全部 标题 作者
关键词 摘要

OALib Journal期刊
ISSN: 2333-9721
费用:99美元

查看量下载量

相关文章

更多...
-  2019 

探寻葡萄树遥感辨识的最佳时相与方法
The Best Phase and the Optimal Identification Method for Identifying Grape Tree Species

DOI: 10.3969/j.issn.1001-7461.2019.05.28

Keywords: 葡萄树,遥感影像,最佳时相,果树辨识,光谱指数
,grape tree,remote sensing image,the best phase,fruit tree identification,spectral index

Full-Text   Cite this paper   Add to My Lib

Abstract:

利用41景卫星影像探寻葡萄树遥感辨识的最佳时相与方法。首先对各景影像分别进行了预处理,随后利用6类探试性方法(即地物光谱比较、波段差分或比值分析、光谱指数求算与分析、光谱指数变化追踪、影像复合与辨识方法协同分析)对探试组影像进行了辨识分析,并从中选出3种较佳的辨识方法,最后利用优选的3种方法对验证组影像进行了辨识验证。结果表明:1)在4月中旬至5月初的影像中,仅采用NDVI阈值便可将葡萄树与别种成年果树、小麦予以区分,但低龄的别种果树、有草撂荒地等与葡萄树易混;2)在5月底至6月中旬的影像中,采用(ρBLUE+ρGREEN+ρRED)与NDVI的双重阈值可将葡萄树与其他地物较好的区分开来;3)在上述2时段影像复合、所用辨识方法协同的情况下,可显著提高葡萄树的辨识精度(葡萄树类的正确识别率可达90.42%,总体分类精度可达96.53%);4)葡萄树遥感辨识的较佳时相为4月中旬至5月初,最佳时相为5月底至6月中旬。
The paper attempted to find out the best phase and the optimal identification method for identifying grape trees by using 41 satellite remote sensing images.Firstly,the images of each period were preprocessed by using image processing technology.Then,the three better identification methods were selected from the six kinds of tentative identification methods which were used in the test group of image identification and analysis.Finally,the verification group images were verified by using the three better identification methods.The conclusions were as follows.1) From mid-April to early May when grape trees had lower NDVI values,they could be distinguished from other adult fruit tree species and wheat by using NDVI threshold.But they were difficult to be distinguished from other young fruit tree species and abandoned land with grass.2) From the end of May to the middle of June,grape trees had high reflectivity in visible and near infrared bands.They could be distinguished from other ground objects by the double threshold of (ρBLUE+ρGREEN+ρRED) and NDVI.3) The identification precision of grape trees could be improved significantly in the circumstances of the above two periods by the recombination of the images and the collaboration of the identification method (the improved correct recognition rate of grape trees could reach 90.42%,the overall classification accuracy was 96.53%).4) The better phase for remote sensing identification of grape trees was from the middle of April to the beginning of May,from the end of May to the middle of June,and the best phase was from the end of May to the middle of June

Full-Text

Contact Us

service@oalib.com

QQ:3279437679

WhatsApp +8615387084133