对东沙岛基于融合信息高分辨遥感影像分类研究
Study on high resolution image classification for Dongsha Island based on fuse information
-
摘要: 以SPOT510m多光谱高分辨率影像为数据基础,建立东沙岛地物信息提取的时空属性和地学规律等先验知识,在此基础上选取地物样本,并进行东沙岛地物光谱分析,从而进行东沙岛的融合信息分析,减少信息提取的误差,使解译的精度提高,结果表明,以先验知识为基础选取的样本光谱能够较好地反映地物光谱的信息含量;高分辨率影像能反映东沙岛的地物特征,并且在影像上能够反应东沙岛潟湖、口门等的发育状态,有效地揭示了地物的区分和地貌发育特征.以此进行的分类结果具有合理性;Kappa分析的分类精度达78.33%.Abstract: Information extraction and geographic law knowledge were established based on SPOT5 10m multi-bands high resolution images.In order to reduce error of the information extraction and to improve int erpreting precision,a spectral analysis for the choosed samples,and a fuse information analysis were made.The results show that(1) the samples based on fuse information can indicate spectral information, (2) high resolution images make Dong sha Island characteristics clear and show lagoons,and sea mouth developed states and(3) the fuse information of high resolution images can clearly show land use type On the basis of the Kappa analysis,the error of the method reaches 78.33%,which can indicate Dongsha Island land use well.
-
Key words:
- fuse information /
- classification /
- method /
- Dongsha Island
-
赵英时.遥感应用分析原理与方法[M].北京:科学出版社,2003.207-225. 刘宝银.南沙群岛遥感融合信息特征分析与计量[M].北京:海洋出版社,2001.24-49. 邹亚荣,王其茂,马超飞.HY-1/COCTS波段信息含量与波段相关性分析[J].海洋科学进展,2005,23(1):70-75. 彭望琭,余先川.遥感与图像解译[M].周涛等译.北京:电子工业出版社,2003.401-410. 刘宝银,杨晓梅.环中国岛链[M].北京:海洋出版社,2003.9-11. 马超飞,蒋兴伟,唐军武等.HY-1CCD宽波段水色要素反演算法[J].海洋学报,2005.27(4):38-44. BARNSLEY M J,BARR S L.Inferring urban land use from satallite sensor images using kernal-based spatial reclassification[J].Phote Engin and Remote Sens,1996,62(8):949-958. MARK J C.Spectral shape classification of Landsat TM imagery[J] ,Photo Engin and Remote Sensing,1998,64(9):905-913. ROBERT A Schowengerdt.Remote Sensing Models and Methods for Image Processing[M].San Diego:Academic Press,1997.398-432. POHL C,VAN GENDEREN J L.Multisensor image fusion in remote sensing:concepts,methods and applications[J].International Journal of Remote Sensing,1998,19(5):823-854. -
点击查看大图
计量
- 文章访问数: 755
- HTML全文浏览量: 4
- PDF下载量: 1017
- 被引次数: 0
下载:
