基于RS的浙江省余杭区森林与湿地动态研究
摘要
森林和湿地两大生态系统分别作为“地球之肺”和“地球之肾”,共同为地球构筑了强大的生态安全屏障。同时两者之间息息相关、相互依存,他们的动态变化常常是具有一定的联系性,因而只研究其中一种的变化是片面的。建设和保护好森林与湿地生态系统是林业部门承担的重要职责,而森林与湿地的动态监测则是森林与湿地保护的一项重要内容。本文以浙江省余杭区为研究区,利用1994年与2008年TM影像,探索森林与湿地动态监测的方法以及余杭区从1994至2008年将近二十年内森林与湿地的变化情况。研究结果表明：
(1)本文主要采用三种变化监测的方法,包括分类后比较法、多时相数据直接分类技术、变化信息的提取技术。分类后比较法的监测精度主要取决于分类方法的选择。本文引入纹理信息,对两期数据分别进行最大似然法和决策树法进行分类,结果显示后者的精度高于前者。最后通过类别叠加分析得到变化类别和变化统计结果。多时相数据直接分类法相对前者合理性较强,减少出现实际不存在变化类别,并且只进行一次分类。难点在于变化类别的确定,有时相当繁琐。变化信息的提取技术主要采用图像差值法和变化向量分析法,难点在于变化阈值的确定,本文采用基于Kappa系数的阈值计算方法以及双窗口变步长阈值搜寻算法来确定阈值,并且选取精度最高的变化—未变化二值图,最后再对变化信息进行分类。
(2)变化结果显示：1994年森林占地40146.66hm2；2008年森林占地41618.97 hm2；1994年湿地占地4380.75hm2；2008年湿地占地9680.85hm2。从1994年到2008年,森林覆盖较为稳定,而湿地覆盖面积显著增加。森林一部分转化为农田以及建筑用地,增加部分由农田转化而来；湿地的增加主要来自于农田的转化。
(3)对两期森林状况采用Riitters模型进行景观破碎化分析,1994年到2008年,尽管森林覆盖面积有所上升,但森林内部破碎化程度加重,2008年穿孔森林是1994年的7.32倍、斑块森林是1994年的1.84倍、过渡森林是1994年的1.27倍。造成森林破碎化程度加剧的主要驱动力为人为干扰。根据AI指数的分析,从1994年到2008年随着建筑用地的急剧扩张,对森林的干扰逐渐上升,而人为干扰中农田干扰则有逐渐减弱的趋势。
(4)从1994年到2008年,余杭区人工湿地增加明显而天然湿地比重下降。研究区内人工湿地主要包括水库及养殖水面,人工湿地的破碎化剧烈升高,主要由于余杭区重视水产养殖的发展,积极发展渔业,人工养殖水面相应增加且彼此之间缺少连通性。
Forest and wetland as two big ecological systems of the earth are regarded as "lung" and "kidneys" of the earth.They are closely related and interdependent and construct a powerful ecological security barrier for the earth.The dynamic changes of them have a certain relation, therefore,only researching one change is not comprehensive.It is an important responsibility of foestry department to Construct and protect forests and wetlands ecosystems, and the dynamic monitoring forests and wetlands is the important content of forest and the wetland protection.Using TM images data of 1994 and 2008, this paper aims at comparing different change detection techniques to assess their effectiveness for detecting changes of forests and wetlands from 1994 to 2008 inYuhang,Zhejiang provice. Following are the conclusions from this study:
(1)This paper describes three change detection methods, including the post-classification comparison method,the multitemporal composition classification and the change detection techniques based on pixels.The detection accuracy of post-clsssification method depends on the choice of classification methods. Taking advantage of the approach of texture analysis,this study compares Maximum likelihood classification and decision tree classification and the result shows that the accuracy of the latter is better than the former. Finally, we get change category and change statistical results by overlay analysis. Compared to post-classification comparison method, multitemporal composition classification,which is more reasonable, only requires classification once and reduces the occurrence of the Unrealistic changes. But it is difficult to determinate the change category. Change detection techniques based on pixels mainly use the image difference method and the Change Vector Analysis(CVA), which is difficult to determine the appropriate changing threshold. In this paper, we choose Double Thresholding algorithm based on the Kappa and a new method named Double-Windows Flexible Pace Searching algorithm to determine image segmentation threshold, and select the most accurate change binary image,to classify
(2)The changed results show that the area of forest and Wetland were 40146.66 hm2 and 4380.75hm2 in 1994 while there were 41618.97 hm2 and 9680.85 hm2 in 2008. From 1994 to 2008, forest cover was more stable,while wetland coverage area increased significantly. Part of the forest changed into farmland or building lands,changed from farmland. The increase of Wetland is mainly because of farmland transformation.
