Deep Convolutional Networks for Cloud Detection Using Resourcesat-2 Data

Debvrat Varshney; Prasun Kumar Gupta; C. Persello; Bhaskar Ramachandra Nikam

doi:10.1109/IGARSS.2019.8898628

Deep Convolutional Networks for Cloud Detection Using Resourcesat-2 Data

Debvrat Varshney, Prasun Kumar Gupta, C. Persello, Bhaskar Ramachandra Nikam

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

2 Citations (Scopus)

51 Downloads (Pure)

Abstract

Cloud cover creates obstruction in Earth Observation studies. The obstruction is harder to distinguish from features having similar reflectance on the ground, such as snow. To distinguish clouds from snow in a VNIR image, we use an additional SWIR band. The images were fed into a deep Fully Convolutional Network that can fuse the multiresolution SWIR and VNIR bands together, in order to produce pixelwise classification. The accuracy obtained by the model on the test image was 93.35%. We compare the performance of this model with a more commonly used technique, Random Forests. To analyze the effect of SWIR, we use another deep learning model, trained only on the VNIR image, and compare the accuracies obtained.

Original language	English
Title of host publication	2019 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019
Subtitle of host publication	Proceedings
Place of Publication	Piscataway, NJ
Publisher	IEEE
Pages	9851-9854
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.2019.8898628
Publication status	Published - Jul 2019
Event	39th IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019 - Yokohama, Japan Duration: 28 Jul 2019 → 2 Aug 2019 Conference number: 39

Publication series

Name	Proceedings IEEE International Geoscience and Remote Sensing Symposium (IGARSS)
Publisher	IEEE
Volume	2019
ISSN (Print)	2153-6996
ISSN (Electronic)	2153-7003

Conference

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

Keywords

Cloud detection
Data fusion
Deep convolutional networks
LISS-4
Snow
SWIR
22/3 OA procedure

Access to Document

10.1109/IGARSS.2019.8898628

ArticleFinal published version, 1.96 MBLicence: Taverne

Cite this

Varshney, D., Gupta, P. K., Persello, C., & Nikam, B. R. (2019). Deep Convolutional Networks for Cloud Detection Using Resourcesat-2 Data. In 2019 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019: Proceedings (pp. 9851-9854). Article 8898628 (Proceedings IEEE International Geoscience and Remote Sensing Symposium (IGARSS); Vol. 2019). IEEE. https://doi.org/10.1109/IGARSS.2019.8898628

@inproceedings{16aa8b2b688144988cec5c83374942f2,

title = "Deep Convolutional Networks for Cloud Detection Using Resourcesat-2 Data",

abstract = "Cloud cover creates obstruction in Earth Observation studies. The obstruction is harder to distinguish from features having similar reflectance on the ground, such as snow. To distinguish clouds from snow in a VNIR image, we use an additional SWIR band. The images were fed into a deep Fully Convolutional Network that can fuse the multiresolution SWIR and VNIR bands together, in order to produce pixelwise classification. The accuracy obtained by the model on the test image was 93.35%. We compare the performance of this model with a more commonly used technique, Random Forests. To analyze the effect of SWIR, we use another deep learning model, trained only on the VNIR image, and compare the accuracies obtained.",

keywords = "Cloud detection, Data fusion, Deep convolutional networks, LISS-4, Snow, SWIR, 22/3 OA procedure",

author = "Debvrat Varshney and Gupta, {Prasun Kumar} and C. Persello and Nikam, {Bhaskar Ramachandra}",

year = "2019",

month = jul,

doi = "10.1109/IGARSS.2019.8898628",

language = "English",

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

series = "Proceedings IEEE International Geoscience and Remote Sensing Symposium (IGARSS)",

publisher = "IEEE",

pages = "9851--9854",

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

address = "United States",

note = "39th IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019, IGARSS 2019 ; Conference date: 28-07-2019 Through 02-08-2019",

}

Varshney, D, Gupta, PK, Persello, C & Nikam, BR 2019, Deep Convolutional Networks for Cloud Detection Using Resourcesat-2 Data. in 2019 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019: Proceedings., 8898628, Proceedings IEEE International Geoscience and Remote Sensing Symposium (IGARSS), vol. 2019, IEEE, Piscataway, NJ, pp. 9851-9854, 39th IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019, Yokohama, Japan, 28/07/19. https://doi.org/10.1109/IGARSS.2019.8898628

Deep Convolutional Networks for Cloud Detection Using Resourcesat-2 Data. / Varshney, Debvrat; Gupta, Prasun Kumar; Persello, C. et al.
2019 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019: Proceedings. Piscataway, NJ: IEEE, 2019. p. 9851-9854 8898628 (Proceedings IEEE International Geoscience and Remote Sensing Symposium (IGARSS); Vol. 2019).

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

TY - GEN

T1 - Deep Convolutional Networks for Cloud Detection Using Resourcesat-2 Data

AU - Varshney, Debvrat

AU - Gupta, Prasun Kumar

AU - Persello, C.

AU - Nikam, Bhaskar Ramachandra

N1 - Conference code: 39

PY - 2019/7

Y1 - 2019/7

N2 - Cloud cover creates obstruction in Earth Observation studies. The obstruction is harder to distinguish from features having similar reflectance on the ground, such as snow. To distinguish clouds from snow in a VNIR image, we use an additional SWIR band. The images were fed into a deep Fully Convolutional Network that can fuse the multiresolution SWIR and VNIR bands together, in order to produce pixelwise classification. The accuracy obtained by the model on the test image was 93.35%. We compare the performance of this model with a more commonly used technique, Random Forests. To analyze the effect of SWIR, we use another deep learning model, trained only on the VNIR image, and compare the accuracies obtained.

AB - Cloud cover creates obstruction in Earth Observation studies. The obstruction is harder to distinguish from features having similar reflectance on the ground, such as snow. To distinguish clouds from snow in a VNIR image, we use an additional SWIR band. The images were fed into a deep Fully Convolutional Network that can fuse the multiresolution SWIR and VNIR bands together, in order to produce pixelwise classification. The accuracy obtained by the model on the test image was 93.35%. We compare the performance of this model with a more commonly used technique, Random Forests. To analyze the effect of SWIR, we use another deep learning model, trained only on the VNIR image, and compare the accuracies obtained.

KW - Cloud detection

KW - Data fusion

KW - Deep convolutional networks

KW - LISS-4

KW - Snow

KW - SWIR

KW - 22/3 OA procedure

U2 - 10.1109/IGARSS.2019.8898628

DO - 10.1109/IGARSS.2019.8898628

M3 - Conference contribution

AN - SCOPUS:85077703256

SN - 978-1-5386-9155-7

T3 - Proceedings IEEE International Geoscience and Remote Sensing Symposium (IGARSS)

SP - 9851

EP - 9854

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

PB - IEEE

CY - Piscataway, NJ

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

Y2 - 28 July 2019 through 2 August 2019

ER -

Varshney D, Gupta PK, Persello C, Nikam BR. Deep Convolutional Networks for Cloud Detection Using Resourcesat-2 Data. In 2019 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2019: Proceedings. Piscataway, NJ: IEEE. 2019. p. 9851-9854. 8898628. (Proceedings IEEE International Geoscience and Remote Sensing Symposium (IGARSS)). doi: 10.1109/IGARSS.2019.8898628

Deep Convolutional Networks for Cloud Detection Using Resourcesat-2 Data

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this