Deep learning-based semantic segmentation in remote sensing

Devis Tuia; Diego Marcos; Konrad Schindler; Bertrand Le Saux

doi:10.1002/9781119646181.ch5

Deep learning-based semantic segmentation in remote sensing

Devis Tuia^*, Diego Marcos, Konrad Schindler, Bertrand Le Saux

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Chapter › Academic › peer-review

1 Citation (Scopus)

Abstract

Semantic segmentation consists of the generation of a categorical map, given an image in which each pixel of the image is automatically assigned a class. Deep learning allows the influence of the pixel's context to be learned by capturing the non-linear relationships between surrounding image features at multiple scales, leading to large improvements in performance and opening up the door to new applications. This chapter explores the use of deep learning-based semantic segmentation in Earth observation imagery and presents in detail three approaches specifically aimed at Earth observation applications.

Original language	English
Title of host publication	Deep Learning for the Earth Sciences
Subtitle of host publication	A Comprehensive Approach to Remote Sensing, Climate Science and Geosciences
Editors	Gustau Camps-Valls, Devis Tuia, Xiao Xiang Zhu, Markus Reichstein
Publisher	Wiley
Chapter	5
Pages	46-66
Number of pages	21
ISBN (Electronic)	9781119646181
ISBN (Print)	9781119646143
DOIs	https://doi.org/10.1002/9781119646181.ch5
Publication status	Published - 20 Aug 2021

Keywords

Deep learning architectures
Image classification
Remote sensing
Semantic segmentation

Access to Document

10.1002/9781119646181.ch5

Cite this

@inbook{3790f0a57c25479ea3aabac51093f068,

title = "Deep learning-based semantic segmentation in remote sensing",

abstract = "Semantic segmentation consists of the generation of a categorical map, given an image in which each pixel of the image is automatically assigned a class. Deep learning allows the influence of the pixel's context to be learned by capturing the non-linear relationships between surrounding image features at multiple scales, leading to large improvements in performance and opening up the door to new applications. This chapter explores the use of deep learning-based semantic segmentation in Earth observation imagery and presents in detail three approaches specifically aimed at Earth observation applications.",

keywords = "Deep learning architectures, Image classification, Remote sensing, Semantic segmentation",

author = "Devis Tuia and Diego Marcos and Konrad Schindler and {Le Saux}, Bertrand",

year = "2021",

month = aug,

day = "20",

doi = "10.1002/9781119646181.ch5",

language = "English",

isbn = "9781119646143",

pages = "46--66",

editor = "Gustau Camps-Valls and Devis Tuia and {Xiang Zhu}, Xiao and Markus Reichstein",

booktitle = "Deep Learning for the Earth Sciences",

publisher = "Wiley",

address = "United States",

}

Deep learning-based semantic segmentation in remote sensing. / Tuia, Devis; Marcos, Diego; Schindler, Konrad et al.
Deep Learning for the Earth Sciences: A Comprehensive Approach to Remote Sensing, Climate Science and Geosciences. ed. / Gustau Camps-Valls; Devis Tuia; Xiao Xiang Zhu; Markus Reichstein. Wiley, 2021. p. 46-66.

Research output: Chapter in Book/Report/Conference proceeding › Chapter › Academic › peer-review

TY - CHAP

T1 - Deep learning-based semantic segmentation in remote sensing

AU - Tuia, Devis

AU - Marcos, Diego

AU - Schindler, Konrad

AU - Le Saux, Bertrand

PY - 2021/8/20

Y1 - 2021/8/20

N2 - Semantic segmentation consists of the generation of a categorical map, given an image in which each pixel of the image is automatically assigned a class. Deep learning allows the influence of the pixel's context to be learned by capturing the non-linear relationships between surrounding image features at multiple scales, leading to large improvements in performance and opening up the door to new applications. This chapter explores the use of deep learning-based semantic segmentation in Earth observation imagery and presents in detail three approaches specifically aimed at Earth observation applications.

AB - Semantic segmentation consists of the generation of a categorical map, given an image in which each pixel of the image is automatically assigned a class. Deep learning allows the influence of the pixel's context to be learned by capturing the non-linear relationships between surrounding image features at multiple scales, leading to large improvements in performance and opening up the door to new applications. This chapter explores the use of deep learning-based semantic segmentation in Earth observation imagery and presents in detail three approaches specifically aimed at Earth observation applications.

KW - Deep learning architectures

KW - Image classification

KW - Remote sensing

KW - Semantic segmentation

U2 - 10.1002/9781119646181.ch5

DO - 10.1002/9781119646181.ch5

M3 - Chapter

AN - SCOPUS:85132387639

SN - 9781119646143

SP - 46

EP - 66

BT - Deep Learning for the Earth Sciences

A2 - Camps-Valls, Gustau

A2 - Tuia, Devis

A2 - Xiang Zhu, Xiao

A2 - Reichstein, Markus

PB - Wiley

ER -

Deep learning-based semantic segmentation in remote sensing

Abstract

Keywords

Access to Document

Fingerprint

Cite this