选取华北地区为研究区,利用MODIS遥感数据多光谱、多时相优势进行分类特征提取,依据土地覆盖分类特征如地表反射率、植被指数、纹理特征等,并对这些分类特征分别从光谱维、时间维、空间维三个角度进行阐述,结合DEM数据,使用最大似然法进行土地覆盖遥感分类特征的评价与比较。结果表明,不同分类特征对分类精度影响不同,将多种分类特征结合能够有效提高区域尺度土地覆盖分类精度,但分类特征的加入不一定能提高某些类别的分类精度。
This paper briefly introduced the application of some classification features, based on the current research status of the regional land cover classification. Classification features play a chief and basic role in the research of land cover, and it is significant for the classification accuracy to select the features. Due to the advantages of integration of multi-temporal and multi-spectral MODIS data in regional land cover, this paper presented the research on land cover classification in area of North China, selected the features such as surface reflectance (MODIS 7-band reflectance), vegetation index (MODIS-NDVI, MODIS-EVI), two characters of texture (homogeneity, entropy), and DEM to classify by the method of MLC, finally compared and evaluated the classification accuracy by using different features. The result indicates that it has higher overall classification accuracy using MODIS 7-band reflectance data than using vegetation index. And combining the two features above, the classification accuracy of cropland and grassland can be improved obviously. The result also shows the classification accuracy varies with the different features. It can increase the overall classification accuracy to integrate MODIS 7-band reflectance data with some other features, which contain vegetation index, texture and DEM. However, classification accuracy of some types can not be always enhanced by the combination of the above classification characters.
[1] Tucker C J,Townshend J R G,Goff T E.Continental land cover classification using meteorological satellite data[J].Science,1984,227:369-375.
[2] 张佳华,徐祥德,延晓冬,等.多圈层陆面过程参数化研究中遥感信息应用的进展和方向[J].应用气象学报,2003,14(6):346~355.
[3] 吴文斌,杨 鹏,柴崎亮介,等.基于Agent的土地利用/土地覆盖变化模型的研究进展[J].地理科学,2007,27(4):573~578.
[4] DeFries R S,Townshed,J G R.NDVI derived land cover classification at a globe scale[J].Int.J.Remote Sensing,1994,15(17):3567-3586.
[5] 刘勇洪,牛 铮.基于MODIS遥感数据的宏观土地覆盖特征分类方法与精度分析研究[J].遥感技术与应用,2004,19(4):217~224
[6] 刘爱霞,王 静,吕春艳.基于MODIS数据的北京西北部地区土地覆盖分类研究[J].地理科学进展,2006,25(2):96~103.
[7] 潘耀忠,李晓兵,何春阳.中国土地覆盖综合分类研究——基于NOAA/AVHRR和HoldridgePE[J].第四纪研究,2000,20(3):270~281.
[8] 张友水,谢元礼.MODIS影像的NDVI和LSWI植被水分含量估算[J].地理科学,2008,28(1):72~76.
[9] 陈 辉,刘劲松,王 卫.冀北地区植被指数变化特征及影响因素分析[J].地理科学,2008,28(6):794~798.
[10] 历 华,柳钦火,邹 杰.基于MODIS数据的长株潭地区NDBI和NDVI与地表温度的关系研究[J].地理科学,2009,29(2):262~267.
[11] 王爱玲,朱文泉,李 京,等.内蒙古生态系统服务价值遥感测量[J].地理科学,2007,27(3):325~330.
[12] 李月臣,刘春霞.北方13省土地利用/覆盖动态变化分析[J].地理科学,2007,27(1):45~52.
[13] 曾永年,向南平,冯兆东,等.Albedo-NDVI特征空间及沙漠化遥感监测指数研究[J].地理科学,2006,26(1):75~81.
[14] Gopal S,Woodcock C E,Strahler A H.Fuzzy Neural Network Classification of Global Land Cover from a 1° AVHRR Data Set[J].Remote Sensing of Environment,1999,67:230-243.
[15] Hansen M C,DeFries R S,Townshend J R G,et al.Global land cover classification at 1 km resolution using a decision tree classifier[J].Int.J.Remote Sensing,2000,21:1331-1365.
[16] Wen C G,Tateishi R.30-second degree grid land cover classification of Asia[J].Int.J.Remote Sensing,2001,22(18):3845-3854.
[17] 卢 玲,李 新,董庆罕,等.SPOT VEGETATION中国西北地区土地覆盖制图与验证[J].遥感学报,2003,7(5):214~220.
[18] 宫 攀,唐华俊,陈仲新,等.基于MODIS LST修正NDVI时序数列的土地覆盖分类[J].资源科学,2006,28(4):106~110.
[19] 李金莲,刘晓玫,李恒鹏.SPOT5影像纹理特征提取与土地利用信息识别方法[J].遥感学报,2006,10(6):926~931.
[20] 刘晓辉,吕宪国,董贵华.分维模型在土地利用研究中的应用[J].地理科学,2008,28(6):765~769.
[21] 王正兴,刘 闯,Huete Alfredo.植被指数研究进展:从AVHRR-NDVI 到MODIS-EVI[J].生态学报,2003,23(5):979~987.
[22] 赵庚星,李玉环,徐春达.遥感和GIS支持的土地利用动态监测研究——以黄河三角洲利县为例[J].应用生态学报,2000,11(4):573~576.
[23] Muchoney D,Borak J.Application of the MODIS Global Supervised Classification Refel to Vegetation and Land Cover Mapping of Central America[J].Int.J.Remote Sensing,2000,21:1115-1138.
