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

FindBook

Google Book

Amazon

博客來

Leveraging High-Resolution Imagery and New Technologies in Machine Learning to Map Forest Disturbances.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Leveraging High-Resolution Imagery and New Technologies in Machine Learning to Map Forest Disturbances./
作者:	Wegmueller, Sarah A.
面頁冊數:	1 online resource (152 pages)
附註:	Source: Dissertations Abstracts International, Volume: 84-11, Section: B.
Contained By:	Dissertations Abstracts International84-11B.
標題:	Forestry. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30490054click for full text (PQDT)
ISBN:	9798379507121

Leveraging High-Resolution Imagery and New Technologies in Machine Learning to Map Forest Disturbances.
Wegmueller, Sarah A.

Leveraging High-Resolution Imagery and New Technologies in Machine Learning to Map Forest Disturbances. - 1 online resource (152 pages)

Source: Dissertations Abstracts International, Volume: 84-11, Section: B.

Thesis (Ph.D.)--The University of Wisconsin - Madison, 2023.

Includes bibliographical references

Remotely sensed imagery, satellite and airborne, has great potential to provide increased monitoring and mapping capabilities for applications in forest health and management. In this dissertation, I investigated ways to leverage newly available satellite imagery, machine learning techniques, and recently collected ground data to develop new methods for monitoring and mapping forest health at various scales and for a range of purposes. The efforts my collaborators and I resulted in two new software programs, named Astrape and Tree Condition and Analysis Program (TreeCAP), that collectively map disturbances ranging from large, stand-replacing derechos to individual tree mortality in a mixed forest with accuracies typically over 80%. Both of these systems were designed to be scaled up for operational use across the contiguous US, and maybe internationally. Astrape is capable of using nearly any imagery source but was designed with Sentinel-2 imagery and Dove imagery. It implements a machine learning framework to produce thematic maps of damage severity in four classes (high severity, moderate severity, low severity, and little to no damage) with limited need for ground-truthing. TreeCAP was built to leverage the National Agricultural Imagery Program (NAIP) data that has a spatial resolution of 0.6-1 m, suitable for differentiating individual trees. TreeCAP uses a machine learning model to create thematic maps of healthy, morbid, and dead trees with high accuracy. Further, I pioneered a vital method to normalize the highly radiometrically variable Dove data called LOESS Radiometric Correction for Contiguous Scenes (LORACCS). The output of LORACCS can be used to create seamless Dove imagery mosaics that can then be used with the aforementioned systems, greatly expanding their temporal and spatial potential. Finally, I conducted a reinvestigating of oak wilt spread in Wisconsin using a time series of NAIP imagery and ground-confirmed incidents (courtesy of the Wisconsin Department of Natural Resources Forest Health Team). The results of this study indicate that oak wilt may be far more prevalent on the landscape than is currently known, highlighting the value of using remote sensing to better understand the patterns of insects and disease regionally. The chapters in this dissertation are ordered chronologically based on when projects were completed. Chapter 1 features LORACCS, followed by Astrape and TreeCAP, in Chapters 2 and 3, respectively, and ending with Chapter 4 on the reinvestigation of oak wilt spread in Wisconsin. An overarching introduction follows this abstract and a section discussing future directions and implications concludes this dissertation.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2023

Mode of access: World Wide Web

ISBN: 9798379507121Subjects--Topical Terms:

895157
Forestry.
Subjects--Index Terms:

AstrapeIndex Terms--Genre/Form:

542853
Electronic books.

Leveraging High-Resolution Imagery and New Technologies in Machine Learning to Map Forest Disturbances.
LDR:04092nmm a2200397K 4500 001 2363213
005 20231116093826.5
006 m o d
007 cr mn ---uuuuu
008 241011s2023 xx obm 000 0 eng d
020 $a 9798379507121
035 $a (MiAaPQ)AAI30490054
035 $a AAI30490054
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Wegmueller, Sarah A. $3 3703965
245 1 0 $a Leveraging High-Resolution Imagery and New Technologies in Machine Learning to Map Forest Disturbances.
264 0 $c 2023
300 $a 1 online resource (152 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 84-11, Section: B.
500 $a Advisor: Townsend, Philip A.
502 $a Thesis (Ph.D.)--The University of Wisconsin - Madison, 2023.
504 $a Includes bibliographical references
520 $a Remotely sensed imagery, satellite and airborne, has great potential to provide increased monitoring and mapping capabilities for applications in forest health and management. In this dissertation, I investigated ways to leverage newly available satellite imagery, machine learning techniques, and recently collected ground data to develop new methods for monitoring and mapping forest health at various scales and for a range of purposes. The efforts my collaborators and I resulted in two new software programs, named Astrape and Tree Condition and Analysis Program (TreeCAP), that collectively map disturbances ranging from large, stand-replacing derechos to individual tree mortality in a mixed forest with accuracies typically over 80%. Both of these systems were designed to be scaled up for operational use across the contiguous US, and maybe internationally. Astrape is capable of using nearly any imagery source but was designed with Sentinel-2 imagery and Dove imagery. It implements a machine learning framework to produce thematic maps of damage severity in four classes (high severity, moderate severity, low severity, and little to no damage) with limited need for ground-truthing. TreeCAP was built to leverage the National Agricultural Imagery Program (NAIP) data that has a spatial resolution of 0.6-1 m, suitable for differentiating individual trees. TreeCAP uses a machine learning model to create thematic maps of healthy, morbid, and dead trees with high accuracy. Further, I pioneered a vital method to normalize the highly radiometrically variable Dove data called LOESS Radiometric Correction for Contiguous Scenes (LORACCS). The output of LORACCS can be used to create seamless Dove imagery mosaics that can then be used with the aforementioned systems, greatly expanding their temporal and spatial potential. Finally, I conducted a reinvestigating of oak wilt spread in Wisconsin using a time series of NAIP imagery and ground-confirmed incidents (courtesy of the Wisconsin Department of Natural Resources Forest Health Team). The results of this study indicate that oak wilt may be far more prevalent on the landscape than is currently known, highlighting the value of using remote sensing to better understand the patterns of insects and disease regionally. The chapters in this dissertation are ordered chronologically based on when projects were completed. Chapter 1 features LORACCS, followed by Astrape and TreeCAP, in Chapters 2 and 3, respectively, and ending with Chapter 4 on the reinvestigation of oak wilt spread in Wisconsin. An overarching introduction follows this abstract and a section discussing future directions and implications concludes this dissertation.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2023
538 $a Mode of access: World Wide Web
650 4 $a Forestry. $3 895157
650 4 $a Remote sensing. $3 535394
650 4 $a Wildlife management. $3 571816
653 $a Astrape
653 $a Forest health
653 $a Oak wilt
653 $a Remote sensing
653 $a TreeCAP
655 7 $a Electronic books. $2 lcsh $3 542853
690 $a 0478
690 $a 0799
690 $a 0286
710 2 $a ProQuest Information and Learning Co. $3 783688
710 2 $a The University of Wisconsin - Madison. $b Forestry. $3 3170937
773 0 $t Dissertations Abstracts International $g 84-11B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30490054 $z click for full text (PQDT)