Deep recurrent neural networks for land-cover classification using sentinel-1 insar time series

Shaojia Ge; Oleg Antropov; Weimin Su; Hong Gu; Jaan Praks

doi:10.1109/IGARSS.2019.8900088

Deep recurrent neural networks for land-cover classification using sentinel-1 insar time series

Shaojia Ge, Oleg Antropov, Weimin Su, Hong Gu, Jaan Praks

Not published at VTT

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

5 Citations (Scopus)

Abstract

To date, the potential of multitemporal interferometric SAR (InSAR) data in land-cover mapping has not been fully explored despite suitable time series increasingly acquired from SAR sensors. Here, we suggest to use an LSTM (Long Short Term Memory) based land-cover classifier to address this problem. Spatial context is preserved by using grey-level spatial dependencies and morphological profiles. Further, a 4-LSTM-based model was trained to capture the temporal dynamics of InSAR coherence. Altogether 39 Sentinel-1 interferometric coherence pairs acquired over Donana in Spain were used to evaluate the method performance. Achieved more than 90% overall accuracy indicates the strong potential of developed InSAR recurrent approach in improving differentiation between various land cover classes.

Original language	English
Title of host publication	IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium
Publisher	IEEE Institute of Electrical and Electronic Engineers
Pages	473-476
Number of pages	4
ISBN (Electronic)	978-1-5386-9154-0, 978-1-5386-9153-3
ISBN (Print)	978-1-5386-9155-7
DOIs	https://doi.org/10.1109/IGARSS.2019.8900088
Publication status	Published - 2019
MoE publication type	A4 Article in a conference publication
Event	39th IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019 - Yokohama, Japan Duration: 28 Jul 2019 → 2 Aug 2019

Conference

Conference	39th IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019
Country/Territory	Japan
City	Yokohama
Period	28/07/19 → 2/08/19

Keywords

Coherence
GLCM
Land-cover mapping
LSTM
Morphological profile
Recurrent neural networks
Sentinel- 1
Synthetic aperture radar (SAR)
Time series

Access to Document

10.1109/IGARSS.2019.8900088

Cite this

@inproceedings{96dd14ad2d7b4d549dc4b6f15755927f,

title = "Deep recurrent neural networks for land-cover classification using sentinel-1 insar time series",

abstract = "To date, the potential of multitemporal interferometric SAR (InSAR) data in land-cover mapping has not been fully explored despite suitable time series increasingly acquired from SAR sensors. Here, we suggest to use an LSTM (Long Short Term Memory) based land-cover classifier to address this problem. Spatial context is preserved by using grey-level spatial dependencies and morphological profiles. Further, a 4-LSTM-based model was trained to capture the temporal dynamics of InSAR coherence. Altogether 39 Sentinel-1 interferometric coherence pairs acquired over Donana in Spain were used to evaluate the method performance. Achieved more than 90% overall accuracy indicates the strong potential of developed InSAR recurrent approach in improving differentiation between various land cover classes. ",

keywords = "Coherence, GLCM, Land-cover mapping, LSTM, Morphological profile, Recurrent neural networks, Sentinel- 1, Synthetic aperture radar (SAR), Time series",

author = "Shaojia Ge and Oleg Antropov and Weimin Su and Hong Gu and Jaan Praks",

note = "Funding Information: This study was supported by the National Natural Science Foundation of China (Grant Nos. 61671246). InSAR coherence time series were provided in the framework of the ESA SEOM SInCohMap project. Publisher Copyright: {\textcopyright} 2019 IEEE.; 39th IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019 ; Conference date: 28-07-2019 Through 02-08-2019",

year = "2019",

doi = "10.1109/IGARSS.2019.8900088",

language = "English",

isbn = "978-1-5386-9155-7",

pages = "473--476",

booktitle = "IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium",

publisher = "IEEE Institute of Electrical and Electronic Engineers",

address = "United States",

}

Ge, S, Antropov, O, Su, W, Gu, H & Praks, J 2019, Deep recurrent neural networks for land-cover classification using sentinel-1 insar time series. in IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium. IEEE Institute of Electrical and Electronic Engineers, pp. 473-476, 39th IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019, Yokohama, Japan, 28/07/19. https://doi.org/10.1109/IGARSS.2019.8900088

Deep recurrent neural networks for land-cover classification using sentinel-1 insar time series. / Ge, Shaojia; Antropov, Oleg; Su, Weimin et al.

IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium. IEEE Institute of Electrical and Electronic Engineers, 2019. p. 473-476.

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

TY - GEN

T1 - Deep recurrent neural networks for land-cover classification using sentinel-1 insar time series

AU - Ge, Shaojia

AU - Antropov, Oleg

AU - Su, Weimin

AU - Gu, Hong

AU - Praks, Jaan

N1 - Funding Information: This study was supported by the National Natural Science Foundation of China (Grant Nos. 61671246). InSAR coherence time series were provided in the framework of the ESA SEOM SInCohMap project. Publisher Copyright: © 2019 IEEE.

PY - 2019

Y1 - 2019

N2 - To date, the potential of multitemporal interferometric SAR (InSAR) data in land-cover mapping has not been fully explored despite suitable time series increasingly acquired from SAR sensors. Here, we suggest to use an LSTM (Long Short Term Memory) based land-cover classifier to address this problem. Spatial context is preserved by using grey-level spatial dependencies and morphological profiles. Further, a 4-LSTM-based model was trained to capture the temporal dynamics of InSAR coherence. Altogether 39 Sentinel-1 interferometric coherence pairs acquired over Donana in Spain were used to evaluate the method performance. Achieved more than 90% overall accuracy indicates the strong potential of developed InSAR recurrent approach in improving differentiation between various land cover classes.

AB - To date, the potential of multitemporal interferometric SAR (InSAR) data in land-cover mapping has not been fully explored despite suitable time series increasingly acquired from SAR sensors. Here, we suggest to use an LSTM (Long Short Term Memory) based land-cover classifier to address this problem. Spatial context is preserved by using grey-level spatial dependencies and morphological profiles. Further, a 4-LSTM-based model was trained to capture the temporal dynamics of InSAR coherence. Altogether 39 Sentinel-1 interferometric coherence pairs acquired over Donana in Spain were used to evaluate the method performance. Achieved more than 90% overall accuracy indicates the strong potential of developed InSAR recurrent approach in improving differentiation between various land cover classes.

KW - Coherence

KW - GLCM

KW - Land-cover mapping

KW - LSTM

KW - Morphological profile

KW - Recurrent neural networks

KW - Sentinel- 1

KW - Synthetic aperture radar (SAR)

KW - Time series

UR - http://www.scopus.com/inward/record.url?scp=85104600356&partnerID=8YFLogxK

U2 - 10.1109/IGARSS.2019.8900088

DO - 10.1109/IGARSS.2019.8900088

M3 - Conference article in proceedings

AN - SCOPUS:85104600356

SN - 978-1-5386-9155-7

SP - 473

EP - 476

BT - IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium

PB - IEEE Institute of Electrical and Electronic Engineers

T2 - 39th IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019

Y2 - 28 July 2019 through 2 August 2019

ER -

Deep recurrent neural networks for land-cover classification using sentinel-1 insar time series

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this