Improving Satellite Imagery Masking Using Multitask and Transfer Learning

Many remote sensing applications require masking of pixels in satellite imagery for further analysis. For instance, estimating water quality variables such as suspended sediment concentration (SSC) requires isolating pixels depicting water bodies unaffected by clouds, their shadows, terrain shadows,...

Full description

Saved in:
Bibliographic Details
Main Authors: Rangel Daroya, Luisa Vieira Lucchese, Travis Simmons, Punwath Prum, Tamlin Pavelsky, John Gardner, Colin J. Gleason, Subhransu Maji
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10925631/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850271643009548288
author Rangel Daroya
Luisa Vieira Lucchese
Travis Simmons
Punwath Prum
Tamlin Pavelsky
John Gardner
Colin J. Gleason
Subhransu Maji
author_facet Rangel Daroya
Luisa Vieira Lucchese
Travis Simmons
Punwath Prum
Tamlin Pavelsky
John Gardner
Colin J. Gleason
Subhransu Maji
author_sort Rangel Daroya
collection DOAJ
description Many remote sensing applications require masking of pixels in satellite imagery for further analysis. For instance, estimating water quality variables such as suspended sediment concentration (SSC) requires isolating pixels depicting water bodies unaffected by clouds, their shadows, terrain shadows, and snow and ice formation. A significant bottleneck is the reliance on multiple data products (e.g., satellite imagery and elevation maps) and lack of precision in individual processing steps, which degrade estimation accuracy. We propose a unified masking system that predicts all necessary masks from harmonized landsat and sentinel (HLS) imagery. Our model leverages multitask learning to improve accuracy while sharing computation across tasks for added efficiency. In this article, we explore recent deep learning architectures, demonstrating that masking performance benefits from pretraining on large satellite imagery datasets. We present a range of models offering different speed/accuracy tradeoffs: MobileNet variants provide the fastest inference while maintaining competitive accuracy, whereas transformer-based architectures achieve the highest accuracy, particularly when pretrained on large-scale satellite datasets. Our models provide a 9% <inline-formula><tex-math notation="LaTeX">$F1$</tex-math></inline-formula> score improvement compared to previous work on water pixel identification. When integrated with an SSC estimation system, our models result in a 30&#x00D7; speedup while reducing estimation error by 2.64 mg/L, allowing for global-scale analysis. We also evaluate our model on a recently proposed cloud and cloud shadow estimation benchmark, where we outperform the current state-of-the-art model by at least 6% in <inline-formula><tex-math notation="LaTeX">$F1$</tex-math></inline-formula> score.
format Article
id doaj-art-41fcf4df68a24df9a90756882c8d6719
institution OA Journals
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-41fcf4df68a24df9a90756882c8d67192025-08-20T01:52:10ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-01188777879610.1109/JSTARS.2025.355162010925631Improving Satellite Imagery Masking Using Multitask and Transfer LearningRangel Daroya0https://orcid.org/0009-0007-5309-6359Luisa Vieira Lucchese1https://orcid.org/0000-0002-8737-7171Travis Simmons2Punwath Prum3https://orcid.org/0009-0009-7085-7695Tamlin Pavelsky4John Gardner5https://orcid.org/0000-0002-1454-5074Colin J. Gleason6Subhransu Maji7https://orcid.org/0000-0002-3869-9334College of Information and Computer Sciences, University of Massachusetts, Amherst, MA, USADepartment of Geology and Environmental Science, University of Pittsburgh, Pittsburgh, PA, USADepartment of Civil and Environmental Engineering, University of Massachusetts, Amherst, MA, USADepartment of Geology and Environmental Science, University of Pittsburgh, Pittsburgh, PA, USADepartment of Earth, Marine and Environmental Science, University of North Carolina, Chapel Hill, USADepartment of Geology and Environmental Science, University of Pittsburgh, Pittsburgh, PA, USADepartment of Civil and Environmental Engineering, University of Massachusetts, Amherst, MA, USACollege of Information and Computer Sciences, University of Massachusetts, Amherst, MA, USAMany remote sensing applications require masking of pixels in satellite imagery for further analysis. For instance, estimating water quality variables such as suspended sediment concentration (SSC) requires isolating pixels depicting water bodies unaffected by clouds, their shadows, terrain shadows, and snow and ice formation. A significant bottleneck is the reliance on multiple data products (e.g., satellite imagery and elevation maps) and lack of precision in individual processing steps, which degrade estimation accuracy. We propose a unified masking system that predicts all necessary masks from harmonized landsat and sentinel (HLS) imagery. Our model leverages multitask learning to improve accuracy while sharing computation across tasks for added efficiency. In this article, we explore recent deep learning architectures, demonstrating that masking performance benefits from pretraining on large satellite imagery datasets. We present a range of models offering different speed/accuracy tradeoffs: MobileNet variants provide the fastest inference while maintaining competitive accuracy, whereas transformer-based architectures achieve the highest accuracy, particularly when pretrained on large-scale satellite datasets. Our models provide a 9% <inline-formula><tex-math notation="LaTeX">$F1$</tex-math></inline-formula> score improvement compared to previous work on water pixel identification. When integrated with an SSC estimation system, our models result in a 30&#x00D7; speedup while reducing estimation error by 2.64 mg/L, allowing for global-scale analysis. We also evaluate our model on a recently proposed cloud and cloud shadow estimation benchmark, where we outperform the current state-of-the-art model by at least 6% in <inline-formula><tex-math notation="LaTeX">$F1$</tex-math></inline-formula> score.https://ieeexplore.ieee.org/document/10925631/Deep learningglobal surface water detectionmultitask learningsuspended sedimenttransfer learning
spellingShingle Rangel Daroya
Luisa Vieira Lucchese
Travis Simmons
Punwath Prum
Tamlin Pavelsky
John Gardner
Colin J. Gleason
Subhransu Maji
Improving Satellite Imagery Masking Using Multitask and Transfer Learning
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Deep learning
global surface water detection
multitask learning
suspended sediment
transfer learning
title Improving Satellite Imagery Masking Using Multitask and Transfer Learning
title_full Improving Satellite Imagery Masking Using Multitask and Transfer Learning
title_fullStr Improving Satellite Imagery Masking Using Multitask and Transfer Learning
title_full_unstemmed Improving Satellite Imagery Masking Using Multitask and Transfer Learning
title_short Improving Satellite Imagery Masking Using Multitask and Transfer Learning
title_sort improving satellite imagery masking using multitask and transfer learning
topic Deep learning
global surface water detection
multitask learning
suspended sediment
transfer learning
url https://ieeexplore.ieee.org/document/10925631/
work_keys_str_mv AT rangeldaroya improvingsatelliteimagerymaskingusingmultitaskandtransferlearning
AT luisavieiralucchese improvingsatelliteimagerymaskingusingmultitaskandtransferlearning
AT travissimmons improvingsatelliteimagerymaskingusingmultitaskandtransferlearning
AT punwathprum improvingsatelliteimagerymaskingusingmultitaskandtransferlearning
AT tamlinpavelsky improvingsatelliteimagerymaskingusingmultitaskandtransferlearning
AT johngardner improvingsatelliteimagerymaskingusingmultitaskandtransferlearning
AT colinjgleason improvingsatelliteimagerymaskingusingmultitaskandtransferlearning
AT subhransumaji improvingsatelliteimagerymaskingusingmultitaskandtransferlearning