基于多特征SVM的昆明市城市森林时空分布信息提取
|
摘要
采用2017年昆明市主城区Landsat TM/OLI及DEM数据,采用SVM分类方法,对比不同多特征组合的分类精度筛选出森林提取的最佳特征组合,并由此得到2000、2010及2017年昆明市主城区森林分布,分析3期森林的总面积、不同海拔森林面积分布和植被覆盖度变化。结果表明,光谱、纹理以及地形特征的多特征组合为城市森林提取的最佳组合(精度为92.69%);2000—2017年昆明市主城区森林总面积呈上升趋势,海拔低于2 000 m区域的森林面积逐年减少,而高于2 000 m区域的森林面积逐年增加;随时空变化呈现出低植被覆盖度及高植被覆盖度面积增加,中植被覆盖度及较高植被覆盖度面积减少的趋势。
Based on the data of Landsat TM/OLI and DEM in main urban area of Kunming in 2017,the SVM classification method was used by comparing the classification accuracy of different multi-feature combinations to select the best feature combination for forest extraction and obtain the distribution of urban forests in 2000,2010 and 2017. The analysis for the differences of total area,the distributions of forests at different altitudes,and the changes of vegetation coverage in three phases showed that the combination of spectral,texture,and topographic features was the best for urban forest extraction(with an accuracy of 92.69%). The total area of forests in main urban area gradually increased from 2000 to 2017,the area at elevations below 2000 m decreased and the area above 2000 m increased gradually. The area of low vegetation coverage and high coverage showed the increasing trend,and the middle coverage and higher coverage was decreasing.
Based on the data of Landsat TM/OLI and DEM in main urban area of Kunming in 2017,the SVM classification method was used by comparing the classification accuracy of different multi-feature combinations to select the best feature combination for forest extraction and obtain the distribution of urban forests in 2000,2010 and 2017. The analysis for the differences of total area,the distributions of forests at different altitudes,and the changes of vegetation coverage in three phases showed that the combination of spectral,texture,and topographic features was the best for urban forest extraction(with an accuracy of 92.69%). The total area of forests in main urban area gradually increased from 2000 to 2017,the area at elevations below 2000 m decreased and the area above 2000 m increased gradually. The area of low vegetation coverage and high coverage showed the increasing trend,and the middle coverage and higher coverage was decreasing.
引文
[1] 郑国强,赵清,朱鹏,等.济南市城市森林信息提取及景观格局分析[J].资源科学,2008,30(5):786-791.
[2] 王成,蔡春菊,陶康华.城市森林的概念、范围及其研究[J].世界林业研究,2004,17(2):23-27.
[3] 唐泽,任志彬,郑海峰,等.城市森林群落结构特征的降温效应[J].应用生态学报,2017,28(9):2823-2830.
[4] 马志勇,沈涛,张军海,等.基于植被覆盖度的植被变化分析[J].测绘通报,2007(3):45-48.
[5] 陈晋,陈云浩,何春阳,等.基于土地覆盖分类的植被覆盖率估算亚像元模型与应用[J].遥感学报,2001,5(6):416-422.
[6] 张锦水,何春阳,潘耀忠,等.基于SVM的多源信息复合的高空间分辨率遥感数据分类研究[J].遥感学报,2006,10(1):49-57.
[7] 郭佳忱.基于SVM方法的长白山森林植被信息提取的研究[J].吉林林业科技,2010,39(1):14-17.
[8] 刘海红.遥感监测城市扩张与“精明增长”策略分析研究[D].昆明:昆明理工大学,2012.
[9] 张琴.昆明市主城区城市森林建设定量评价[J].Geographical Science Research,2015,4(3):127-135.
[10] 田传召,于阳,肖虹雁,等.基于TM影像的昆明市区土地利用遥感动态监测[J].林业资源管理,2014(4):103-108.
[11] COHEN W B.,GOWARD S N.Landsat′s Role in Ecological Applications of Remote Sensing[J].Bioscience,2004,54(6):535-545.
[12] 饶萍,王建力,王勇.基于多特征决策树的建设用地信息提取[J].农业工程学报,2014,30(12):233-240.
[13] 初庆伟,张洪群,吴业炜,等.Landsat-8卫星数据应用探讨[J].遥感信息,2013,28(4):110-114.
[14] 李梦颖,邢艳秋,刘美爽,等.基于支持向量机的Landsat-8影像森林类型识别研究[J].中南林业科技大学学报,2017,37(4):52-58.
[15] RUNDQUIST B C.The influence of canopy green vegetation fraction on spectral measurements over native tallgrass prairie[J].Remote Sensing of Environment,2002,81(1):129-135.
[16] 徐涵秋.利用改进的归一化差异水体指数(MNDWI)提取水体信息的研究[J].遥感学报,2005,9(5):589-595.
[17] 胡玉福,邓良基,匡先辉,等.基于纹理特征的高分辨率遥感图像土地利用分类研究[J].地理与地理信息科学,2011,27(5):42-45.