(3) By Riitters Model, Compared to 1994, although the overall forest cover rate in Yuhang District has increased in 2008, forest fragmentation has accelerated due to a decreased proportion of interior forest and an increased proportion of perforated, patch and transitional forests. Perforated forest in 2008 was 7.32 times that of 1994, patch forest was 1.84 times that of 1994 and transitional forest was 1.27 times that of 1994. That human disturbance accelerates forest fragmentation from the 1980s to 2008.According to the analysis of Association Index, Rapid urbanization during 1994 to 2008 is considered to be the main driving force for this deterioration,while the interference of field is a gradual weakening trend.
(4)From 1994 to 2008, artificial wetland increased obviously and natural wetland proportion dropped in yuhang district. Artificial wetland mainly includes reservoirs and aquaculture, and artificial wetland fragmentation dramatically rise, mainly due to the development of attention for aquaculture in yuhang district,and water aquaculture increase but lack of connectedness.
(1)本文主要采用三种变化监测的方法,包括分类后比较法、多时相数据直接分类技术、变化信息的提取技术。分类后比较法的监测精度主要取决于分类方法的选择。本文引入纹理信息,对两期数据分别进行最大似然法和决策树法进行分类,结果显示后者的精度高于前者。最后通过类别叠加分析得到变化类别和变化统计结果。多时相数据直接分类法相对前者合理性较强,减少出现实际不存在变化类别,并且只进行一次分类。难点在于变化类别的确定,有时相当繁琐。变化信息的提取技术主要采用图像差值法和变化向量分析法,难点在于变化阈值的确定,本文采用基于Kappa系数的阈值计算方法以及双窗口变步长阈值搜寻算法来确定阈值,并且选取精度最高的变化—未变化二值图,最后再对变化信息进行分类。
(2)变化结果显示：1994年森林占地40146.66hm2；2008年森林占地41618.97 hm2；1994年湿地占地4380.75hm2；2008年湿地占地9680.85hm2。从1994年到2008年,森林覆盖较为稳定,而湿地覆盖面积显著增加。森林一部分转化为农田以及建筑用地,增加部分由农田转化而来；湿地的增加主要来自于农田的转化。
(3)对两期森林状况采用Riitters模型进行景观破碎化分析,1994年到2008年,尽管森林覆盖面积有所上升,但森林内部破碎化程度加重,2008年穿孔森林是1994年的7.32倍、斑块森林是1994年的1.84倍、过渡森林是1994年的1.27倍。造成森林破碎化程度加剧的主要驱动力为人为干扰。根据AI指数的分析,从1994年到2008年随着建筑用地的急剧扩张,对森林的干扰逐渐上升,而人为干扰中农田干扰则有逐渐减弱的趋势。
(4)从1994年到2008年,余杭区人工湿地增加明显而天然湿地比重下降。研究区内人工湿地主要包括水库及养殖水面,人工湿地的破碎化剧烈升高,主要由于余杭区重视水产养殖的发展,积极发展渔业,人工养殖水面相应增加且彼此之间缺少连通性。
Forest and wetland as two big ecological systems of the earth are regarded as "lung" and "kidneys" of the earth.They are closely related and interdependent and construct a powerful ecological security barrier for the earth.The dynamic changes of them have a certain relation, therefore,only researching one change is not comprehensive.It is an important responsibility of foestry department to Construct and protect forests and wetlands ecosystems, and the dynamic monitoring forests and wetlands is the important content of forest and the wetland protection.Using TM images data of 1994 and 2008, this paper aims at comparing different change detection techniques to assess their effectiveness for detecting changes of forests and wetlands from 1994 to 2008 inYuhang,Zhejiang provice. Following are the conclusions from this study:
