Using unmanned aerial systems to classify land cover and assess productivity in forested wetlands in Nova Scotia

Show simple item record

dc.contributor.advisor Harper, Karen Amanda, 1969-
dc.creator Wilson, Iain M. J. C.
dc.date.accessioned 2019-05-01T14:55:44Z
dc.date.available 2019-05-01T14:55:44Z
dc.date.issued 2019
dc.identifier.uri http://library2.smu.ca/handle/01/28364
dc.description 1 online resource (vi, 50 p.) : colour illustrations
dc.description Includes abstract and appendix.
dc.description Includes bibliographical references (p. 41-46).
dc.description.abstract Forested wetlands in Nova Scotia are an understudied and ephemeral ecosystem type with high predicted ecological value. They are thought to cover a broad geographical area; however, their distribution is difficult to quantify, partly due to their similarity to drier forested landscapes in ordinary RGB aerial imagery. This study used unmanned aerial systems (UAS) imagery to attempt to classify and quantify the distribution of forested wetland communities, differentiate forested wetlands from drier forested communities, and assess productivity levels using the normalized difference vegetation index (NDVI). This study is one of the first known examples of UAS use in forested wetland ecosystems. NDVI imagery was captured in the Musquodoboit River valley during the summer of 2018 using a consumer-grade UAS and processed into orthomosaic maps in Pix4D Mapper Pro. The maximum likelihood classifier algorithm was applied to the dataset to group similar pixels into land cover classes based on ground truth data collected in the same time frame as the UAS flights. The classification scheme was then put through a confusion matrix to assess its accuracy. Based on this assessment, the classification was not accurate. This may be due to several factors, including flaws in the ground sampling method, and the fact that forests are generally difficult to classify through pixel-based classification methods. NDVI values did not differ greatly across land cover classes, which may have played a role in the unsuccessful classification. Suggestions for future studies include using a more rigorous and quantitative ground sampling protocol and considering different classification methods, such as object-based image analysis (OBIA). en_CA
dc.description.provenance Submitted by Greg Hilliard (greg.hilliard@smu.ca) on 2019-05-01T14:55:44Z No. of bitstreams: 1 Wilson_Iain_Honours_2019.pdf: 2009903 bytes, checksum: a5db5e0f9542837b7ce25fc927ac8d08 (MD5) en
dc.description.provenance Made available in DSpace on 2019-05-01T14:55:44Z (GMT). No. of bitstreams: 1 Wilson_Iain_Honours_2019.pdf: 2009903 bytes, checksum: a5db5e0f9542837b7ce25fc927ac8d08 (MD5) Previous issue date: 2019-04-12 en
dc.language.iso en en_CA
dc.publisher Halifax, N.S. : Saint Mary's University
dc.title Using unmanned aerial systems to classify land cover and assess productivity in forested wetlands in Nova Scotia en_CA
dc.type Text en_CA
thesis.degree.name Bachelor of Environmental Studies (Honours Environmental Studies)
thesis.degree.level Undergraduate
thesis.degree.discipline Geography and Environmental Studies
thesis.degree.grantor Saint Mary's University (Halifax, N.S.)
 Find Full text

Files in this item

 
 

This item appears in the following Collection(s)

Show simple item record