東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Crop Classification Using Sentinel-2...

Murlender, Ana Wegman.

FindBook

Google Book

Amazon

博客來

Crop Classification Using Sentinel-2A Satellite Observations and Support Vector Machine Algorithm Over a Highly Diverse Agricultural Region.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Crop Classification Using Sentinel-2A Satellite Observations and Support Vector Machine Algorithm Over a Highly Diverse Agricultural Region./
作者:	Murlender, Ana Wegman.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2019,
面頁冊數:	39 p.
附註:	Source: Masters Abstracts International, Volume: 81-02.
Contained By:	Masters Abstracts International81-02.
標題:	Geography. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=13424141
ISBN:	9781085577762

Crop Classification Using Sentinel-2A Satellite Observations and Support Vector Machine Algorithm Over a Highly Diverse Agricultural Region.
Murlender, Ana Wegman.

Crop Classification Using Sentinel-2A Satellite Observations and Support Vector Machine Algorithm Over a Highly Diverse Agricultural Region. - Ann Arbor : ProQuest Dissertations & Theses, 2019 - 39 p.

Source: Masters Abstracts International, Volume: 81-02.

Thesis (M.A.)--University of California, Davis, 2019.

This item must not be sold to any third party vendors.

Accurate crop maps are important for agricultural monitoring and management. Multispectral data from moderate spatial resolution satellite sensors such as Landsat has been widely used for crop mapping, but producing accurate classifications remains a challenge. Sentinel-2A presents several advantages in terms of temporal, spatial and spectral resolution in comparison to Landsat-8. Sentinel-2A has higher spectral resolution than Landsat-8 and also includes three additional bands in the "red edge" spectrum. Red edge reflectance contains potential useful information about vegetation and thus may improve vegetation classification. However, the degree to which Sentinel-2A data can improve the accuracy of crop classification has not been established. I evaluated the value of the additional band by comparing classification with "Landsat-like bands only" with classification with all Sentinel bands. I assessed the accuracy of crop classification for the Sacramento-San Joaquin Delta, an agriculturally diverse region in California, USA. A Support Vector Machine algorithm with both linear and radial kernel functions was used to classify crop types at a 10 m spatial resolution. Training data was randomly selected from the pixels with agreement between two existing classifications ("CropScape" and California Department of Water Resources by Land IQ). Sentinel-2A images from 2016 summer and winter seasons were included to account for seasonal differences. SVM with a radial kernel performed the best, with an overall accuracy of 92.1% when using all Sentinel bands plus selected vegetation indices from both seasons. The accuracy of the SVM with linear kernel function was 7.6% lower on average. The addition of Sentinel's red edge spectral channels compared to Landsat-like Sentinel bands, when using a linear kernel SVM model and Sentinel data obtained in summer, increased the overall accuracy by 6.0% for all crops and 42.5% on average for fruit and nut crops (such as cherries, pears, almonds, walnuts, and olives). When including the winter scene, the overall accuracy increased on average by 5.8% for all crops and by 29.5% for fruit and nut crops. These results demonstrated the added value of Sentinel-2A images in mapping diverse crop types, in comparison with Landsat-type observations.

ISBN: 9781085577762Subjects--Topical Terms:

524010
Geography.
Subjects--Index Terms:

Crop classification

Crop Classification Using Sentinel-2A Satellite Observations and Support Vector Machine Algorithm Over a Highly Diverse Agricultural Region.
LDR:03580nmm a2200385 4500 001 2269187
005 20200910100153.5
008 220629s2019 ||||||||||||||||| ||eng d
020 $a 9781085577762
035 $a (MiAaPQ)AAI13424141
035 $a AAI13424141
040 $a MiAaPQ $c MiAaPQ
100 1 $a Murlender, Ana Wegman. $3 3546491
245 1 0 $a Crop Classification Using Sentinel-2A Satellite Observations and Support Vector Machine Algorithm Over a Highly Diverse Agricultural Region.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2019
300 $a 39 p.
500 $a Source: Masters Abstracts International, Volume: 81-02.
500 $a Advisor: Jin, Yufang.
502 $a Thesis (M.A.)--University of California, Davis, 2019.
506 $a This item must not be sold to any third party vendors.
506 $a This item must not be added to any third party search indexes.
520 $a Accurate crop maps are important for agricultural monitoring and management. Multispectral data from moderate spatial resolution satellite sensors such as Landsat has been widely used for crop mapping, but producing accurate classifications remains a challenge. Sentinel-2A presents several advantages in terms of temporal, spatial and spectral resolution in comparison to Landsat-8. Sentinel-2A has higher spectral resolution than Landsat-8 and also includes three additional bands in the "red edge" spectrum. Red edge reflectance contains potential useful information about vegetation and thus may improve vegetation classification. However, the degree to which Sentinel-2A data can improve the accuracy of crop classification has not been established. I evaluated the value of the additional band by comparing classification with "Landsat-like bands only" with classification with all Sentinel bands. I assessed the accuracy of crop classification for the Sacramento-San Joaquin Delta, an agriculturally diverse region in California, USA. A Support Vector Machine algorithm with both linear and radial kernel functions was used to classify crop types at a 10 m spatial resolution. Training data was randomly selected from the pixels with agreement between two existing classifications ("CropScape" and California Department of Water Resources by Land IQ). Sentinel-2A images from 2016 summer and winter seasons were included to account for seasonal differences. SVM with a radial kernel performed the best, with an overall accuracy of 92.1% when using all Sentinel bands plus selected vegetation indices from both seasons. The accuracy of the SVM with linear kernel function was 7.6% lower on average. The addition of Sentinel's red edge spectral channels compared to Landsat-like Sentinel bands, when using a linear kernel SVM model and Sentinel data obtained in summer, increased the overall accuracy by 6.0% for all crops and 42.5% on average for fruit and nut crops (such as cherries, pears, almonds, walnuts, and olives). When including the winter scene, the overall accuracy increased on average by 5.8% for all crops and by 29.5% for fruit and nut crops. These results demonstrated the added value of Sentinel-2A images in mapping diverse crop types, in comparison with Landsat-type observations.
590 $a School code: 0029.
650 4 $a Geography. $3 524010
650 4 $a Agriculture. $3 518588
650 4 $a Geographic information science. $3 3432445
653 $a Crop classification
653 $a LULC mapping
653 $a Red edge
653 $a Sentinel-2
653 $a Support vector machine
690 $a 0366
690 $a 0370
690 $a 0473
710 2 $a University of California, Davis. $b Geography. $3 1677408
773 0 $t Masters Abstracts International $g 81-02.
790 $a 0029
791 $a M.A.
792 $a 2019
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=13424141