(1)This paper describes three change detection methods, including the post-classification comparison method,the multitemporal composition classification and the change detection techniques based on pixels.The detection accuracy of post-clsssification method depends on the choice of classification methods. Taking advantage of the approach of texture analysis,this study compares Maximum likelihood classification and decision tree classification and the result shows that the accuracy of the latter is better than the former. Finally, we get change category and change statistical results by overlay analysis. Compared to post-classification comparison method, multitemporal composition classification,which is more reasonable, only requires classification once and reduces the occurrence of the Unrealistic changes. But it is difficult to determinate the change category. Change detection techniques based on pixels mainly use the image difference method and the Change Vector Analysis(CVA), which is difficult to determine the appropriate changing threshold. In this paper, we choose Double Thresholding algorithm based on the Kappa and a new method named Double-Windows Flexible Pace Searching algorithm to determine image segmentation threshold, and select the most accurate change binary image,to classify
(2)The changed results show that the area of forest and Wetland were 40146.66 hm2 and 4380.75hm2 in 1994 while there were 41618.97 hm2 and 9680.85 hm2 in 2008. From 1994 to 2008, forest cover was more stable,while wetland coverage area increased significantly. Part of the forest changed into farmland or building lands,changed from farmland. The increase of Wetland is mainly because of farmland transformation.
(3) By Riitters Model, Compared to 1994, although the overall forest cover rate in Yuhang District has increased in 2008, forest fragmentation has accelerated due to a decreased proportion of interior forest and an increased proportion of perforated, patch and transitional forests. Perforated forest in 2008 was 7.32 times that of 1994, patch forest was 1.84 times that of 1994 and transitional forest was 1.27 times that of 1994. That human disturbance accelerates forest fragmentation from the 1980s to 2008.According to the analysis of Association Index, Rapid urbanization during 1994 to 2008 is considered to be the main driving force for this deterioration,while the interference of field is a gradual weakening trend.
(4)From 1994 to 2008, artificial wetland increased obviously and natural wetland proportion dropped in yuhang district. Artificial wetland mainly includes reservoirs and aquaculture, and artificial wetland fragmentation dramatically rise, mainly due to the development of attention for aquaculture in yuhang district,and water aquaculture increase but lack of connectedness.
引文
[1]刑元军,旦增.基于面向对象的高分辨率影像的森林植被变化信息提取研究[J].中南林业调查规划,2011,29(4)：34-37.
[2]张涵韬.湿地生态环境动态监测模型的研究[J].林业勘察设计(福建),2008(1)：173-174.
[3]冯德俊,缪文,兰燕.遥感动态监测自动发现方法研究现状及发展趋势[J].四川测绘,2007,30(4)：147-151.
[4]李敏.基于模糊C均值算法的遥感图像[D].湖南：湖南大学,2009.
[5]王海波,马明国.基于遥感的湖泊水域动态变化监测研究进展[J].遥感技术与应用,2009,24(5)：674-684.
[6]Chen J, Gong P, He C Y, et al. Land-use/land-cover change detection using improved change vector analysis. Photogrammetric Engineering & Remote Sensing,2003,69(40):369-379.
[7]杨存建,欧晓昆,党承林,张增祥.森林植被动态变化信息的遥感监测[J].地球信息科学.2000(4)：71-74.
[8]陈振清,董玉娟,李久刚.基于模板阈值法的土地利用变化信息提取方法研究[J].遥感应用,2009.3：64-67.
