|
|
面向对象的SPOT 5影像城区水体信息提取研究 |
曹 凯1,2, 江 南2, 吕 恒2,周连义2,刘 新2 |
1.南京大学地理与海洋科学学院,南京210093; 2.中国科学院南京地理与湖泊研究所,南京 210008 |
|
THEEXTRACTIONOFWATERINFORMATIONINURBANAREAS
BASEDONSPOT5IMAGEUSINGOBJECT-ORIENTEDMETHOD |
CAOKai 1,2,JIANGNan 2,LüHeng 2,ZHOULian-yi 2,LIUXin 2 |
1.SchoolofGeographyandOceanographicSciences,NanjingUniversity,Nanjing210093,China;2.NanjingInstituteof
GeographyandLimnology,ChineseAcademyofSciences,Nanjing210008,China |
|
摘要 以南京市部分主城区为研究对象,首先利用基于多尺度图像分割方法对图像进行分割; 然后利用对象所包含的光谱、形状及纹理等特征确定地物识别中所需的各种特征参数; 最后,通过规则的建立,实现研究区地物的逐级分层分类。研究结果表明,该方法不仅使分类结果具有丰富的语义规则信息,而且还有效减少了“分类椒盐现象”的产生,分类精度显著提高。
|
|
关键词 :
高原湖泊动态,
监测气象卫星
|
Abstract: Withthemaincityzonesastheresearcharea,thispaperstudiedthemethodforextractingwaterinfor
mationinSPOT5image.First,theimageissegmentedbymulti-segmentationmethods.Afterthat,incombina
tionwithsomespectral,shapeandtexturalcharacteristics,themulti-levelclassificationmethodcanbesetup.
Theresultshowsthatthiskindofclassificationmethodusedtoextractthewaterinformationincityzoneshasnoton
lyabundantrules-meaninginformationthatcanreducethesalt-phenomenonbutalsoobviouslyhighaccuracy. |
Key words:
Plateau lake
Dynamic monitoring
Meteorological satellite
|
收稿日期: 2006-10-18
出版日期: 2009-07-24
|
|
基金资助: 中国科学院南京地理与湖泊研究所所长基金资助。 |
引用本文: |
曹凯, 江南, 吕恒, 周连义, 刘新. 面向对象的SPOT 5影像城区水体信息提取研究[J]. 国土资源遥感, 2007, 19(2): 27-30.
CAO Kai, JIANG Nan, LV Heng, ZHOU Lian-Yi, LIU Xin. THEEXTRACTIONOFWATERINFORMATIONINURBANAREAS
BASEDONSPOT5IMAGEUSINGOBJECT-ORIENTEDMETHOD. REMOTE SENSING FOR LAND & RESOURCES, 2007, 19(2): 27-30.
|
|
|
|
链接本文: |
https://www.gtzyyg.com/CN/10.6046/gtzyyg.2007.02.07
或
https://www.gtzyyg.com/CN/Y2007/V19/I2/27
|
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
|
Shared |
|
|
|
|
|
Discussed |
|
|
|
|