955-963.