DSpace Repository

Comparison of Single Channel Indices for U-Net Based Segmentation of Vegetation in Satellite Images

Show simple item record

dc.contributor.author Ülkü, İrem
dc.contributor.author Barmpoutis, P.
dc.contributor.author Stathaki, T.
dc.contributor.author Akagündüz, Erdem
dc.date.accessioned 2020-05-18T12:54:30Z
dc.date.available 2020-05-18T12:54:30Z
dc.date.issued 2020
dc.identifier.citation Ulku, I.; Barmpoutis, P.; Stathaki, T.; Akagunduz, E.,"Comparison of Single Channel Indices for U-Net Based Segmentation of Vegetation in Satellite Images",Proceedings of Spıe - the International Society for Optical Engineering, Vol. 11433, (2020). tr_TR
dc.identifier.isbn 978-151063643-9
dc.identifier.issn 0277786X
dc.identifier.uri http://hdl.handle.net/20.500.12416/3904
dc.description.abstract Hyper-spectral satellite imagery, consisting of multiple visible or infrared bands, is extremely dense and weighty for deep operations. Regarding problems related to vegetation as, more specifically, tree segmentation, it is difficult to train deep architectures due to lack of large-scale satellite imagery. In this paper, we compare the success of different single channel indices, which are constructed from multiple bands, for the purpose of tree segmentation in a deep convolutional neural network (CNN) architecture. The utilized indices are either hand-crafted such as excess green index (ExG) and normalized difference vegetation index (NDVI) or reconstructed from the visible bands using feature space transformation methods such as principle component analysis (PCA). For comparison, these features are fed to an identical CNN architecture, which is a standard U-Net-based symmetric encoder-decoder design with hierarchical skip connections and the segmentation success for each single index is recorded. Experimental results show that single bands, which are constructed from the vegetation indices and space transformations, can achieve similar segmentation performances as compared to that of the original multi-channel case tr_TR
dc.language.iso eng tr_TR
dc.publisher SPIE tr_TR
dc.relation.isversionof 10.1117/12.2556374 tr_TR
dc.rights info:eu-repo/semantics/closedAccess tr_TR
dc.subject Deep Convolutional Neural Networks tr_TR
dc.subject Hyper-Spectral Imagery tr_TR
dc.subject Vegetation Segmentation tr_TR
dc.title Comparison of Single Channel Indices for U-Net Based Segmentation of Vegetation in Satellite Images tr_TR
dc.type conferenceObject tr_TR
dc.relation.journal Proceedings of Spıe - the International Society for Optical Engineering tr_TR
dc.contributor.authorID 233834 tr_TR
dc.identifier.volume 11433 tr_TR
dc.contributor.department Çankaya Üniversitesi, Mühendislik Fakültesi, Elektrik Elektronik Mühendisliği Bölümü tr_TR


Files in this item

Files Size Format View

There are no files associated with this item.

This item appears in the following Collection(s)

Show simple item record