[9]孙晓霞,张继贤.土地利用动态遥感监测的误差分析[J].测绘科学,2003,28(4)：45-46.
[10]张静.遥感影像变化监测算法及应用研究[D].陕西：西安电子科技大学,2009.
[11]朱朝杰,基于特征分析的遥感影像变化监测方法研究[D].中国人民解放军信息工程大学,2007.
[12]Silapaswan,C.S.,Verbyla,D.L., and Mcguire, A.D.,Land cover change on the Seward Peninsula:the use of remote sensing to evaluate the potential influences of climate warming on historical vegetation dynamics.Canadian Journal of Remote Sensing,2001,27:542-554.
[13]Petit, C, Scudder, T., and Lambin, E.,Quantifying processes of land-cover change by remote sensing: resettlement and rapid land-cover change in southeastern Zambia International Journal of Remote Sensing,2001,22:3435-3456.
[14]Jensen J R. Introductory digital image processing, a remote sensing perspective.Upper Saddle River, New Jersey:Prentice Hall,2004.
[15]漆星,许江涛,郝成元,王海宁.基于GIS空间分析的孟州市土地利用/覆盖变化研究[J].2009,37(2)：730-731,750.
[16]张怡春.基于GIS的汤旺河林业局森林资源动态分析[D].黑龙江：东北林业大学,2006.
[17]徐天蜀.森林资源动态监测技术研究——以云南省高黎贡山自然保护区为例[D].云南省：西南林学院,2008.
[18]李向军,遥感土地利用变化监测方法探讨[D]：博士论文.北京：中国科学院研究生院.2006.
[19]张友水,冯学智,周成虎.多时相TM影像相对辐射校正研究[J].测绘学报,2006,35(2)：122-127.
[20]彭望碌.遥感与图像解译(第四版)[M].北京：电子工业出版社,2003.8：335-338.
[21]阿布都瓦斯提.吾拉木,秦其明,朱黎江.基于6S模型的可见光、近红外遥感数据的大气校正[J].北京大学学报(自然科学版),2004,40(4)：611-618.
[22]李海鹰,苗放,孔祥生,胡玉枝,李海燕.基于6S模型的土法炼焦污染区TM遥感图像大气校正[J].内蒙古石油化工,2006(10)：140-144.
[23]韩改新.一种新的遥感影像变化监测相对辐射校正方法[J].测绘与空间地理信息,2008,31(3)：84-86,
[24]丁丽霞,周斌,王人潮.遥感监测中5种相对辐射校正方法研究[J].浙江大学学报,2005,31(3)： 269-276.
[25]易善祯,王志辉.不同方法在水体遥感提取中的应用实现(J).红外,28(4)：1-9.
[26]王刚,李小曼,田杰。几种TM影像的水体自动提取方法比较[J].测绘科学,2008,33(3)：141-142.
[27]肖玉环,黄伦春,刘晓燕,王媛媛,张红英.基于HIS变换的多源遥感影像融合方法研究[J].工程地球物理学报,2010,7(2)：248-252.
[28]李荣,梁学强,曾丹.遥感影像数据几种HIS融合方法的比较[J].科技经济市场,2009(10)：4-5.
[29]梁玉喜.TM图像用于川西地区森林面积自动提取的研究[D].四川农业大学,2004.
[30]冯德俊,李永树,兰燕.基于主成分变换的动态监测变化信息自动发现[J].计算机工程与应用,2004(36)：199-202.
[31]高程程,惠晓惠.基于灰度共生矩阵的纹理特征提取[J].计算机系统应用,2010,19(6)：195-198.
[32]宁顺刚,白万民,喻钧.基于灰度共生矩阵的图像分割方法研究[J].图像.编码与软件,2009,22(11)：69-71.
[33]苑丽红,付丽,杨勇,苗静.灰度共生矩阵提取纹理特征的实验结果分析[J].计算机应用,2009,29(4)：1018-1021.