[18] 魏莉莉,寇卫利,向兰兰,等.西双版纳热带森林景观破碎化地形差异性分析[J].西南林业大学学报,2018,38(2).
[19] CORTES C,Vapnik Vladimir Support Vector Network[J].Machine learning,1995,20(3):273-297.
[20] 周晓宇,陈富龙,姜爱辉.基于SVM雷达卧龙大熊猫栖息地森林成图[J].国土资源遥感,2017,29(3):85-91.
[21] 陈波,张友静,陈亮.结合纹理的SVM遥感影像分类研究[J].测绘工程,2007,16(5):23-27.
[22] 奉国和.SVM分类核函数及参数选择比较[J].计算机工程与应用,2011,47(3):123-124.
[23] 陈涛,李平湘,张良培.武汉地区1988-2002年植被覆盖度变化动态分析[J].遥感技术与应用,2008,23(5):511-516.
[24] 李贺颖,王艳慧.城市热岛效应与植被覆盖度定量关系特征研究[J].地理空间信息,2013,11(5):78-81.
[25] 赵雷,杨品红,戴文新,等.城乡建设用地规模与合理布局研究[C]//中国土地学会学术年会论文集,2009.
[2] 王成,蔡春菊,陶康华.城市森林的概念、范围及其研究[J].世界林业研究,2004,17(2):23-27.
[3] 唐泽,任志彬,郑海峰,等.城市森林群落结构特征的降温效应[J].应用生态学报,2017,28(9):2823-2830.
[4] 马志勇,沈涛,张军海,等.基于植被覆盖度的植被变化分析[J].测绘通报,2007(3):45-48.
[5] 陈晋,陈云浩,何春阳,等.基于土地覆盖分类的植被覆盖率估算亚像元模型与应用[J].遥感学报,2001,5(6):416-422.
[6] 张锦水,何春阳,潘耀忠,等.基于SVM的多源信息复合的高空间分辨率遥感数据分类研究[J].遥感学报,2006,10(1):49-57.
[7] 郭佳忱.基于SVM方法的长白山森林植被信息提取的研究[J].吉林林业科技,2010,39(1):14-17.
[8] 刘海红.遥感监测城市扩张与“精明增长”策略分析研究[D].昆明:昆明理工大学,2012.
[9] 张琴.昆明市主城区城市森林建设定量评价[J].Geographical Science Research,2015,4(3):127-135.
[10] 田传召,于阳,肖虹雁,等.基于TM影像的昆明市区土地利用遥感动态监测[J].林业资源管理,2014(4):103-108.
[11] COHEN W B.,GOWARD S N.Landsat′s Role in Ecological Applications of Remote Sensing[J].Bioscience,2004,54(6):535-545.
[12] 饶萍,王建力,王勇.基于多特征决策树的建设用地信息提取[J].农业工程学报,2014,30(12):233-240.
[13] 初庆伟,张洪群,吴业炜,等.Landsat-8卫星数据应用探讨[J].遥感信息,2013,28(4):110-114.
[14] 李梦颖,邢艳秋,刘美爽,等.基于支持向量机的Landsat-8影像森林类型识别研究[J].中南林业科技大学学报,2017,37(4):52-58.
[15] RUNDQUIST B C.The influence of canopy green vegetation fraction on spectral measurements over native tallgrass prairie[J].Remote Sensing of Environment,2002,81(1):129-135.
[16] 徐涵秋.利用改进的归一化差异水体指数(MNDWI)提取水体信息的研究[J].遥感学报,2005,9(5):589-595.
[17] 胡玉福,邓良基,匡先辉,等.基于纹理特征的高分辨率遥感图像土地利用分类研究[J].地理与地理信息科学,2011,27(5):42-45.
[18] 魏莉莉,寇卫利,向兰兰,等.西双版纳热带森林景观破碎化地形差异性分析[J].西南林业大学学报,2018,38(2).
[19] CORTES C,Vapnik Vladimir Support Vector Network[J].Machine learning,1995,20(3):273-297.
[20] 周晓宇,陈富龙,姜爱辉.基于SVM雷达卧龙大熊猫栖息地森林成图[J].国土资源遥感,2017,29(3):85-91.
[21] 陈波,张友静,陈亮.结合纹理的SVM遥感影像分类研究[J].测绘工程,2007,16(5):23-27.
[22] 奉国和.SVM分类核函数及参数选择比较[J].计算机工程与应用,2011,47(3):123-124.
[23] 陈涛,李平湘,张良培.武汉地区1988-2002年植被覆盖度变化动态分析[J].遥感技术与应用,2008,23(5):511-516.
[24] 李贺颖,王艳慧.城市热岛效应与植被覆盖度定量关系特征研究[J].地理空间信息,2013,11(5):78-81.
[25] 赵雷,杨品红,戴文新,等.城乡建设用地规模与合理布局研究[C]//中国土地学会学术年会论文集,2009.