[34]冯建辉,杨玉静.基于灰度共生矩阵提取纹理特征图像的研究[J].北京测绘,2007(3),1922.
[35]侯海苗,冀小平.基于灰度共生矩阵的纹理特征[J].长治学院学报,2008,25(5)：31-32.
[36]谭颖.翁源县基于纹理信息及CART决策树技术的遥感影像分类研究[D].江苏：南京林业大学,2008.
[37]郭德俊,宋蜇存.基于灰度共生矩阵的纹理图像分类研究[J].林业机械与木工设备,2005,33(7)：21-23.
[38]薄华,马缚龙,焦李成.图像纹理的灰度共生矩阵计算问题的分析[J].电子学报,2006(1)：155-157.
[39]刘淼,秦大庸,刘家宏,吴发启.基于NDVI的山西省植被覆盖度变化研究[J].人民黄河,2009,31(5)：17-18.
[40]许定成,游先祥,韩熙春.中国遥感进展[M].北京：万国学术出版社,1992.205-210.
[41]FasART模糊神经网络用于遥感图像监督分类的研究[J].中国图象图形学报,2002,7(12)：1263-1268.
[42]陈华,陈书海,张平,严卫东K.means算法在遥感分类中的应用[J].红外与激光工程,2000,29(2)：26-30.
[43]刘鹏,姚正,尹俊杰.一种有效的C4.5改进模型[J].清华大学学报(自然科学版).2006.46(S1)：996-1001.
[44]PalM, MatherPM. An assessment of the effectiveness of decision tree methods of land cover classification [J]. Remote Sensing of Environment,2003 (86):554-565.
[45]余晶,蒋平安,高敏华.基于决策树的土地利用分类方法研究[J].新疆农业科学,2009,46(2)：430-434.
[46]罗开华.遥感图像分类后处理3中实现方法比较[J].林业调查规划,2008,33(2)：24-25.
[47]周斌,杨柏林.运用多时相直接分类法对土地利用进行遥感动态监测的研究[J].自然资源学报2001,16(3)：263-268.
[48]R.N.Siwe.B.Koch.Change vector analysis to categorise land cover change processes using the tasselled cap
as biophysical indicator[J]. Environ Monit Assess (2008) 145:227-235.
[49]William A. Malila.Change Vector Analysis:An Approach for Detecting Forest Changes with Landsat[J]. Laboratory for Applications of Remote Sensing(1980):326-335.
[50]S.Berberoglu,A.Akin.Assessing different remote sensing techniques to detect land use/cover changes in the eastern Mediterranean[J]. International Journal of Applied Earth Observation and Geoinformation (2009) 11:46-53.
[51]邓小炼.基于变化矢量分析的土地利用变化监测方法研究[D]：博士论文.北京：中国科学院,2006.
[52]邸凤萍,朱重光,丁玲.方向矢量法在城市土地利用变化监测中的应用[J].计算机工程,2008,34(2)：253-254.
[53]史榕,许惠平,陈华根.遥感影像数据融合技术及融合质量评价研究[J].鲁东大学学报(自然科学版),2008,24(2)：130-134.
[54]孙家抦等.遥感原理与应用[M].武汉：武汉大学出版社.2009.
[55]陈晋,何春阳,史培军,陈云浩,马楠.基于变化向量分析的土地利用/覆盖变化动态监测---变化阈值的确定方法[J].遥感学报,2001,5(4)：259-266.
[56]郭徵,江洪,陈健,程苗苗,江子山,余树全,李土生.基于遥感的杭州余杭森林景观格局变化[J].浙江林学院学报,2010,27(1)：36-43.
[57]邓劲松,李君,余亮,等.快速城市化过程中杭州市土地利用景观格局动态[J].应用生态学报,2008,19(9)：2003-2008.
[58]Wade T G, Riitters K H, Wickham J D, et al. Distribution and causes of Global forest fragmentation[J/OL].Conservation Ecology,2003,7(2):7[2009].
[59]Riitters K H,Wickham J D,O'Neill R V,et al.Global-Scale Patterns of Forest Fragmentation[J]. ConservationEcology,2000,4(2):3.
[60]方晰,唐代生,杨乐,等.湖南省林科院试验林场森林植被景观格局及破碎化分析[J].中南林业科技大学学报,2008,28(4)：107-112.
[61]Abdullah S A,Nakagoshi N.Changes in agricultural landscape pattern and its spatial relationship with forestland in the State of Selangor,peninsular Malaysia[J].Landscape and Urban Planning,2008,87:147-155.
[62]Abdullah S A,Nakagoshi N.Forest fragmentation and its correlation to human land use Change in the state of Selangor,peninsular Malaysia[J].Forest Ecology and Management,2007,241:39-48.
[63]宗秀影,刘高焕,乔玉良,林松.黄河三角洲湿地景观格局动态变化分析[J].地球信息科学学报,2009,11(1)：91-97.
[64]唐继刚.南京市湿地资源动态遥感监测与分析[J].ecological economy:40-45.
[65]周昕薇.基于3S技术的北京湿地动态监测与评价方法研究[D].首都师范大学,2006.
[66]邓书斌,武红敢,江涛.遥感动态监测中的相对辐射校正方法研究[J].遥感应用,2008.4,71-75.
[67]梅洋,陆苗.基于遥感影像的变化监测研究动态[J].地理信息世界,2009.04：42-47.
[2]张涵韬.湿地生态环境动态监测模型的研究[J].林业勘察设计(福建),2008(1)：173-174.
[3]冯德俊,缪文,兰燕.遥感动态监测自动发现方法研究现状及发展趋势[J].四川测绘,2007,30(4)：147-151.
[4]李敏.基于模糊C均值算法的遥感图像[D].湖南：湖南大学,2009.
[5]王海波,马明国.基于遥感的湖泊水域动态变化监测研究进展[J].遥感技术与应用,2009,24(5)：674-684.
[6]Chen J, Gong P, He C Y, et al. Land-use/land-cover change detection using improved change vector analysis. Photogrammetric Engineering & Remote Sensing,2003,69(40):369-379.
[7]杨存建,欧晓昆,党承林,张增祥.森林植被动态变化信息的遥感监测[J].地球信息科学.2000(4)：71-74.
[8]陈振清,董玉娟,李久刚.基于模板阈值法的土地利用变化信息提取方法研究[J].遥感应用,2009.3：64-67.
[9]孙晓霞,张继贤.土地利用动态遥感监测的误差分析[J].测绘科学,2003,28(4)：45-46.
[10]张静.遥感影像变化监测算法及应用研究[D].陕西：西安电子科技大学,2009.
[11]朱朝杰,基于特征分析的遥感影像变化监测方法研究[D].中国人民解放军信息工程大学,2007.
[12]Silapaswan,C.S.,Verbyla,D.L., and Mcguire, A.D.,Land cover change on the Seward Peninsula:the use of remote sensing to evaluate the potential influences of climate warming on historical vegetation dynamics.Canadian Journal of Remote Sensing,2001,27:542-554.
[13]Petit, C, Scudder, T., and Lambin, E.,Quantifying processes of land-cover change by remote sensing: resettlement and rapid land-cover change in southeastern Zambia International Journal of Remote Sensing,2001,22:3435-3456.
[14]Jensen J R. Introductory digital image processing, a remote sensing perspective.Upper Saddle River, New Jersey:Prentice Hall,2004.
[15]漆星,许江涛,郝成元,王海宁.基于GIS空间分析的孟州市土地利用/覆盖变化研究[J].2009,37(2)：730-731,750.
[16]张怡春.基于GIS的汤旺河林业局森林资源动态分析[D].黑龙江：东北林业大学,2006.
[17]徐天蜀.森林资源动态监测技术研究——以云南省高黎贡山自然保护区为例[D].云南省：西南林学院,2008.
[18]李向军,遥感土地利用变化监测方法探讨[D]：博士论文.北京：中国科学院研究生院.2006.
[19]张友水,冯学智,周成虎.多时相TM影像相对辐射校正研究[J].测绘学报,2006,35(2)：122-127.
[20]彭望碌.遥感与图像解译(第四版)[M].北京：电子工业出版社,2003.8：335-338.
[21]阿布都瓦斯提.吾拉木,秦其明,朱黎江.基于6S模型的可见光、近红外遥感数据的大气校正[J].北京大学学报(自然科学版),2004,40(4)：611-618.
[22]李海鹰,苗放,孔祥生,胡玉枝,李海燕.基于6S模型的土法炼焦污染区TM遥感图像大气校正[J].内蒙古石油化工,2006(10)：140-144.
[23]韩改新.一种新的遥感影像变化监测相对辐射校正方法[J].测绘与空间地理信息,2008,31(3)：84-86,
[24]丁丽霞,周斌,王人潮.遥感监测中5种相对辐射校正方法研究[J].浙江大学学报,2005,31(3)： 269-276.
[25]易善祯,王志辉.不同方法在水体遥感提取中的应用实现(J).红外,28(4)：1-9.
[26]王刚,李小曼,田杰。几种TM影像的水体自动提取方法比较[J].测绘科学,2008,33(3)：141-142.
[27]肖玉环,黄伦春,刘晓燕,王媛媛,张红英.基于HIS变换的多源遥感影像融合方法研究[J].工程地球物理学报,2010,7(2)：248-252.
[28]李荣,梁学强,曾丹.遥感影像数据几种HIS融合方法的比较[J].科技经济市场,2009(10)：4-5.
[29]梁玉喜.TM图像用于川西地区森林面积自动提取的研究[D].四川农业大学,2004.
[30]冯德俊,李永树,兰燕.基于主成分变换的动态监测变化信息自动发现[J].计算机工程与应用,2004(36)：199-202.
[31]高程程,惠晓惠.基于灰度共生矩阵的纹理特征提取[J].计算机系统应用,2010,19(6)：195-198.
[32]宁顺刚,白万民,喻钧.基于灰度共生矩阵的图像分割方法研究[J].图像.编码与软件,2009,22(11)：69-71.
[33]苑丽红,付丽,杨勇,苗静.灰度共生矩阵提取纹理特征的实验结果分析[J].计算机应用,2009,29(4)：1018-1021.
[34]冯建辉,杨玉静.基于灰度共生矩阵提取纹理特征图像的研究[J].北京测绘,2007(3),1922.
[35]侯海苗,冀小平.基于灰度共生矩阵的纹理特征[J].长治学院学报,2008,25(5)：31-32.
[36]谭颖.翁源县基于纹理信息及CART决策树技术的遥感影像分类研究[D].江苏：南京林业大学,2008.
[37]郭德俊,宋蜇存.基于灰度共生矩阵的纹理图像分类研究[J].林业机械与木工设备,2005,33(7)：21-23.
[38]薄华,马缚龙,焦李成.图像纹理的灰度共生矩阵计算问题的分析[J].电子学报,2006(1)：155-157.
[39]刘淼,秦大庸,刘家宏,吴发启.基于NDVI的山西省植被覆盖度变化研究[J].人民黄河,2009,31(5)：17-18.
[40]许定成,游先祥,韩熙春.中国遥感进展[M].北京：万国学术出版社,1992.205-210.
[41]FasART模糊神经网络用于遥感图像监督分类的研究[J].中国图象图形学报,2002,7(12)：1263-1268.
[42]陈华,陈书海,张平,严卫东K.means算法在遥感分类中的应用[J].红外与激光工程,2000,29(2)：26-30.
[43]刘鹏,姚正,尹俊杰.一种有效的C4.5改进模型[J].清华大学学报(自然科学版).2006.46(S1)：996-1001.
[44]PalM, MatherPM. An assessment of the effectiveness of decision tree methods of land cover classification [J]. Remote Sensing of Environment,2003 (86):554-565.
[45]余晶,蒋平安,高敏华.基于决策树的土地利用分类方法研究[J].新疆农业科学,2009,46(2)：430-434.
[46]罗开华.遥感图像分类后处理3中实现方法比较[J].林业调查规划,2008,33(2)：24-25.
[47]周斌,杨柏林.运用多时相直接分类法对土地利用进行遥感动态监测的研究[J].自然资源学报2001,16(3)：263-268.
[48]R.N.Siwe.B.Koch.Change vector analysis to categorise land cover change processes using the tasselled cap
as biophysical indicator[J]. Environ Monit Assess (2008) 145:227-235.
[49]William A. Malila.Change Vector Analysis:An Approach for Detecting Forest Changes with Landsat[J]. Laboratory for Applications of Remote Sensing(1980):326-335.
[50]S.Berberoglu,A.Akin.Assessing different remote sensing techniques to detect land use/cover changes in the eastern Mediterranean[J]. International Journal of Applied Earth Observation and Geoinformation (2009) 11:46-53.
[51]邓小炼.基于变化矢量分析的土地利用变化监测方法研究[D]：博士论文.北京：中国科学院,2006.
[52]邸凤萍,朱重光,丁玲.方向矢量法在城市土地利用变化监测中的应用[J].计算机工程,2008,34(2)：253-254.
[53]史榕,许惠平,陈华根.遥感影像数据融合技术及融合质量评价研究[J].鲁东大学学报(自然科学版),2008,24(2)：130-134.
[54]孙家抦等.遥感原理与应用[M].武汉：武汉大学出版社.2009.
[55]陈晋,何春阳,史培军,陈云浩,马楠.基于变化向量分析的土地利用/覆盖变化动态监测---变化阈值的确定方法[J].遥感学报,2001,5(4)：259-266.
[56]郭徵,江洪,陈健,程苗苗,江子山,余树全,李土生.基于遥感的杭州余杭森林景观格局变化[J].浙江林学院学报,2010,27(1)：36-43.
[57]邓劲松,李君,余亮,等.快速城市化过程中杭州市土地利用景观格局动态[J].应用生态学报,2008,19(9)：2003-2008.
[58]Wade T G, Riitters K H, Wickham J D, et al. Distribution and causes of Global forest fragmentation[J/OL].Conservation Ecology,2003,7(2):7[2009].
[59]Riitters K H,Wickham J D,O'Neill R V,et al.Global-Scale Patterns of Forest Fragmentation[J]. ConservationEcology,2000,4(2):3.
[60]方晰,唐代生,杨乐,等.湖南省林科院试验林场森林植被景观格局及破碎化分析[J].中南林业科技大学学报,2008,28(4)：107-112.
[61]Abdullah S A,Nakagoshi N.Changes in agricultural landscape pattern and its spatial relationship with forestland in the State of Selangor,peninsular Malaysia[J].Landscape and Urban Planning,2008,87:147-155.
[62]Abdullah S A,Nakagoshi N.Forest fragmentation and its correlation to human land use Change in the state of Selangor,peninsular Malaysia[J].Forest Ecology and Management,2007,241:39-48.
[63]宗秀影,刘高焕,乔玉良,林松.黄河三角洲湿地景观格局动态变化分析[J].地球信息科学学报,2009,11(1)：91-97.
[64]唐继刚.南京市湿地资源动态遥感监测与分析[J].ecological economy:40-45.
[65]周昕薇.基于3S技术的北京湿地动态监测与评价方法研究[D].首都师范大学,2006.
[66]邓书斌,武红敢,江涛.遥感动态监测中的相对辐射校正方法研究[J].遥感应用,2008.4,71-75.
[67]梅洋,陆苗.基于遥感影像的变化监测研究动态[J].地理信息世界,2009.04：42-47.